Systems between information knowledge

Department of Political and Economic Studies University of Helsinki

Finland

Systems between information and knowledge

In a memory management model of an extended enterprise

Jouni Meriluoto

ACADEMIC DISSERTATION

To be presented, with the permission of the Faculty of Social Sciences of the University of Helsinki, for public examination in the Auditorium XII, University main building, on June 21, 2011, at 12 noon.

Helsinki 2011

© Jouni Meriluoto 2011, 2012 E‐mail: [email protected]

ISBN 978‐952‐10‐8249‐8 (Online)

ISBN 978‐952‐92‐9171‐7 (Print) Unigrafia Oy, Helsinki 2011

CONTEN

ABSTRAC

CREDITS

1 INT

1.1 1.2 1.3 1.41.4.1.4.

HMOO

1.4.1.4.

SIn

1.4. 1.5 1.6 1.7 1.81.8.1.8.1.8.1.8.1.8.

PART I – T

1 CAS

1.1 1.2 1.3

2 INFO

2.1 2.2 2.32.3.

PS

2.3. 2.4 2.5 2.6 2.7

3 ORG

3.13.1.3.1. 3.23.2.

NTS

CT .................

....................

RODUCTION .

CURRENT RESDEFINITIONS OROOTS OF KNOTRADITIONAL

1 Data and2 RepresenHuman‐readablMachine‐readabOntologies ........Other knowledg3 Knowledg4 SearchingSearch engines .ntelligent agent5 ApplicatiWHAT ARE DASOCIOTECHNICSECI AND BA .THE MISSING

1 Do we kn2 Where is3 SECI is ov4 Extended5 Processes

THE MODEL B

SE STUDY MET

GROUNDED THCASE STUDY ORESEARCH DES

ORMATION A

PARADIGMS FTHE RATIONALINFORMATION

1 InformatPhysical informaSemantics and s2 InformatPRACTICAL KNTACIT AND EXSEARCH AND RUNDERSTAND

GANISATIONA

DISCIPLINES O1 Evolution2 OrganisaMEANS FOR O

1 Organisa

.....................

.....................

.....................

EARCH – MISSINOF KNOWLEDGE OWLEDGE MANAKM SYSTEMS ANd knowledge wnting knowlede knowledge....ble knowledge .........................ge description age filtering ....g for knowled........................ts and visualisaions and their ATA? ...............CAL VIEW AND T......................LINKS ..............now our conte the memory versimplified, d enterprise ins should end –

BUILDING TOO

THOD BASED

HEORY PRINCIPLEON GROUNDED THSIGN – THE CASE

AND KNOWLE

ROM PHILOSOPHL LOAD OF KM ..N BETWEEN DATAtion species baation and syntasemiotics ..........tion species baNOWLEDGE CAN NPLICIT KNOWLEDRETRIEVAL IN THDING RETRIEVED I

AL MEMORY .

OF ORGANISATIONn or design? ...ational memorORGANISATIONALational Memor

.....................

.....................

.....................

NG LINKS FROM TMANAGEMENT .AGEMENT ........ND THEIR PREDECwarehouses, adge .........................................................................................attributes ...............................dge .........................................ation models ....development.....................HE ACTOR‐NETW..........................................ent – what is twe have lost?ba has got cun a real actor‐n– all knowledg

OLS...............

ON GROUND

ES ..................HEORY ............ES ..................

DGE SPECIES

HY AND SCIENCE .....................A AND KNOWLEDased on probaax .............................................ased on qualitNOT BE PASSED BDGE ARE BOTH DYE CONTEXT OF INNFORMATION ..

.....................

NAL MEMORY .........................ry mechanismL MEMORY .......ry Processes ..

.....................

.....................

.....................

THE STATE OF THE..........................................CESSORS ..........and knowledg...............................................................................................................................................................................................................t tools ................................

WORK THEORY .............................................the difference ? ....................ultural boundanetwork withoge is not the s

.....................

DED THEORY P

.....................

.....................

.....................

....................

.....................

.....................DGE ................ability ............................................................tative interpreBY ..................YNAMIC AND CONNFORMATION SPE.....................

.....................

.....................

.....................m and content ..........................................

.....................

.....................

.....................

E ART .............................................................................e discovery ...........................................................................................................................................................................................................................................................................................................................between data.....................aries ..............out explicit‐tasame .............

.....................

PRINCIPLES ....

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................................................................etation ...............................NSTRUCTED ......ECIES AND MEAN.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

........................

........................

........................

........................

.....................

.....................

........................

........................

.....................

.....................

.....................

.....................

.....................a, information..........................................acit dichotomy.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

........................

........................

.....................

.....................

.....................NING ....................................

.....................

.....................

.....................

.....................

.....................

.....................

....................

....................

....................

......................

......................

......................

.......................................................................................................................................................................................................................................................................................................................................................................n and knowled..........................................y .........................................

....................

....................

......................

......................

......................

....................

......................

......................

........................................................................................................................................................................................................

....................

...........................................................................................................

.................. 7

.................. 9

.................. 8

................. 11

................. 12

................. 15

................. 17

................. 20

................. 21.................... 21.................... 21.................... 22.................... 22................. 22................. 23.................... 23.................... 24................. 25................. 26................. 28................. 29................. 32dge? ......... 33................. 33................. 35................. 35................. 36

................ 37

................ 39

................. 39

................. 41

................. 45

................ 49

................. 50

................. 51

................. 52

................. 53.................... 53.................... 54................. 56................. 61................. 64................. 66................. 67

................ 69

................. 70

................. 70

................. 71

................. 74

................. 74

7

9

8

1 7 0 1 1 2 2 2 3 4 5 6 8 9 3 3 5 5 6

7

9

9 1

9

0 1 3 4 6 1 4 6 7

9

0 0 4 4

CONTEN

3.2. 3.33.3. 3.43.4.

PART II ‐

INITIAL C

HOW THTHE FO

1 CAS

1.1 1.2 1.31.3.1.3.1.3.1.3. 1.41.4.1.4.

2 CAS

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.82.8.

DInCWL

2.8.WWSWAWDSL

2.8.DTCVTSC

NTS

2 Bins of orORGANISATIO

1 User goaCONCLUSIONS

1 Summary

INITIAL CASE

CASE STUDIES

HE CASES WERE RCAL COMPANY A

SE I – INDIVID

SCOPE OF CASCONCEPTS ....CURRENT SITU

1 Phase 1 –2 Records a3 Phase 2 –4 ConclusioFUTURE ‐ TRE

1 Future tr2 Threats a

SE II ‐ ORGANI

CASE II, PHASCASE II, PHASCONCEPTS ....BACKGROUNDCASE II PHASESELECTING DOCASE II, PHASCASE II, PHAS

1 CommonDocument manantegrated vs. nCompound vs. nWorkflow (7.1alLifecycle ...........2 Very impWindows ClientWeb servers andSearch results dWorking and puAudit trails (6.1aWorkflow (7.2alDocument manaSecurity integraLifecycle (7.2all)3 ExceptionDevelopment toTypes of data (3Check‐out / cheVendors of docuTools for filterinSearch from comCross‐repository

rganisational ONAL MEMORY SYals and media .S FROM ORGANISy and missing

STUDIES ......

BASED ON G

REACHED..........AS AN EXTENDED

UAL KNOWLE

SE I .......................................UATION OF APPLI– System persand archiving– Users’ perspons from the cNDS, THREATS Arends ..............and solutions .

ISATIONAL KN

E 1 – USERS’ PERE 2 – SYSTEM PE......................D ....................ES ...................OCUMENT MANAE 1: USERS’ PERSE 2: SYSTEM PERn questions .....agement systemon‐integrated dnon‐compound ll) .............................................

portant features (3.1.1a) – Arcd thin clients (3isplay options (ublic versions (5a) – control ......ll) as a very impagement systemtion (6.2a) – co) as a very impons ..................ools (8u) – an in3.1u) – an invenck‐in (5.1u) – aument manageng (4.2m) – an ampound documy searching (4.1

memories .....YSTEMS .................................SATIONAL MEMOlinks concern

.....................

ROUNDED TH

.....................D ORGANISATION

EDGE FROM D

.....................

.....................CATIONS AND COspective .......... .....................pective ...........current situatiND SOLUTIONS .............................................

NOWLEDGE F

RSPECTIVE .......ERSPECTIVE .....................................................................AGEMENT SYSTEMSPECTIVE – RESURSPECTIVE – RESU......................ms in general (2document mandocument man................................................es ..................chitecture ........3.1m) – architec(4.2u) – docum5.1m) – library S........................portant featurem in general (2ontrol ...............ortant feature ........................nvention procesntion and learni legal process ement systems (administrative pments (4.1a) – a1m) – legal, lear

.....................

.....................

.....................ORY ................ning this resea

.....................

HEORY ...........

.....................N ....................

DATA MANAG

.....................

.....................ONTENT .........................................................................ion of case I a...............................................................

ROM INFORM

.....................

.....................

.....................

.....................

.....................MS AND A PROJECULTS FROM THE INULTS FROM STRU.....................2.1all) ‐ Systemagement systenagement syste.............................................................................................cture ................

ment list: searchServices ................................... ........................1all) .....................................................................................ss and projects ing process excexception: View(8.2m) – a learnprocess exceptin invention prorning and admi

.....................

.....................

.....................

.....................arch ...............

.....................

.....................

.....................

.....................

GEMENT SYST

.....................

.....................

.....................

.....................

.....................

.....................applications ..................................................................

MATION MAN

.....................

.....................

.....................

.....................

.....................CT SCOPE .........NTERVIEWS ......UCTURED NUMER..................... background ...m. (2.2all) ‐ sysem (2.3all) ‐ sys..................................................................................................................... and retrieval .......................................................................................................................................................................exception: cuseption .............w checked‐out dning process exion ..................ocess exceptionnistrative proce

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

TEMS .............

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

NAGEMENT SY

.....................

.....................

.....................

.....................

.....................

.....................

.....................RIC EVALUATION .............................................stem backgrounstem backgroun..................................................................................................................................................................................................................................................................................................................tomised applic........................documents ......ception ...................................n .......................ess exception ..

...........................................................................................................

....................

....................

......................

......................

....................

......................

......................

..........................................................................................................................................................................

YSTEM ..........

......................

......................

......................

......................

......................

......................

......................QUESTIONNAIRE.............................................nd ....................nd......................................................................................................................................................................................................................................................................................................................................ations .............................................................................................................................................................

................. 76

................. 77

................. 78

................. 80

................. 82

................ 84

................ 85

................. 86

................. 87

................ 90

................. 91

................. 92

................. 95

................. 97

................. 99

............... 100

............... 101

............... 101

............... 101

............... 102

.............. 104

............... 104

............... 105

............... 106

............... 106

............... 107

............... 108

............... 109ES ............ 114............... 116.................. 116.................. 117.................. 119.................. 121.................. 122............... 124.................. 125.................. 126.................. 126.................. 127.................. 128.................. 128.................. 130.................. 130.................. 130............... 131.................. 131.................. 132.................. 133.................. 133.................. 133.................. 134.................. 134

6 7 8 0 2

4

5

6 7

0

1 7 9 0 1 2

4

4 6 6 7 8 9 4 6 6 7 9 1 2 4 5 6 6 7 8 8 0 0 0 1 2 4 4

CONTEN

SSNOS

2.8.InLPALCT

3 CAS

3.1 3.23.2.3.2.3.2. 3.33.3.3.3. 3.43.4.3.4.3.4.3.4.3.4.3.4.3.4.3.4.3.4. 3.53.5.3.5.

CSNDA

3.5.EC

3.5.

PART III –

1 FIND

1.1 1.21.2.1.2.1.2. 1.31.3.

M

NTS

Security Levels (Standards (8a) –Notifications (6.Objects as attribSearch methods4 Phase 2 snvention proceLearning procesProjects ............Administrative pLegal process....Case II, phase 2 The case organi

SE III ‐ EXTEND

OBJECTIVES ...EXTENDED OR

1 Social co2 Culture a3 Findings CASE III, INITI

1 Who kno2 ConclusioCASE III, PHAS

1 From org2 Processes3 Extended4 Organisa5 Knowledg6 Knowledg7 Knowledg8 Trust, cul9 SummaryCASE III PHAS

1 Requirem2 KnowledgCollaboration bySystems behind Notifications anDetective work Access to docum3 CommunEveryday presenCorrespondence4 Conclusio

– MODELLING

DINGS FROM

CASE I – MANCASE II – MA

1 Suggestio2 Suggestio3 Case II fuCASE III – MA

1 Results ofMeta‐data on sk

(6.1m) – a Lega– an invention, .1u) – a learningbutes (2.4a) – as (4.1u) – no exsummary .......ss.....................ss ..............................................process ....................................exceptions ......sation on avera

DED ORGANIS

......................RGANISATION .....nstructivism aand cooperatiowithin the socAL SURVEY – STAows what – preons from the sSE 1: INTERVIEWganisational tos and practiced organisationational memorge capture ange creation ange sharing .....lture, educatioy of most impE 2: ONE DEDICAments for casege sharing – wy systems ......... the portal .......d traceability ...– searching andment bases .......nication – fromnce vs. virtual me via e‐mail andon on most im

G THE RESULT

THE PREVALE

NAGEMENT OF DANAGEMENT OF INons and resultons and resultuture ..............ANAGING THE KNof case III part kills ..................

l process exceplegal and learng process excepn Invention, legception .................................................................................................................................................................................age ..................

SATIONAL KN

.....................

.....................and learning ..on ..................cial loci of tecATE OF THE ART .eliminary intestate of the arWS INCLUDING EXo social structes in new internal networks ..ry content in pnd reuse .........nd feedback .........................on and languaortant notes ..ATED KM SYSTEMe III KM systemwith or withou........................................................................d finding ..................................m internal techmeeting via comd lighter commumportant obse

S ..................

ENT USAGE O

ATA FROM BOTHNFORMATION INts from case IIts from case II......................NOWLEDGE OF AN1 – extended........................

ption ................ing process excption ................gal and adminis.....................................................................................................................................................................................................................

OWLEDGE AN

.....................

.....................

.....................

.....................chnological pra.....................erviews ..........rt of the focal XTERNAL COLLABtures through r‐organisation.....................practice .......................................................................age ....................................M – WISE PORTAm ...................ut systems ............................................................................................................................hnology and mmmunication teunication chanrvations ........

.....................

OF SYSTEMS ...

H THE INDIVIDUALN AN ORGANISATII, phase 1 ......I, phase 2 ...........................N EXTENDED ENTd organisation........................

........................ception. ...................................strative process.....................................................................................................................................................................................................................

ND A KM SYST

.....................

.....................

.....................

.....................actice ......................................................company ......ORATORS ........the extendednal networks ....................................................................................................................................................AL ....................................................................................................................................................................................meetings towaechnologies ......nels ......................................

.....................

.....................

L AND GROUP VIIONAL UNIT .....................................................................TERPRISE ..........al cooperation........................

........................

........................

........................s exception ...........................................................................................................................................................................................................................

TEM ..............

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................d enterprise ...................................................................................................................................................................................................................................................................................................................................................................ards extended.....................................................................

.....................

.....................

EWPOINT ..................................................................................................................n ...........................................

........................

........................

........................

........................

.....................................................................................................................................................................................................................

....................

......................

................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................d organisation.....................................................................

....................

....................

......................

........................................................................................................................................................

.................. 135

.................. 135

.................. 136

.................. 136

.................. 137............... 137.................. 137.................. 139.................. 140.................. 142.................. 144.................. 145.................. 146

.............. 148

............... 148

............... 150

............... 151

............... 152

............... 153

............... 156

............... 156

............... 158

............... 160

............... 161

............... 162

............... 163

............... 164

............... 165

............... 166

............... 167

............... 168

............... 169

............... 169

............... 171

............... 173.................. 174.................. 174.................. 176.................. 177.................. 179n ............. 179.................. 180.................. 181............... 183

.............. 184

.............. 185

............... 185

............... 188

............... 189

............... 190

............... 191

............... 192

............... 193.................. 194

5 5 6 6 7 7 7 9 0 2 4 5 6

8

8 0 2 3 6 6 8 0 2 3 4 5 6 7 8 9 9 3 4 4 6 7 9 9 0 1 3

4

5

8 9 0 3 4

CONTEN

C1.3.

SCS

1.3.

2 BUI

2.12.1. 2.22.2.2.2.2.2.to n2.2.2.2.2.2.2.2.2.2.an e2.2.2.2.2.2. 2.2.2.2.2.2. 2.3 2.4

3 STR

3.1 3.2 3.3

CONCLUS

SYSTEMS

REFEREN

APPENDI

APPENDI

APPENDI

ProcNeeCurr

APPENDI

NTS

Conclusions from2 Results frStoring and retrCommunicationSystems ............3 Case III –

LDING THE M

TOWARDS EXT1 The mostSTATEMENTS

1 Memory 2 The set o3 Informatnormalised da4 The know5 Knowform6 Commun7 The bliss 8 Knowledgend ..............9 Culture e10 A pres11 A socio

.........12 Mobile13 An ext14 The SEA MEMORY MADAPTING TH

RATEGY BASED

COMMUNICATTACTICAL APPSTRATEGY AN

SIONS ...........

S BETWEEN IN

CES ..............

X 1: ACRONY

X 2: SYSTEMS

X 3: QUESTIO

cesses, informeds and problerent tools and

X 4: LIST OF T

m case III, phasrom case III, prieving informatn media and ‘kn........................– final conclusi

MODEL – INFER

TENDED ORGANISt important orON A MEMORY Mis the right m

of informationtion systems sta and informwformation comation is mainication can onof ‘knowledgge always has......................expands to colsentation layeotechnical arc......................e communicattended enterpECI model andMANAGEMENT MOE MODEL TO SYS

D ON THE MO

TION CREATES KNLICATION OF THED TACTICS ........

.....................

NFORMATION

.....................

YMS AND ABB

S INVESTIGAT

ONNAIRES .....

mation search ems ...............d the future ....

TABLES AND F

se 1 – users’ viephase 2 – KM stion .................owledge sharin........................ions ...............

RENCES FROM

SATIONAL MEMOrganisational MANAGEMENT Mmetaphor ........n systems shouhould be integ

mation ............oncept is requinly organisatinly be suppore sharing’ is ds a context, w......................llaborators aner is needed ....chitecture is n......................tion technologprise should ind ba concept aODEL OF A AN EXTEMS BETWEEN

ODEL: LONG T

NOWFORMATIONE MEMORY MAN.....................

.....................

N AND KNOWL

.....................

BREVIATIONS .

TED ...............

.....................

and distributi............................................

FIGURES ........

ewpoint ............systems ................................ng’ .................................................................

M CASES STRU

ORY – INFORMATbin is informa

MODEL OF AN EXT.....................uld be extendegrated to offe.....................ired between ional short terted, not manadubious .........while informati.....................nd knowforma.....................eeded for new.....................gy enables ubnteract for learare valid for onXTENDED ENTERPINFORMATION A

TERM SYSTEM

N FOR SYSTEMS .AGEMENT MODE.....................

.....................

LEDGE: DISCU

.....................

.....................

.....................

.....................

ion ..........................................................

.....................

..........................................................................................................................................

UCTURING TH

TION SYSTEMS ASation systemsTENDED ENTERPR.....................ed .................er role‐based, .....................information arm memory ...aged...................................ion can be mo.....................ation ...................................w search and r.....................biquitous interrning and usene kind of knoPRISE ..............AND KNOWLEDGE

MS AND FURTH

.....................EL .......................................

.....................

USSION ..........

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

........................

.....................

........................

........................

........................

.....................

HE MODEL ......

S THE NUCLEUS ......................RISE (M3.EXE) ............................................ubiquitous ac.....................and knowledg...............................................................odular; the cre...............................................................retrieval and .....................raction to forme standards foowledge creat.....................E ....................

HER RESULTS .

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

.....................

..........................................................................................................................................

....................

...........................................................................................................ccess, and on‐l.....................ge .................................................................................eation process...............................................................other future a.....................mulate knowfor collaboratiotion ............................................................

....................

......................

......................

......................

....................

....................

....................

....................

....................

....................

.....................

.....................

.....................

....................

.................. 194............... 195.................. 196.................. 197.................. 199............... 200

.............. 202

............... 204

............... 206

............... 207

............... 208

............... 209line visibility ............... 209............... 210............... 214............... 216............... 217s should have ............... 219............... 220............... 221applications ................. 223ormation 225on ............ 225............... 227............... 229............... 235

.............. 238

............... 238

............... 239

............... 242

.............. 246

.............. 251

.............. 254

.............. 265

.............. 267

.............. 270

............... 270

............... 271

............... 271

.............. 278

4 5 6 7 9 0

2

4 6 7 8 9

9 0 4 6 7

9 0

3 5 5 7 9

8

8 9

6

1

4

5

7

0

0

8


ABSTRACT

What is the gap between information processed in systems, and organisational human knowledge? Is the gap insurmountable, or can the information subspecies be analysed and selected, so that a section of these two sets will be found, or so that the union of information and knowledge complement each other?

The research question of this thesis is how can knowledge be managed with information systems? The purpose is to investigate knowledge management (KM) and information systems (IS) in order to discover organisational models and strategies of knowledge in an extended organisational memory. The challenge is to find a golden mean between knowledge management and information systems. The result is a model of managing knowledge with information systems in an extended organisation. It constructs findings from various information systems and types of knowledge. Three case studies are based on an application of grounded theory. They lead to an extrapolated organisational strategy based on a model.

The research involves planning an applicable model for a future based on existing, historical reality. The scientific model is instrumentalistic, not explanatory, and based on practical case studies. Although a theory will not be formulated, new concepts and propositions – together with rules in the model – will be presented.

Human knowledge is always the final aim of all data and information. A philosophical classification of knowledge is employed as a background. Positivism, post‐positivism, and critical theory are briefly presented. Although this thesis starts from an objective comparison, it recursively regards constructivism as the most appropriate viewpoint on knowledge.

This research extends to socially constructed knowledge. It deals with the organisational level of knowledge in discourse communities by default. Because knowledge is not synonymous with information, information systems alone can not deliver knowledge management. The social nature of knowledge gives uncodifiable content and contexts to the creation and utilisation of knowledge with information systems. Information systems, however, can support management of knowledge. To illustrate this, the research will present a memory management model of an extended enterprise (See Figure 1).

Figure 1. Introducing a memory management model of an extended enterprise (M3.EXE)


The key dimensions of the model are horizontal and vertical: vertical processes include data and information that can be managed, while communication can only be supported by infrastructure. Horizontally, the right hand side contains systems (applications, platforms, technology, systemised human knowledge), and the left hand side content. Both are independent from each other.

The model is based on long research, which covered 1) the theoretical background of knowledge management, 2) case studies on organisational knowledge management systems 3) the results of systems for the management of knowledge. Information systems were examined mainly from the users’ point of view. The results provide advice on how to 1) combine theories of management of knowledge with information systems, and 2) progress via existing systems and development platforms towards organisational applications that are generic enough for collaboration. The model could be used to analyse what kinds of systems and knowledge are compatible as a framework for forming a strategy. The types of information and knowledge types are modelled more deeply than in the existing literature on KM, and information systems from data, information and knowledge categories are analysed to show which features are useful for management of knowledge.

In the case studies, the current status of the literature and field of knowledge management information systems are reviewed. An evaluation of management applications from the knowledge workers’ point of view is the main theme. The question is, how can so called data, information and knowledge management applications be used for management, and is there actually any difference between them? The focus moves from organisational knowledge management and administration towards information system strategies and models. In the organisational context, knowledge management has always existed by some name or another. It is, however, a current interest for many organisations, partly because of the long awaited communication and knowledge storing tools. Here the 'system' concept could be replaced with ‘infrastructure’, ‘platform’, or the like, because now one must distinguish between the system as a tool, represented in the first case, and the system as an organisation‐specific application, as in the last case.

Knowledge management systems could be divided into categories in many ways. Information systems, as well as knowledge management itself, are divided into management of information and management of people. This research focuses more on management of information from the organisational perspective. Management of people is dealt with through systems such as communication, collaboration and knowledge creation. In this research, systems and the cultural environment for knowledge sharing will be combined. Knowledge storing tools play the main role here; while computer mediated communication systems are also significant. Information sharing within and between organisations is another approach to the systems.

The systems are meant for management of various kinds of knowledge, information and data, while their content is classified as knowledge. Here this is seen from two sides: First, the acquisition of information systems from knowledge management theory, and then the selection of knowledge management from information systems. Knowledge is one of enterprises’ most important assets. The recognition of this has fuelled interest in comprehensive approaches to the basic activities of knowledge management: the identification, acquisition, development, dissemination, utilisation and preservation of the enterprise's knowledge. So far, enterprises have addressed knowledge management from either a management or a technological point of view.

This research is based on a hybrid solution, where effective knowledge management models are reached with both people and technology. It will also concentrate on how theoretically valuable conceptualisations from the management of knowledge in an extended organisational memory could be used in practice. Organisational memory contains information and knowledge, and the systems have proved to be the backbone for certain kinds of knowledge. However, in so far as the systems deal mostly with containers, and the knowledge management literature with applied knowledge. This practical knowledge can not be managed with systems alone. In normal knowledge‐related work, the information species are not recursively analysed. Also tacit vs. explicit knowledge is not the right division; a new concept is needed. Knowformation will formulate in communication through the information in the system, as it is combined with personal and organisational memory, together with other information from the environment, and finally interpreted or even constructed differently.


CREDITS

I would like to thank the following people, who have supported this work in one way or another: Turo Virtanen, Ilkka Heiskanen, Niilo Kauppi, Ilkka Tuomi, Arnd Rothe, Guy Boy, Osmo Vikman, Marko Nieminen, Mikael Johnson, Päivi Pöyry, Markus Dürstewitz, Markus Helmke, Frithjof Weber, Jonathan Feinberg, Jonathan Sage, my family, friends, colleagues, fellow students, and many others.

Copyright © 2011 by Jouni Meriluoto. All rights reserved. No part of this material may be reproduced or used without prior written permission from the author. To order your copy, please contact [email protected].


8

1 INTRODUCTION

This research on management analyses the various categories of knowledge management theory. The research topic is relevant to administration and organisational science, focusing on information systems. It is an interface between organisational administration and information technology. Knowledge management research has been developing the foundations for a theory with different focal points. Still, there are varying interpretations of the knowledge management discipline, and various information systems and application development tools have been marketed for knowledge management. The gap between fuzzy theories and oversimplified practice demonstrates the need for studies such as this.

The history of knowledge management has reached the third wave, where the content is the main issue. Information systems for the knowledge management of an extended organisation can be successfully researched with grounded theory based case study method. The methodology used for this kind of research is meant to build rather than test a model or a theory (Strauss & Corbin 1998: 13).

Various species of information will be analysed more deeply in this study. ICT consists of information processing and communication technologies. The same themes can be found in philosophy. Information theory gives us quantitative classes based on probability, while semantics leads us to qualitative information categories. Information and communication theory exist side by side with systems theory. Systems analysis is an engineering discipline based on this theory of the nature of systems. This framework of analysis for studying and modifying the world is used as the background for the case systems.

Knowledge management has become increasingly important for both academics and practitioners. The subject is multi‐disciplinary and covers a broad range of fields. As the relationship between information, knowledge and management has become closer, contributions have been made by disciplines such as information systems, strategy, organisation theory and human resources (Alavi & Leidner 2001). The term knowledge management became more widely used in the mid eighties, but the first article on it was written ten years earlier, and attempts to implement it in organisations ten years later. Now KM is a long term strategic field. (Examples from Barnes 2002 to Hackney 2008.) The current approach to the convergence of knowledge management and information systems is mainly referred to as content management.

Managing organisational knowledge has become increasingly attractive especially in economic life, as companies’ potential for success is closely tied to their ability to innovate and develop their methods of production; all this is based on the knowledge assets they possess (Davenport & Prusak 1998). KM focuses on facilitating and managing knowledge‐related activities, such as systematic, explicit and deliberate building, renewal and application of knowledge in order to maximise an enterprise’s knowledge‐related effectiveness and returns from its knowledge assets (Kemp et al. 2002). For inter‐organisational networks KM should include a process enabling successful collaboration (Klueber et al. 2000). As this research will show, however, organisational knowledge management – even with information systems – is more than managing business resources.

The ability to use intellectual resources and create new solutions for human needs is central to the global information economy. Human knowledge and capabilities have always been at the core of value‐creation, but now during the information age the intellectual component of work has become increasingly important (Zuboff 1988). The ‘resource’ based thinking is no longer relevant when we deal with dynamic organisational knowledge and do not see personnel merely as containers of information. However, when information systems are seen as a core part of that knowledge, it is well justified to research the substance of KM with methods used here. Evolution of the substance area requires this kind of research. The crucial need is to build a model or strategy for an extended organisation – with which it can utilise various systems to manage the unexplored zone between information and knowledge. Business benefits are a consequence of – not the primary reason – for knowledge management.

Knowledge management research has emphasised the importance of learning and it is true that contemporary organisations need to be flexible and responsive in our changing world (Drucker 1995, 1988; Teece, Pisano and Shuen 1997). According to Peter Drucker, land, labour and capital – the classical factors of production – have (long ago) become secondary to knowledge as the primary


9

resource for the new economy (Drucker 1992). Industrial research and development in particular are facing the challenges of a global economy and rapid technological change. A global organisation with its collaborators and other interest groups is influenced by rapid political, commercial, and market changes (Dam 2001).

As the importance of the work force grows (Naisbitt 1994, Mazarr 2001), organisations are assessing the value of employees’ knowledge. Knowledge is seen as outstripping the previously mentioned traditional resources and is considered the key source of a comparative or competitive advantage (Grant 1996; Swan and Newell 2000). One of the hypes of the 1990s was measuring human resource value (e.g. Journal of Human Resource Costing and Accounting); another ‘intellectual capital’ (e.g. Hudson 1994). These business‐driven materialisations contradict the generic importance of the work force and lead to artificial calculations on maximised profit. They are not discussed in this book. It is solely extrapolated that efficiency and effectiveness of knowledge work also bring financial results.

Technology also plays an increasingly important role among organisations and their employees. Organisational knowledge and new technologies give impetus to strategy. (Barnes 2002: 1) However, current research lacks the prerequisite of learning, knowledge creation, and all knowledge management – memory. Memorising the right issues is claimed to be as important as learning in general. On the other hand, organisational memory can utilise technology more effectively.

To compare organisational assets with the modern structure of corporations, one case study in this research is investigating extended organisational knowledge. The development of information and communication technologies its part has contributed to the creation of new forms of working. It has become possible to distribute work not only between different offices but also between different companies. People may belong to several teams or work groups that more often are formed on an inter‐organisational basis.

Cooperative organisations and also virtual communities have significant possibilities of enhancing the way acquisition and application of knowledge occur (Dyer, J.H. & K. Nobeoka 2000; Bieber et al. 2002). Today’s environment of multi‐organisation collaboration in technology alliances calls for improved knowledge sharing and processing capabilities, while information technology has been steadily increasing in importance as a tool of KM. (Stein & Zwass 1995). Virtual work is usually done in virtual teams that are created to serve a specific purpose; the teams are dissolved as soon as the goals have been met (Watson‐Manheim & Belanger 2000). In virtual teams the social processes are relatively similar to the ones in traditional teams, but the distance and the use of technology as a mediator of information and communication create challenges (Blanchard & Markus 2000). This kind of complex, distributed or virtual work requires specific technologies providing support for efficient communication (Watson‐Manheim & Belanger 2000). The risk involved with this new kind of instant appearance and disappearance of collaborating groups is organisational amnesia.

What then is the relationship between information systems and organisational knowledge? This dissertation is the result of over ten years of theoretical and empirical study of knowledge management and information systems, combining theoretical literature with three practical case studies of systems for management of knowledge. The theory – together with the experimental case study and constant ethnographic field observations – deals with the organisational level. The case study method, based on grounded theory, is utilised to reach the optimum results of both the interviews and research on the systems. Grounded theory, with ethnographic case studies like these, has not been used to research the information systems and extended organisational knowledge and memory. According to Myers and Newman (2007) even the qualitative interview has been a largely unexamined craft in IS research. In this study, the ethnographic discipline is also objective and pure because of the distance in time: The researcher is independent from the cases and organisations, and complements the retrospective results from the cases with constant observations. The case studies of this research used evaluative methods to guide towards deeper interviews and field observations. The scientific paradigms thus vary from positivist to constructivist (or naturalist) paradigms. Evaluative parts can be seen as design‐science studying construction, where constructs form new concepts of a domain, and a model is a set of propositions or statements expressing relationships among constructs. On the other hand, the whole research question could be interpreted as ‘is it possible at all to manage knowledge with information systems’, what was studied in more holistic level during the years.


10

Even though the number of relevant studies has increased since 1990’s, the subject, information systems for management of knowledge, has not been previously studied on an extended organisational level with this kind of hybrid solution. Although some valuable findings exist (as presented in this introduction) the case study method based on grounded theory has not been used to study information systems from the viewpoint of extended organisations and information systems – including the memory concept. Also, based on the constant follow‐up of scientific, technical and business articles, knowledge concept is analysed more deeply than in earlier KM studies. Existing research either concerns technical models mainly based on engineering or computer science, or concentrates on the human being on an individual, organisational or social level, and ignores information systems as a subject. Knowledge is also confused with communication and the media. However, some tangential articles have been released, and they are referenced to and criticised in the next section, Current research – Missing links from the state of the art.

When beginning this study by literature review, in the end of 1990’s, there was a big demand for empirical case study based on other than computer technology. Even though some review articles, practical case studies, and business KM studies on market prospects, business needs and technological trends have been published during the years, there is still this demand.

We can take a look on a typical review and interpretation of KM literature, such as Review: Knowledge management and knowledge management systems: conceptual foundations and research issues (Alavi & Leidner 2001). It is a representative sample on high‐quality article on knowledge management. The year of publication is quite irrelevant here, as technology is in minor role and knowledge in major role in this kind of research.

Compared with this study, there are some differences. The article is based on business. A firm should have a utility value from application of a ‘resource’. This resource called knowledge is first seen in the top of a traditional data‐information‐knowledge hierarchy. Alternative perspectives are somewhat presented, first referring to ‘justified belief that increases an entity’s capacity for effective action’. Six perspectives lead to ten taxonomies of knowledge. Different kinds of concepts, categorisations and results are derived from this research.

According to the article, impact of IT on knowledge exchange is questionable. Agreed, and impact of IT is researched in this study, but observations and interviews during this research have led to avoid such concepts as ‘exchange’ or ‘knowledge transfer’. Also, systems in this research include long term perceptions of the use of data, information and KM systems, together with investigation on vast amount of alternative systems. KM branch seldom impugn the meaning of ‘knowledge management system’.

Together with knowledge sharing related ‘transfer’, other organisational KM processes are presented in the article. ‘Knowledge creation’ is based on Nonaka’s tacit‐explicit, SECI (Socialisation, Externalisation, Combination and Internalisation) and ba concepts. All of these have been investigated in this research. Knowledge storage and retrieval processes use organisational memory research results as they are. That is fine, but these OM concepts are upgraded in this research. The last process in the article is knowledge application. Already the name shows that this research is required, as applied knowledge is not a presentation of epistemic knowledge. Even the ‘justified true belief’ in most KM literature is in fact doxastic – not necessarily justified. All of these processes in the article depend on the ‘type of knowledge’. Whether the reason is a categorisation, that should be renewed, is a big question mark.

Also other concepts and systems mentioned in the article – such as individual, group, and organisational levels (extended here to inter‐organisational level) meta‐data, Lotus Notes, informally organised knowledge in communications — have been empirically researched in this study, with a value added. In their summary, Alavi and Leidner state that ‘to improve knowledge management, utilising information technology implies attention not only to improving the individual and group level processes of knowledge creation and storage, but also improving the linkages among and between groups’. Might be veracious, but how ‘knowledge creation’ or such ‘linkages’ actually relate to knowledge, will be discussed in this study. Alavi and Leidner present five categories of research issues. Some of those details (collective knowledge base – applications like wikis; contribution barriers, hegemonies of knowledge, incentives; contextual details – including here links and tags; ‘dynamic’ knowledge; transfer

Systems

in the insearchinand techperformincluded– and no

Cu1.1

the concthreats. managerway to pbetter acthe currecritical oinformat

Knowled1999), aquestionshould bknowled

– knowinimportanthe value

philosopguide uslanguage

culture ainformateverythi

research(managesociotec

proposethrough context fmoderatbe comb‘justifiedhelps to

systems knowledare pres

between inf

itiative – nowng for knowlehnical attribuance, probled in this studyot only this p

urrent resea

Basedcept of knowIt is natural trs have stronproceed. It coccepted, whent situationoverview of ttion systems

First, tdge is seen and is talked aning what it ibe categorisedge species.

The seng what the nce (Swan &e of informa

Third,phy and sciens to make orges, the word

Knowand process,tion systemsng.

Amonh. It combineement of infohnical appro

To hed a model coconversionsfor knowledgtors for the kbined to cread true belief’generate be

Otherwith a histo

dge, as well aented first.

formation an

w in extendeedge – now iutes, knowleem solving any. The main particular art

rch – Missin

d on interviewwledge managto distance ong opinions oould be that en they becon of knowledthe missing lis used for it. the current rs just ‘econoabout in termis about and ed for better

econd missinorganisation

& Newell 200tion is also p knowledge nce. In this reganisational ‘memory’ isledge manag supported bs is often mis

ng others, a ses the advantormation) anoach has its olp organisationsisting of es between tage creation. knowledge cate knowledg as the episteliefs. r issues covery of knowleas all relevan

nd knowledg

ed organisatincluding knodge ‘transfend decision mpoint is thaticle, but also

ng links from

ws and othegement, mooneself fromon knowledgalternative wome more fage managemink between

research doeomic ideas’ (Wms of stockpwhether it c results. This

ng link is the n knows and 0). The valuepresented in storages havesearch, botmemory inds homonymogement shouby technologsinterpreted.

sociotechnicatages of exisnd the humaown roots, wons understelements likecit and expliTheir researreating procge. They try temological d

red here areedge managent open ques

ge

11

ional level; ITowledge prerred’ into onmaking, IT ba little empirio many missi

m the state of

r discussionsstly because

m something ge managemworldviews –amiliar. To cment researc organisation

es not sufficiWiig 1997) opiles, reservocan actually bs categorisat

lack of undeunderstande of meta‐knChapter 2.ve not been h content andependent frous for both tuld be accomgy. Knowledg. The system

al approach ting approacnistic approa

which are brieand knowlede the SECI pricit knowledgrch emphasisess. Accordito manage ‘kdefinition of

e data and infement systemtions. Defini

T usage – nosentation insn organisatioased ‘knowlecal work hasng links dem

f the art

s, it looks likee they lack exunfamiliar. Fent, and ass– with their flarify genericch: theories, nal knowledg

ently define or ‘intellectuaoirs, exchangbe managedion is offere

erstanding thing the contenowledge wil

analysed as nd its media rom the contthe media an

mplished throge is seen as ms are either

to knowledgches, such asach (leadersefly presentedge creationrocess, whichge. They alsoses knowledgng to their reknowledge cknowledge.

formation, pms, actor‐netions of know

ow under a crstead of metn, ‘knowledgedge creatios been undermand new res

e many manaxpertise on itFor the sameume that thefortes and wc backgroundpractices ange managem

and analyseal capital’ (Ste, capturing . Knowledged in Chapter

he epistemolext of knowlll be present

knowledge iare seen as tents of the mnd the conteough changean asset. Buignored, or s

ge managems the systemhip, manageed here. Nonaka, Toyh is a processo use a concege assets as esearch, all treation’ refeHowever, th

practical knowtwork theorywledge man

ritical compata‐data; socige application’, and so fortaken. That search.

agers dislikets opportunie reason, indeir way is the

weak points –d, this sectiod systems. It

ment and the

e what knowtewart 1997;and utilisati

e and informr 2, Informat

logy of metaedge and, heed recurring

tself, with itcritical factomemory. (In ent.) s in the orgaut the supporseen as a sol

ent was use‐theoretical ement of peo

yama and Kos of knowledept ‘ba’ for athe inputs, othree elemenerring to the his approach

wledge, infoy, tacit and eagement and

arison; al, cultural on for orth) are is very true

or avoid ties and ividual e only right – will be on presents t will give a e

ledge is. ; Van Buren on, without ation ion and

‐knowledge ence, its gly, while

s roots in ors. This will many

anisational rting role of ution to

d in this approach ople). The

onno have dge creation a shared outputs and nts have to classical mainly

ormation explicit d its origins

Systems

De1.2

have facis not a sknowled

not beenmanagemphysical

managemresearchand not

viewpoinlevels ca

Sveiby a

systems included

Wiig. In following

1 The cruchttp://ww

between inf

efinitions of

Approced many prosimple underdge. In the fo

One pn well definement, but knstorage of in

Knowment of infoh will not prowritten dow

Therents are technan be contras

o Mthcogrha

o Medarsksk

nd many oth

o Ino O

or

In thisand other m

d as it appearOne o

his Managemg still used d

‘Knowwith d

1.

cial terms areww.soberit.h

formation an

f knowledge

oaches to knooblems. Partrtaking. It shollowing sectproblem in kned. Existing knowledge remnformation aledge managormation. Theovide yet anown as isolated are several nology and asted. Accord

Managemenheir educationstruction roupware eandled in in

Managemenducation in re primarilykills or behakills, know‐h

her scientists

ndividual PeOrganisationrganisation

s research thmemory medrs in adminisof the most rment of Knowdefinitions:

wledge manadifferent hor

. Business Porganisati

e defined in thhut.fi/wise/. A

nd knowledg

managemen

owledge ma of the diffichould always tion the concnowledge maknowledge mmains unclarand knowledgement has aere are manyother one. Thd sentences wschools or tradministratioing to Karl E

nt of Informaion in compof informat

etc. To themnformation snt of People.philosophy,y involved inaviour. To thhow etc., th

s divide know

erspective. Tnal Perspect.

he new emphia. The perspstrative systeespected anwledge – Pos

gement in orizons and pu

Perspective –ion must inv

he questionnaAlso, modern d

ge

12

nt

nagement aulty stems frbe recogniscepts relevananagement i

management rified. Circuladge (i.e. memalways beeny poor definhe definitionwithout a coracks of knowon. Two track. Sveiby (200

ation. Reseauter or infotion manag

m knowledgesystems. . Researchey, psychologn assessing, hem knowlehat is consta

wledge mana

The focus intive. The foc

hasis is on mpective is orgems or commd earliest ressition Statem

rganisationsurposes:

– focusing onest in or exp

ires. Case III tdata dictionar

nd the use orom the veryed that therent to this resis that so fardefinitions gar referencemory) is forgo divided intoitions of knon in this reseaontext.1 wledge manaks of thought00), knowled

archers andormation scigement syste is objects

ers and pracgy, sociologychanging a

edge is procantly changi

agement into

research acus in resea

anagement ganisational,munication asearchers ofment Februa

s must be con

n why, whereploit knowled

erminology wries, meta‐libr

of knowledgey nature of ke are many tearch are clar knowledge give differens are lackingotten. o managemeowledge manarch will be i

agement. Twt on the subjdge managem

d practitioneience. They tems, AI, reethat can be

ctitioners in y, or businesand improvicesses, a coming.

o two levels:

nd practicearch and pra

of knowledg, but the mand collaboraf knowledge ry 6, 1996, h

nsidered fro

e, and to whdge. Strategie

was published raries and e‐li

e managemenowledge. Ittypes of orgaarified. and informat synonyms fg in content.

ent of peoplenagement, animplicitly for

wo prominenject and twoment can be

ers in this fiare involveengineeringe identified a

this field hass manageming human implex set of

is on the inactice is on t

ge with informnagement ofation systemmanagemenhe already pr

m three pers

hat extent thes, products

in net page braries were t

ent systems ts definition anisational

ation have for Also, the

e versus nd this rmulated,

nt o different studied as:

ield have ed in g, and

ave their ment. They individual f dynamic

ndividual. the

mation f people is s. nt is Karl. M. rovided the

spectives

e and

traced.


13

services, alliances, acquisitions, or divestments should be considered from knowledge‐related points of view.

2. Management Perspective – focusing on determining, organizing, directing, facilitating, and monitoring knowledge‐related practices and activities required to achieve the desired business strategies and objectives.

3. Hands‐On Operational Perspective – focusing on applying the expertise to conduct explicit knowledge‐related work and tasks.’

He has defined the concept as follows: ‘Knowledge management is the active, explicit and systematic evaluation, facilitation, creation, direction and monitoring of all knowledge‐related activities and conditions within the enterprise’ (Wiig 14.5.1998), or ‘KM is the systematic, explicit, and deliberate building, renewal, and application of knowledge to maximize an enterprise’s knowledge‐related effectiveness and returns from its knowledge assets.’

The previous definitions are circular. So are most of the definitions of knowledge management. However, Karl M. Wiig’s perspectives begin with ‘business’. The business perspective is not criticised enough. People working in a business are not business objects or knowledge assets. Their knowledge is not a subject of commerce, and it is not even totally manageable, as claimed in the second perspective.

According to James Bair, Gartner Group, the aim of KM is an integrated approach, with which the information properties of an organisation are identified, captured, searched, divided and evaluated. Information property includes databases, functional principles, practices and competencies, but also the unobserved, tacit expertise and knowledge of employees. (Bair 21.11.1997: 21.) Why is information property knowledge?

The definitions of KM seem to be quite superficial. Beckman (1999) has collected a

number of definitions, such as:

‘KM is the process of capturing a company’s collective expertise wherever it resides – in databases, on paper, or in peoples head – and distributing it to wherever it can help produce the biggest payoff.’ — Hibbard 1997

The right knowledge, right people and right time pop up in various definitions:

‘KM is getting the right knowledge to the right people at the right time so they can make the best decisions.’ — Petrash 1996;

‘KM is a conscious strategy of getting the right knowledge to the right people at the right time and helping people share and put information into action in ways that strive to improve organizational performance.’ — O’Dell and Jackson 1998

‘KM is the explicit control and management of knowledge within an organisation aimed at achieving the company’s objectives.’ — van der Spek 1997

‘KM is the formalization of, and access to, experience, knowledge, and expertise that create new capabilities, enable superior performance, encourage innovation and enhance customer value.’ — Beckman 1997

Seventy three per cent of 260 European corporations voted for the business definition of KM as:

‘a collection of processes that govern the creation, dissemination and utilisation of knowledge to fulfil organisational objectives’ (Murray and Myers 1997).


14

Roelof Puit Beijerse’s article ‘Questions in knowledge management: defining and conceptualising a phenomenon’, published in 1999 in the Journal of Knowledge Management (Beijerse, 1999), starts with a look at previous definitions of knowledge. He notes that human beings have a need for knowledge (to structure the world around them, create meaning), and that knowledge has to do with information (Den Hertog and Huizenga 1997; Weggeman 1997; in Ibid.); he reviews Polanyi’s distinction between tacit and explicit knowledge, as well as Nonaka’s and Takeuchi’s knowledge creation processes (socialisation, externalisation, combination, internalisation – discussed further in this chapter). Finally, he ends up with a definition of knowledge as follows:

‘Knowledge is seen here as information; the capability to interpret data and information through a process of giving meaning to these data and information; and an attitude aimed at wanting to do so’ (Beijerse 1999, p. 100).

Beijerse refers to Weggeman’s definition of knowledge management, but wants to emphasise the importance of tacit knowledge; this results in the following definition of knowledge management:

‘Knowledge management is achieving organisational goals through the strategy‐driven motivation and facilitation of (knowledge‐) workers to develop, enhance and use their capability to interpret data and information (by using available sources of information, experience, skills, culture, character, personality, feelings, etc.) through a process of giving meaning to these data and information.’

The problem with these definitions is that all of them are built on, or relate to, typical ‘representationist’ ideas. This research also notes that ‘business’ is too narrow an approach to the wide scope of knowledge – even when it is constricted to organisational knowledge.

One missing link of the above definitions is that they see knowledge as an independent entity. Data and information (in current information science) are independent entities, but knowledge always remains in a memory. The memory can be short term or long term, individual or social, but the emphasis here is on the meaning of extended organisational memory.

The following wider – not only business‐oriented – definition addresses the problems best in terms of independence. It is partially adapted from the definition by the European Knowledge Management Forum which cooperated with the writer during the research of case III.

‘Knowledge management refers to an evolved or planned process through which organisations create and use institutional and corporate collective data, information and knowledge from organisational memory. Knowledge management deals with organising and controlling the operational processes in the knowledge value chain in an efficient and effective way. It focuses on facilitating and managing knowledge and information related activities concerning all information species. It is systematic, explicit and deliberate building, renewal and application of data, information and knowledge to maximise an enterprise’s knowledge‐related effectiveness and returns from its knowledge assets. ‘

Even though this definition comes closest, it also is circular. All the definitions analyse management, but knowledge is treated as a recurring concept of information, expertise or other synonyms.

What can we conclude from the previous definitions? The subject demands research. We do not need to constrict ourselves to one definition; therefore, this research will not redefine knowledge management, but will let the reader construct his or her own understanding of the subject.

Systems

Ro1.3

Greeks. Tresearcheven on further tstarting

started i(March aManageAdminist

PhilosopChapter the buildworkers’culture, focused enterpricodified the receintensivedefinitio

as followresource1990s it sufficienbusinesscontent informatthe signi

combinabut a pravarious sreact an1993), to(Bettis, BA compemanagemwaning omanagema scepticintroduc

Recentlytowards based vicapabilitcapabilit

between inf

oots of know

HumaThere has beh does not cothe first atteto Cuneiformpoint.

Duringn sociology aand Simon 1ment is an atration, publ

Variouphy has given2, Informatiding of know’ cognition, wenvironmenmainly on thse, while ecoknowledge: nt history ofe work and itons and fuzzy

If we tws: In the 197e. In the 1980was realised

nt. Emphasis s processes (managemention manageificance of in

Knowation of new actical, subststrategic prod cope with otal quality mBradley and etitive advanment was deof re‐engineement is a necal point of vction will sho

Stratey, strategic memployees ew of the orties as the prty are differe

formation an

wledge mana

an knowledgeeen even befoncentrate oempts of knom language in

g the previouand econom958), and phrticle by Niclished in the us disciplinesn various defon and know

wledge‐relatewhile social st and individhe organisationomics has The group ef informationts automatey philosophietry to crystal70s informat0s the strated that informwas transferHelokunnas nt and social ement with knformation syledge managand familiartantive respoogrammes fopressure. ThmanagementHamel 1991ntage was peeveloped by ering can be w name for view – every ow what is reegic managemmanagement(Barney 199rganisation. Trincipal sourcent entities (

nd knowledg

agement

e has been sfore a divisioon Tzu or Conowledge captn 3000 BC. Tw

us century remics (Hayek 1hilosophy (Poholas L. HenPublic Admis have influefinitions of inwledge specied capabilitiescience has bduals’ interacional level: omostly deteelicitation men systems, ard routines. (es circulatingllize the histotion (and its egic importanmation managrred to the m26.1.2001).networks. T

knowledge mystems for kgement, like r ideas. Knowonse to real or managemehe techniquet (Edwards D) and empowervasive in alconsultants generalisedearlier informso‐called neequired fromment has evo theories hav1; Grant 199This view recce of achieviLee, Kim and

ge

15

studied in phon between Enfucius from turing, storawentieth cen

evival of the 945, Arrow 1olanyi 1966).ry: Knowledginistration Reenced knowlenformation aes. Closest tes. Cognitive been interesctions and moptimising opermined prioethod was alrtificial intellKakabadse eg in the literaory of knowlgoal‐orientence of informgement as thmanagementThe first decThe forthcommanagement nowledge many system

wledge manasocial and ecent were deves – such as pDeming 1993werment (Lipl of these. (Bto replace d to every newmation and dew approach m the state ofolved along wve been shift91; Lank 1997cognises the ing and sustad Yu 2001.) O

ilosophy andEastern and China, or th

age and retrientury Weste

knowledge d1962, Marsh. The first apge Managemeview (Henryedge managand knowledo philosophyscience has ted in the or

motivation. Mperations anrities. Knowllso used in cigence has bet al. 2003.) Tature and in edge managed managemmation manahe only compt of intellectucade of 21st cming decade, in context o

management.that is valueagement is nconomic trenveloped to eprocess re‐en3), downsizinpnack & StamBarnes 2002.eclining revew movemendata manageis either oldf the art for awith a knowting their foc7). This apprimportance aining a comOrganisation

d epistemoloWestern phieir contempeval, which crn philosoph

debate (releall 1965), orpearance obment: A New y 1974). ement theorge, as will bey has been ininvestigatedrganisationalManagement d integratingledge enginease III in thisbeen used toThis mixtureorganisationgement in onent) was recagement waspetitive comual and sociacentury concand this res

of organisatio. ed, is both olot just the cnds (Prusak 2nhance an ongineering (Hg (Trimmer 1mps 1997) – ) Argumentenues from, nt. Sceptics mement method or wrong (Pan improvedledge‐centricus from orgroach is oftenof organisatpetitive advaal resources

ogy since theilosophers. Tporaries fromcan be tracedhy is more ac

vant for thisrganisationalbserved of KnConcern for

ry and praxise discussed information sd individual kl level: peopscience has g them withieering has els research. In tackle know has createdns. ne paragraphcognised as as emphasised

mpetence wasal capital andcentrated moearch combions and socie

d and new. Ionsultant’s i2001). In therganisation’sHammer and1993), outsooften appeas that knowlfor examplemay say that ods, but simPrusak 2001.d practice in tc view of theganisational cn called a restional resourantage. Resoare the mai

e ancient This m India, nor d even ccurate a

research) theory nowledge r Public

s. n Part I science, and knowledge le, their also n an icited and n regard to wledge‐d many KM

h, it could bea critical d. In the s not d the core ore on ine ety to clarify

t is a nvention e past, s ability to d Champy ourcing red limited. edge , the knowledge ilarly – from) This the future.e firm. capability source‐rces and ources and n source of


16

capabilities, while capabilities are the principal source of competitive advantage (Barnes 2002.) Knowledge‐centric is a wider term than resource‐based.

The resource‐based perspective considers information and knowledge as critical competitive differentiators. Organisational knowledge is acknowledged as the most valuable organisational asset (Spender 1996). The capability to manage knowledge strategically is a significant source of competitive advantage (Grant 1991). In business, initiatives are based on the use of modern information technologies for managing organisation‐wide knowledge resources. This research prefers the knowledge‐centric term, because an analogy between information and knowledge resources is wrong, as shown in the case studies.

The general evolution of knowledge management began in the mid‐1980s, although organisation and management researcher Peter Drucker has been using terms such as knowledge worker and knowledge society from the sixties, as he writes in his book Post‐Capitalist Society (Drucker 1993). In the eighties, when Drucker proclaimed the age of the knowledge worker, he predicted that competition in business would be all about the management of knowledge resources. Even though Drucker did not write that systems could replace knowledge workers, ten years later KM was used in the first essays on commercial system technology.

The first conference devoted to knowledge management, with over 150 participants, was held in Boston in 1994. However, the discipline was not well defined. Many of the speakers tried to define organisational knowledge in order to differentiate it from data and information. Knowledge seemed to be a key residual. It explained internal productivity after everything else was accounted for – even perfectly managed information would not lead to improved productivity.

Various academic communities tried to develop theories of knowledge management. The acronym was also commercialised with different technologies. Once the term knowledge management caught on, search‐engine, portal, artificial intelligence (AI) etc. vendors started re‐branding their solutions as knowledge management. The vendors did KM a great disadvantage. On the academic side, the AI community, and especially the knowledge‐based system community, have traditionally been involved in the development of ontologies to support knowledge management. At the beginning of the millennium, the leading vendors of commercial knowledge management systems included Autonomy, Business Objects. Cognos, Hewlett Packard, Hummingbird and Invention Machine (Carnelley et al. 2001). Some companies have also tailored their KMS solutions. An estimated 80 per cent of global corporations had tailored KMS projects by the turn of the millennium (Lawton 2001).

The proliferation of the web browser between 1995 and 1997 simplified the development of knowledge management applications with a common (not necessarily standard) interface. Also, in the research world, the hypertext community that led to the Web community developed associative memory models that extend human cognition. The computer‐supported cooperative work (CSCW) community, born from the human‐computer interaction (HCI) community, is developing software support for communication, co‐operation and co‐ordination. More recently, safety problems have induced the requirement of traceability of decisions in organisations.

Various universities have Ph.D. programs in KM. Most governmental organisations and private companies have launched KM projects. Individually, each community and market solution contributes a piece to the knowledge‐management puzzle. Altogether, they may constitute an overall theory of knowledge management (Lawton 2000; Boy 2002).

It has been said that ‘knowledge sharing’ among employees, and with customers and business partners has a tremendous potential pay‐off (see e.g. Barnes 2002: 2). In 1998 Thomas Davenport wrote:

Lotus… devotes 25 per cent of the total performance evaluation of its customer support workers to knowledge sharing. Buckman Laboratories recognises its 100 top knowledge sharers with an annual conference at a resort. ABB evaluates managers based not only on the result of their decisions, but also on the knowledge and information applied in the decision‐making process. (Davenport 28.10.1998.)

But the context is crucial here: The benefits are coming from narrow areas, which include improved customer service, shorter delivery‐cycle times and increased collaboration within the

Systems

companyother knvalue of just one the work

be digitisubiquitoincreasinlow costorganisa

even thosees knothe reitea need. (needs tomeasuri

containeto the risknowledUnited Sgeneric ivivid epitopic dowould fiand knowterm ‘co

managemhard to dbook.

Tr1.4

over sevsystems differentdifferentand diffetheir pri

between inf

y or with busnowledge. Evit will be adjknowledge kers with the

Basedsed, codifiedous and transngly have acct. One questiational know

The oough this resowledge as aerated use an(Gibbons et o be accountng it (Sveiby

Not alers in the sysse, fall, and rdge managemStates – fromissues, informsodes that lawnplayed thnally becomwledge couldontent.’

Propoment paradidevelop, and

raditional KM

Informveral decades(IS) literaturt research stt definitions erent methomary results

1. Thco

2. Thec

3. Thfo

formation an

siness partneven though injusted to infmanagemene thought of d on most ind or easily dissparent inforcess to almoon should beledge and skrganisationasearch claims commodity nd the reconal. 2000). Foed for, and a1997; Swanll knowledgestems can bereturn of KMment, a predm technologicmation and kater can be she role of tece a technicad be defined

ortionally, cogms and cond contributed

M systems a

mation techns. This can alre during thetreams. Bankof the manadological pas impact man

he decision somputers in

he value of inconomic ana

he human‐coor effective sy

nd knowledg

ers. It has alnformation clation. The ant illusion, whhappily worterpretationstributed. Burmation systost all the infoe raised herekills rise as wal level of knos that memotraded. In ‘k

nfiguration ofor the cognitia number of and Newell e needs to bee managed. TM. In many reictable reasocal viewpointknowledge (fstored and rechnology, thel issue. For td, and why an

mpared withncepts are ved by various

nd their pre

nologies for mlso be seen fe past decadeker & Kauffmagerial probleradigms usenagerial prac

support and decision sup

nformation ralysis of infor

omputer systystems desig

ge

17

lso been claican be sharebsolute benhich is sold trking in conjuns on informaut the real coems. Access ormation thee: Does the vwhen the valuowledge canory is a more knowledge ef knowledgeive paradigmefforts are b2000). The Se managed, aThe emergenespects, conton which hast. One modefor example,etrieved for e definitions he moment n important

h the long hiery recent endisciplines, a

decessors

managerial afrom the litees reflects th

man (2004) idems that relad to study thctice:

design scienpport, contro

research strermation as a

tems design gn.

med that sod (or ‘reconsefits of knowo managers unction with ation speciesore of knowlto informatey might neevalue of cognue of informa also be seenapplicable mconomy’, ‘kn with other f

m, knowledgebeing made tSECI process as the cognitnce of contentent manages to do with ern kind of co, from blogs)purpose. Whof knowledgit is importapart of them

story of philndeavours. Tas will be pre

and professiorature. The dhe inception,dentify five reate to IS, thehem, and the

ce research ol, and mana

eam reflects commodity

research str

me knowledstructed’) wiwledge sharinwith the idecolleagues. s, there is knedge managion expands ed at any timnitive and soation deprecn as conscioumetaphor. Thnowledge indforms to solve is an asset:to develop pis based on tive paradigmnt managemement is nowthe way KM ontent mana) through pohile the first ge ensured tnt to undersm would even

osophy, all kThey have beesented in va

onal workersdevelopmen, growth, andesearch streae alternate the levels of th

stream studgerial decisio

relationshipin the mana

ream empha

dge can be trithout losingng in the woea of profitab

nowledge thagement comeso that peop

me and in anyociotechnicalciates? usness or cohe cognitive dustries’ addve a problem it is somethrocedures fothe cognitivm claims. Homent is directw standing in was introduagement is est‐processinimportant bhat its manastand how inntually be tie

knowledge aecome very imarious chapt

s have alreadt of informatd maturationams that incheoretical pee organisatio

ies the applion‐making.

s establishedgement of th

sises the cog

aded for g it, the rkplace is bility and to

at can not es from ple y place at

gnition – paradigm d value by m or to meet ing that or e paradigm.wever, tly related for

uced in licitation of g of these ooks on the agement formation ed to the

nd content mportant, ers of this

dy evolved tion n of several orporate erspectives on at which

cation of

d based on he firm.

gnitive basis


18

4. The IS organisation and strategy research stream focuses the level of analysis on the locus of value of the IS investment instead of on the perceptions of a system or its user.

5. The economics of information systems and technology research stream emphasises the application of theoretical perspectives and methods from analytical and empirical economics to managerial problems involving IS and information technologies (IT).

Based on a discussion of these streams, Banker & Kauffman evaluate the IS literature’s core contributions to theoretical and managerial knowledge, and make some predictions about the road that lies ahead for IS researchers.

From these tracks there can be distinguished a road towards information systems, which leads to current KM systems. Information systems before knowledge management were developed under different disciplines. Information technologies for managerial and professional workers have been evolving for several decades. Their origins are in analytical applications:

• Management Information Systems (MIS) • Decision‐Support Systems (DSS) • Executive Information systems (EIS) • Information Management Systems • Artificial Intelligence • Semantic Network • Collaboration (Groupware…) • (Re)engineering Management information systems (MIS) were the first technologies to disseminate

organisation‐wide information from vast amounts of data for management purposes. Management systems were generally referred to as executive information systems (EIS). An EIS contains a portfolio of tools such as drill‐down access to databases, news source alerts and other information – all aimed at supporting managerial decision‐making.

A knowledge management system, however, does not have to be a computer system. It can be a process of finding, selecting, organising, distilling and presenting information in a way that improves comprehension in a specific area of interest, and acquiring, storing and utilizing knowledge for such things as problem solving, strategic planning, decision making and dynamic learning.

In this thesis there are various examples from knowledge management systems: Organisational Memory Systems; Documentation Management; Data Warehousing, Data Marts, Data Mining and Archiving; and Web‐enabled Information Services for Engineering. The last‐mentioned engineering services demonstrate examples from two different industries: thus they are the same kind of systems from two different viewpoints. The names of the companies involved in this study have been omitted, and they are referred to as ‘focal 2company’ or ‘partner companies’. There are references to the following subjects:

• Individual • Group or ‘team’ • Case organisation • Focal company • Extended organisation

2 All reported case studies are concerning an anonymous extended enterprise, referred to as the focal company.


19

In parallel with these official units, there are some results from other communities. Excerpts from the research material (systems and somewhat interviews) have been added in order to illustrate the findings.

Peer to peer vs. collaboration is a crucial distinction when dealing with extended organisational knowledge. Therefore the collaboration systems, groupware, and computer supported cooperative work are basically examined also in this research. Groupware, databases and enterprise web‐based systems are the basis for the cognitive paradigm of knowledge management. They are presented here mainly as the information species based on processing.

In the past, knowledge management was aimed at a single group – managers. More recently, however, knowledge management systems are increasingly designed for entire organisations and even shared between organisations. If executives need access to information and knowledge, their employees are also likely to have an interest in and need for that information. In addition, knowledge management technology is ideally suited for non‐management groups – such as customer support, where customer service requests and their solutions can be codified and entered into a database available to all customer service representatives.

In practice, as organisations store an increasing amount of information and knowledge in data and knowledge warehouses, and in data and knowledge bases, they are attempting to manage that knowledge in more efficient ways. Historically organisational knowledge has been stored in paper‐based archives and implicitly in people's minds. Unfortunately, paper has limited accessibility and is difficult to update. And when people leave, most of the knowledge vanishes, so reuse is not always feasible. Thus, organisations have moved to data and knowledge bases to improve their ability to access, update and archive data and knowledge.

Knowledge and information resources vary among particular industries and applications, but they generally include manuals, letters, summaries of responses to clients, news, customer information, competitor intelligence and knowledge derived from work processes. A wide range of technologies are being used to implement knowledge management systems: e‐mail, databases and data warehouses; group support systems; browsers and search engines; intranets and Internet; expert and knowledge‐based systems; and intelligent agents.

There are various definitions for a system. Ludwig Von Bertalanffy’s general system theory and biology could be the original roots, but the current system definitions from engineering can be summarised as follows:

‘A set of interrelated components working together to accomplish a common purpose’ (Conference on System Integration, New York 12.3.2003). ‘System is a collection of parts, which interact with each other to function as a whole’ (Martin, James N. ‘How to Implement Systems Engineering – Principles & Practises’ – a lecture at FINSE seminar, Helsinki 10.3.2003).

‘A system is a set of interrelated components which interact with one another in an organised fashion toward a common purpose’ (NASA Systems Engineering Handbook).

Based on Russell Ackoff’s articles, a system is a whole that can not be divided into independent parts without losing its essential characteristics as a whole. It follows from this definition that a system’s essential defining properties are the product of the interactions of its parts, not the actions of the parts considered separately. Therefore, when a system is taken apart, or its parts are considered independently of each other, the system losses its essential properties. Furthermore, when the performance of each part taken separately is improved, the performance of the system as a whole may not be, and usually is not’. (Ackoff 2003; Sauser 2007.)


20

In this research the following definition comes closest:

System is a collection of related elements and other systems that exists in an environment and contributes to a common goal. The system guarantees compatibility between the elements that it consists of.

However, this research is not strict to one definition with this concept either. The designed systems themselves tell what they are. The methodology of systems design is crucial. One methodology for engineering (case III) is a good example from Europe. CommonKADS (formerly known as KADS, Knowledge Acquisition and Design Structuring) is a methodology to support structured knowledge in engineering. It is one standard for knowledge analysis and knowledge‐intensive system development.

CommonKADS makes it possible to spot the opportunities and bottlenecks in the process of how organisations develop, distribute and apply their knowledge resources. It also provides the methods to perform a detailed analysis of knowledge‐intensive tasks and processes. Finally, it supports the development of knowledge systems that support selected parts of the business process. It is a top‐down approach for modelling. (Breuker 1994; Schreiber 1999)

Knowledge management systems are, as the word already implies, the application of techniques and methods from the field of knowledge‐based systems. Technology has been around for quite some time, and the application of knowledge technology and infrastructure is well known. The same knowledge technology is used in learning tools. Information management should always support the learning organisation and its memory. The question of knowledge management is most critical for big multinational corporations and their collaborators, where learning is constant and systems used worldwide. Some of the systems included in this research are presented next. Traditional KM systems have had their failures and successes. Useful examples are referred to later, and some of the caveats are mentioned as well.

1.4.1 Data and knowledge warehouses, and knowledge discovery

As many companies, this research also begins from one of the first knowledge management tools: data warehouses. The data warehouse concept is defined in the Part II section 1.2, Concepts, but there are various definitions. For example, Bill Inmon, the creator of Enterprise Data Warehousing, has published a set of descriptions of information management that is named DW 2.0 ™. Do these kinds of version updates or trademarks actually contribute anything new? The features of DW 2.0 are comprehensiveness (read: structured and unstructured data, including references), time scale (from on‐line information to archive) together with lifecycle, meta‐data and its integration. Even knowledge warehouses were planned well before any trademarks.

Rather than the kind of quantitative data typical of data warehouses, knowledge warehouses are aimed more at qualitative data. Knowledge management systems generate knowledge from a wide range of databases, including Lotus Notes databases, data warehouses, work processes, news articles, external databases, Web pages (both internal and external), and people. Thus, knowledge warehouses are likely to be virtual warehouses where the knowledge is dispersed across a number of servers.

In some cases, a web browser can be used as an interface to a relational database. A frequently used corporate application is a human resource knowledge base about employee capabilities and skills. Employee information can include education, specialties, previous experience and other descriptors. However, the skill management system is simply meta‐information.

Historically, Lotus Notes has provided one of the primary tools for storing qualitative and document‐based information and for facilitating virtual groups. With the recent explosion of the Internet, however, low‐cost Web‐based solutions within Intranet environments have become the focus of knowledge management. Lotus has got many new weapons for the struggle. These tools are dealt with in the concrete examples.


21

Knowledge discovery means generating knowledge from data. It is a new and rapidly evolving discipline that uses tools from artificial intelligence, mathematics and statistics to tease knowledge out of data warehouses. Gregory Piatetsky‐Shapiro and William Frawley defined knowledge discovery first as ‘non‐trivial extraction of implicit, previously unknown, and potentially useful information from data’ (Piatetsky‐Shapiro & Frawley 1991). Knowledge discovery is a method that includes different tools and approaches to analyse both text and numeric data. Data mining is related more to knowledge discovery than to data warehouses.

1.4.2 Representing knowledge

Knowledge management systems represent knowledge in both human and machine‐readable forms. Human‐readable knowledge is typically accessed using browsers or intelligent search agents. But some knowledge is accessible for machine‐readable purposes, designed as an expert system's knowledge base to support decision‐making. Meanwhile, ontologies are generally endemic to knowledge management systems because they typically refer to taxonomies of the tasks that define the knowledge for systems.

Human‐readable knowledge

Human‐readable knowledge is represented using a wide range of approaches in knowledge management systems. In many situations, case‐specific information appears to provide the appropriate level of representation required for users to make the best use of knowledge. In practice, this kind of information could be presented as declarative knowledge (facts and assertions), text or rules, but also as pictures and sound. Organisational rules, for example, guide promulgated behaviour, and adaptations of these rules could potentially be used in a rule‐based knowledge management system.

If, on the other hand, information is highly filtered, then it is likely to be represented as a set of declarative statements. For example, a knowledge base on Global Best Practices lists five specific ‘Best Practices for Managing Information Resources,’ including ‘Develop and maintain an IT strategy that is integrated and aligned with the company's business goals’ (Arthur Andersen 19.11.1998). The knowledge is declarative and is independent from particular situation information, meaning that the database lists no particular cases. Although filtering ensures that knowledge is correct and consistent, developing declarative knowledge is ultimately a political process, typically removing context and controversy. As a result, filtered knowledge can be limited in its ability to provide as deep an insight as unfiltered information. Best practices technique can become complicated and problematic, as argued by Wareham and Gerrits (1999). Their analysis is mainly based on theories and theoretical constructs. Observations on practical examples during this empirical study indicate the same results.

Machine‐readable knowledge

Expert systems use their knowledge bases and user responses to guide the user to recommended solutions. Expert systems can be an integral part of a knowledge management system.

Although some knowledge management systems contain such artificial‐intelligence‐based parts, they use artificial intelligence primarily in the form of intelligent agents to search for human‐readable knowledge. Additional research is needed to expand the use of artificial intelligence and knowledge‐based systems in knowledge management. It is necessary to know what forms of knowledge representation appear to work best for particular types of knowledge and how artificial intelligence can be further integrated into knowledge management systems.


22

Ontologies

Ontology is an explicit specification of a conceptualisation. In enterprise knowledge management systems, ontology specifications can refer to taxonomies of the tasks that define the knowledge for the system. Ontologies define the shared vocabulary used in the knowledge management system to facilitate communication, search, storage and representation. Development and maintenance of an enterprise‐wide ontology requires continual effort to evolve the ontology over time.

Ontologies are particularly important in ensuring that best‐practices databases are able to communicate to the user the broadest range of practices and activities and allow the user to recognise when the best practice would fit in their organisation. In addition, since enterprises are often involved in multiple industries, multiple ontologies may be required as part of the knowledge management system.

Out of necessity, virtually all enterprises with a knowledge management system have developed their own ontology. Because these firms have made this investment, ontology construction appears at this point to offer competitive advantages. In this research is found out, whether huge divisions of the focal company, or industries in general, are to form coalitions or subscribe to central services for interests in an ontology shared across multiple organisations in order to cut development costs and to speed system development.

Ontologies are related to modelling. Traditionally three ontologies which span the basic dimensions of information modelling are 1) Enterprise ontology, which gives the context, and 2) Domain ontology, which gives the content to 3) Information ontology (Abecker et al. 1998, p. 44). The latter requires philosophical ontology, described later along with information species.

Other knowledge description attributes

In addition to ontological information, additional descriptive attributes of knowledge can prove critical to its use and maintenance. Contributor, organisational and status information are all viable descriptive attributes. Virtually all knowledge bases capture contact or contributor information, including contact or contributor names, date of contribution, a person's role in generating the knowledge and so on.

Many knowledge bases also include organisational information that can define the department or division in which the project was built or from which knowledge was gathered. Status information about knowledge is also a typical kind of descriptive attribute. This kind of status information can include whether an element of a project is planned, currently being implemented or has been implemented. It can record whether the information is externally available or for internal purposes only.

1.4.3 Knowledge filtering

Quality and importance of the knowledge varies, depending on a number of factors, such as who is providing the knowledge to the system. A lack of information filtering typically leads to multiple and sometimes conflicting responses, such as in informal communication systems, where messages sent to the forum are captured in topic threads consisting of the initial message and subsequent responses.

Because knowledge quality and importance varies from source to source, systems often resort to knowledge filtering to ensure complete and correct knowledge. In practice this means a combination of human and computerised knowledge, when an editor or a team filters messages by their usefulness and relevance (Davenport 01.02.1998).

All filtering is not done by humans. Perhaps the most visible and frequent use of computer‐based filters is the message filtering that categorises and prioritises e‐mail messages. A number of products also help monitor qualitative databases. For example, grapeVine monitors multiple Lotus Notes databases based on a personal interest profile according to individual, group or organisational needs.


23

Cascading as it is accomplished in collaborative systems has some limitations. Collaborative systems have individuals rank the importance of information coming into the company. Users might categorise certain information as ‘very important’, ‘important’ and so on. Other individuals in the enterprise then decide on what level the information must be labelled before it is delivered to them. In the case of a busy manager, information might need to be ‘very important’.

However, a limitation of using this approach is that some information ranked as ‘important’ might turn out to be ‘very important’. The manager, then, might not always see information that is necessary or very important. In such settings, there would be a tendency to rate marginal items higher than other. Meanwhile, managers might set their own level at ‘important’ in order to assure that all ultimately ‘very important’ announcements would be received. Ultimately, this leads not only to importance inflation, but also to a deluge of information that such a system is designed to combat.

1.4.4 Searching for knowledge

Knowledge bases can become quite large and searching them efficiently becomes an extremely critical function. The most dominant search techniques include search engines, intelligent agents and visualisation models.

Search engines

A wide range of well‐known Internet search engines have been investigated during this research3. The classification of engines was as follows:

1. General engines: Google, AltaVista, Excite, HotBot, Infoseek, Kolumbus, LookSmart, Lycos, Northern Light, PlanetSearch. Search.Com, Open Text Index, WebCrawler, Yahoo!;

2. Business search engines: BigBook, YellowPages, Guide; 3. MetaSearch engines: Ask Jeeves, MetaCrawler, Mamma, MetaSearch, ProFusion,

SavvySearch; 4. People search tools: Four11, Switchboard, InfoSpace, Bigfoot, WhoWhere, WorldPages; 5. Usenet search engines: Deja News, HotBot, Reference.COM.

Search engines have been used to guide users to information on the Internet. These and other search engines can be adapted to intranet environments for knowledge management. In addition, a number of firms have developed alternative approaches to the conventional search engines. For example, ‘central repository of interfaces’ has been used for quite a while. These 'knowledge maps' are a link to knowledge (Bernstein 1997); users can select a map and use it to navigate directly to knowledge stored in multiple databases without needing to know which database to access.

It should be noted that search engines are one current instance of information discovery tools. It would be interesting to find out, why these tools are so difficult to develop. Different types of tools during the decade in this research have been investigated, including: <isindex>, WAIS (Z39.50 protocol), TREC, Commercial Search (Verity, PLS, etc.), Yahoo!, AltaVista, Harvest etc., PointCast, Google, Web Portals and Intranet Portals. The list is a reminder of rapid change.

During this research, the discovery tools have changed from search engines toward portals and publishing channels. The tested searching tools covered also meta‐data search in the extended enterprise (See Figure 2. Search based on meta‐data in external documents).

3 Searching was an immanent function in all cases throughout the research, therefore introduced here.


24

Figure 2. Search based on meta‐data in external documents

Separate search engines, however, are not the solution for human (constructivist)

knowledge management in the future. On the basis of this research, the integration of different tools in an extended enterprise, and the independence of content and systems are more predictable.

Intelligent agents and visualisation models

Intelligent agents can be used to connect people to knowledge available on the Internet or Intranets. For example, the IBM4 Corporation releases white papers about its Intelligent Agent Strategy. The first products have been developed since the previous decade (Krulwich & Burkey 1997). Intelligent agents, such as info finders, learn user interests from sets of classified messages or documents, recognizing that people will tend to classify only those examples that interest them. In addition, info finders use heuristics to gather additional insights into a user's interests. Based on message syntax info finders attempt to determine significant phrases that provide insight into user goals.

One of the dominant trends in the search for effective enterprise knowledge management is the use of visualisation models. They are positioned somewhere between presentation and search. Two different ways to represent knowledge space can be differentiated.

First, analysing content using meta‐information derived from source documents – be it structured information in databases and tagged documents such as news feeds, or unstructured information in office documents and Web pages. From unstructured documents a linguistic analysis can be performed and documents automatically tagged. This tagged information can be analysed and relationships between documents identified. The constructed result is a multidimensional information space using a mark‐up language. Data is downloaded to the client using any information streaming transport protocol to conserve resources.

There are also applications for visualising large hierarchies. Their technology offers several visualisation formats, each of which exploit focus and context techniques that foreground objects of interest while preserving the overall structure of even very large data sets.

4 High‐tech organisations including Xerox Corporation, Hewlett‐Packard Company, and IBM were early explorers of knowledge practices (with varied success), trying to apply their undoubted technological capabilities to managing knowledge. All of these collaborating organisations have been constantly followed up during this research.


25

One such tool, the hyperbolic browser display, has been developed over ten years. It exploits hyperbolic geometry to provide exponentially more information space for hierarchies that expand exponentially with depth. Thus, a hyperbolic browser can display 1000 nodes as opposed to the 100 or so nodes displayed in a conventional 2D browser (Lamping, Rao & Pirolli, 1995).

The user navigates the information space by clicking on a node or dragging the mouse over the hyperbolic plane. Current implementations map Web hierarchies identified by URLs, thus forgoing a strong semantic structure, but it is conceivable that the browser could incorporate more semantic information using technologies such as linguistic natural‐language processing tools.

1.4.5 Applications and their development tools

Nowadays the scope of technologies for KM can be seen in the following list, compiled during this research:

o Business Intelligence – Data Warehousing, Data Marts, Data Mining o Groupware and Collaboration Technologies o Search Engines o Document Management o Knowledge Resource Directories o Messaging / e‐mail o Content Management o Web Calendars / Reminders o Portals o Web Casting o Intelligent agents o Customer Relationship Management o Decision Support Systems o Workflow and Tracking o Web / multimedia based training / e‐learning o Social networking utilities o Wikis o Tags o Widgets o Blogs

In this research, existing systems were first investigated (case I), then two of them were

compared in case II, and finally in the third case, the customised application was researched. New system development was done only in case III, but all the off‐the‐shelf systems are evolving as well.

One widely referred to list on the conflict between knowledge and applications is Mc Dermott’s (1999):

o Knowledge is a human act o Knowledge is the residue of thinking o Knowledge is created in the present moment o Knowledge belongs to communities o Knowledge circulates through communities in many ways o New knowledge is created at the boundaries of old

Systems

can suppcover apand havelanguagelearnt thprogramtricky pu

based onof the syin all thr

towards managemmoving fconversathrough possibly

W1.5

categoriOne resuresearch

informatthe systescience aWisdom

managem‘know‐wthe Infodoes notphilosop

knowled(1989), abe classi

lighter siof this re

between inf

Theseport knowledpplications ane varied, for es. Even thouhat cases shommers only. Ouzzles. Know

Some n Intranet (Loystem providree cases for

Internvirtual officement, extenfrom asynchation and obgroupware tin distributio

What are data

In thissed for centult is a picturh and a strate

Later tion and knoems. The claand knowled (DIKW). Tha

Milan ment contexwhy’ (Zeleny1ormation Scit mention daphy hundreds

Even tdge, we musta systems theified into five

1. D2. In

‘w3. Kn4. U5. W The co

ide (first busesearch.

formation an

e issues are sdge managemnd their deveexample, frough the resuould also inclOn the otherledge and apof the knowotus Notes gder (and its ocomparison

net‐based coe. Knowledgded search, ronous to sybject sharing.towards knoon.

a?

s research twuries, and dare of these cegy based onsections preowledge. Thisssification ofdge managemat hierarchy Zeleny wrotxt, as an anal1985). Howeence contexata – and thes of years agthough this rt take a look eorist and pre categories:

ata: symbonformation:what’, ‘whernowledge: anderstandi

Wisdom: eva

oncepts of usiness introd

nd knowledg

omewhat diment; the prelopment. Dom XML, thrults are preseude a normar hand, graphpplications hwledge managroupware, downers) are r. ntent could ge managememessaging, wynchronous w. Collaboratioowledge. The

wo categorisaata and inforategories ban that. sent practicas section exaf informationment both reis used evente the first arlogy to the qver, an articlxt refers to a e informationgo, when cultresearch focat another mrofessor of o:

ls data that are’, and ‘whapplication ng: apprecialuated und

nderstandinuctions) of k

ge

26

scussed in thresumption sevelopmentough UML toented in Partal scenario. Uhic designersave many cogement systdomino and orelevant back

be seen in thent framewoweb applicatwith collaboron tools have value of kno

ations meet:rmation systeased on obse

al knowledgeamines the sun and knowleefer to the h in ITIL, in thrticle which questions ‘knle about the poem by T.Sn, knowledgeture was quiuses more omuch referenorganisationa

are processehen’ questioof data andation of ‘whderstanding

g and, especknowledge m

he results. Thshould not b tools and mowards objet III, it could bUser interfacs are designiommon factotems describother web tokground info

he history froorks include tions and meration keywoe developedowledge is in

: Knowledge ems, which hervations der

e, as well as ubject of theedge began wierarchy of Dhe service traincluded thenow‐nothing’InformationS. Eliot (Clevee and wisdomte different.n the grey anced hierarcal change, th

ed to be useons d informatiohy’ .

cially, wisdommanagement

he main quee made that

methods havect‐oriented pbe notified, tces should neng user interors. ed in this stuool systems)ormation. Th

om back‐offiproducts likeeeting tools. ords such as d from messancreasing in

and informahave been carived from gr

philosophicae research –without the Data, Informaansition conte DIKW hiera’, ‘know‐whan‐Knowledgeeland 1982).m concepts o

rea betweenchy. Accordinhe content of

eful; provid

on; answers

m are mainlyresearch, an

stion is how knowledge e been changprogramminthat from UMever be left trfaces that a

udy are tech. The produce products a

ice, through e document Communicaawareness, aging to the wcreation, dis

ation, which ategorised forounded the

al and scientdata and infdata layer. Iation, Knowltext. rchy in the Kat’, ‘know‐hoe‐Wisdom hie The poem ooriginate fro

n informationng to Russell f the human

des answers

s ‘how’ que

y discussed ond are outsid

systems does not ging rapidly, g ML we to re more like

nically ct strategies are present

front‐office

tion is

web scovery and

have been or decades. ory

ific formation in nformation ledge and

Knowledge ow’ and erarchy in of course om

n and Ackoff mind can

s to ‘who’,

stions

on the de the scope


27

Ackoff says that the DIKW hierarchy can also have many dimensions. One dimension of Ackoff’s hierarchy is temporal. He states that while information ‘ages rapidly’, knowledge ‘has a longer life‐span’ and only understanding ‘has an aura of permanence’. It is wisdom that he considers to be ‘permanent’. This is an American individualistic point of view. It does not suit organisational or social knowledge or information. The situation is just the opposite for an organisation. Data and information remain, while individuals come and go.

A basic model for this research is a hierarchy describing only data, information and knowledge (The origins of the pyramid shape are not analysed, but it is described in many articles – together with wisdom and, sometimes, understanding). To meet their goals, organisations must manage information and knowledge resources effectively. Information and knowledge management focuses on different parts of the same value chain. (See Figure 3. Data, information and knowledge pyramid.) This research will show that all data, information and knowledge models should be interpreted in a new way or replaced with a more appropriate model.

Figure 3. Data, information and knowledge pyramid

The relationship between data, information and knowledge may be seen as a loop. Data

and information are only two opposite ends on a continuum. Concentrating our attention on certain aspects of knowledge makes it focal. This focal knowledge can, sometimes and partially, be articulated and furnished with words and be referred to as information. If the information becomes too de‐contextualised, i.e. too distant from the knowledge required to interpret it, it can be addressed as data. Since a piece of text itself is not sufficient to exhaustively describe the knowledge to which it refers, the reader's tacit knowledge must be compatible with that of the writer in order to interpret and fully comprehend the implications of the information. Hence, what one conceives as information another sees as data (Stenmark 2002). Adopting a social perspective on knowledge leads to different criteria, which are related to the level of stabilisation. Knowledge is widely stabilised among many people through institutional procedures (school, university, skills institutions). Data is significant to a small amount of specialists. Information implies a continuous process of evaluation, while interpretation is more or less stabilised, built in order to cross boundaries. The main result of this situation is clarified in Figure 46. Knowformation in individual and organisational triangles, which is included in the M3.EXE model, and is not simply a metaphor.

Systems

So1.6

social nesociotectheoretiwrote anthe first on knowresearch

Nowadapercentaknowled

privacy. researchemploye

informathumanis‘off‐the‐personadifferencwith theago.

predictioAfter emsociotec

which ofapproacinfluenceor organ

researchexaggeracovers sc

knowledinterpretmental mThe sociabout or

explorinstabilisatactors anpreferenallowed

technololiteratur

between inf

ociotechnica

The cuetworks. Thishnical view hcal approachn analysis of economic ex

wledge manah has not bee

Nurmys the informage of peopldge workers,

One syHow the useh, however, iees, whereas

Most tion technolostic side of inshelf’ informlised – no mce between te fact that inf

Nurmons were evemphasising inhnical resea

The inften treats tehes. It liberae through acnisation, by p

The ach has been usated when cocience from

The kndge itself is atation of knomodels, whically‐orientedrganisationa

The sog the world, tion processnd natural annce to social to be actant

Still mogical practice findings ar

formation an

al view and t

urrent evolus is not a newhas been dishes have alsothese approxpansion of igement – foen based on inen’s viewpmation and ee doing knowthe amountystem‐theore of ICT is cos different frs Nurminen cof the humaogy. This resnformation symation systematter if they the two decaformation sy

inen’s classifen fulfilled. Hnformation mrch is relevanformation syechnology inally assigns pctions. Conseperforming actor‐networksing it, with ompared witother researnowledge palso a speciesowledge. Actch are shaped constructivl knowledgeocially‐orientincluding infes of technicnd technical elements, wts or actors inmore preciselce by Edwardre adapted to

nd knowledg

he actor‐net

tion has prow direction, bscussed in liteo been preseoaches as earinformation cused only oa stable ‘golpoints were sespecially cowledge workt of which waretical issue tntrolled wasrom the probconcentratedanistic side ofsearch needsystems studms could be are based e.ades is commystems are ba

fication was However, infmanagementnt. ystems as a a way that sroperties, poequently, artctions that wk theory canmany interpth the actuarchers, like Maradigm useds of informattors are consd through invist theory (c. ted construcformation sycal and scienphenomena

whereas in thn networks ty for this resd W. Constano these categ

ge

28

twork theor

gressed frombut a constrierature for oented repeatrly as 1986 (Ntechnology. on the dualisden mean’ bstable, but hmmunicatiok has grown eas not forecatreated in bos, and still is,blem presend on citizens f knowledges to include cied in this retailored to in.g. on Softwamunication tased on an a

ahead of its formation tet, KM literatu

case can be superficially owers and eftefacts can awere neither not be passretations. Thl subject of tMichel Callond in this resetion presentasidered to ennterpretationcompared wi

ctivist theoryystems, the antific objects a. Strong soce actor‐netwthrough whicsearch, current II are usedgories. The t

y

m data, inforicted view ofover twenty ytedly. A FinniNurminen 19Ten years la

stic division. Sbetween hume was researn technologyenormously.asted. oth this reseaa dilemma. ted by Nurmand academ

e managemecurrent humaesearch is shandividual orgare as a Servechnology aarchitecture t

time. He alschnology haure went to t

observed thresembles tffects to techlso have uninr intended noed over becahe original cothis researchn or John Lawearch is basedation. Constrnact and conn and discouth cognitive)

y is divided inactor‐netwoas these resial constructwork theory ch technical ent interpretd as the basictheories used

mation and f informationyears. The hish researche986). The boater a similarSo far knowlman issues arching a set oy is accessibl. So the rese

arch and NuThe privacy

minen. We armic world. nt ignores, oan managemaring with Nuganisations avice architectnd mobility. that was not

so had foresis developed the leadersh

rough the ache analyses hnologies, wntended effeor anticipateause scienceoncept by Brh. The concepw. d on the conructivism is anstruct realitirse between) is more app

nto more detrk theory is msults from thtivism is critinatural or teor scientific tations from c categorisatd in this wor

knowledge tn or knowleduman and syer, Markku Nok was writtr kind of trenledge managnd informatiof IT professle to everyonarch subject

rminen’s stuproblem in tre dealing wi

or even despment issues. Turminen theand even be ture. HoweveThis researct predicted t

ght and som in a differenip track. Now

ctor‐networkof social sha

which can havects, as can aed by others.e and technoruno Latour mpt is too narr

nstructivist tha frameworkies based onn different mpropriate wh

tailed branchmost precisee building ofcised for givechnical elemobjects are sthe social lotion model. Ok are mainly

towards dge. The ystem‐Nurminen, ten during nd – based gement ion systems.ions. ne. Also the was only

udy is this th

ises The idea that

er, one ch deals wenty years

me nt direction. w

k theory, ping ve an an individual logy may be row when it

heory, while k for an their embers. hen talking

hes. When e. It studies f human ing special ments are stabilised. cus of Other based on

l

Systems

the sociamanagemmanagem

theory isinformataccuracybecome

strategicforms ofrepeatedstructurewhich enbenefit owide speexceptiostudy.

SE1.7

the orgamodels (capital (W1999) an1998). Tmodels ainterpretworldvie

Knowledclaim thacreationmanagem

between inf

al sciences ament literatument, docum

Figures referred totion systemsy in technolomore impor

Therec alliances, suf co‐operatiod resource aes, as well asnables peoplof inter‐orgaectrum of coon of the exte

ECI and ba

Cognitanisation kno(Carayannis1Wiig 1993; End the modehe most widare based onted differentew.

As in tdge Creationat most orga process, anment.

Bco

formation an

nd philosophure (see e.g. mentation an

e 4. Collabora when evalus. However, togical practicrtant than th are several ubcontractinon vary, somend informatis factors relale to communisational ne

ompetences, ended organ

tive paradigmowledge netw1999), the piEdvinsson anl based on codely known on positivistic tly in differe

this research’ (Nonaka etanisations lacd assert that

Business llaborators

nd knowledg

hy. HoweverBarnes 2002nd engineeri

Fig

ators shows ating how exthis researchce. Latour’s ce research mforms of netng relationshe common faion flows, mated to humaunicate, collaetworking is sharing of rinisation conc

m‐based moworks modelllars and fund Malone 19ognitive scieof these modphilosophy, nt sources b

h, in their artt al. 2001), Nck a general t knowledge

Cs

s

ge

29

r, their long r2). Case specng, but also

gure 4. Collabo

the actors axtended orgah takes a newconcepts conmaterial itseltworking bethips, networkactors in theutual expectan interactioborate, and the competisks, and infocept, the acto

odels of knowl based on onctions of kno997), the inteence, semiotiels is SECI mthe philosopecause SECI

ticle ‘SECI, Baonaka, Toyaunderstandimanagemen

Custome

Componenservice pro

Organisat

roots are leacific observatfrom extend

orators

nalysed in thanisational kw look at thencerning the f (see e.g. Latween compks and virtuae networks catations, and gon within thefinally innovitive advantaormational syor‐network t

wledge includrganisationaowledge maellectual capics and epist

model. This rephical basis iand ba do n

a and Leaderma and Koning of knowlent often redu

ers

nt and viders

ion

ding us also tions are moded organisat

his research.knowledge ca actor‐networesearch of tatour 1996). anies: these l organisatioan be found.governance e networks. Ovate efficientage produceynergies (Ebtheory was n

de state of kl knowledgenagement mital managememological pesearch claims too strict aot originate

rship: a Unifino focus on edge itself auces itself to

Standardassoc

to current kostly about dtional collab

The actor‐nan be managork theory bytechnology o

include joinons. Even tho. These factoor coordinatOne exampletly. The fundd by the utilers 1997). Wnot finally us

nowledge (Ee and meta‐kmodel of intement model pragmatism ms that the cand the concfrom the We

ied Model ofknowledge cnd the knowo mere inform

disationciates

nowledge ata oration.

network ged with y stressing often

t ventures, ough the ors include tion e is trust, amental isation of a

With the ed in this

Earl 1998), knowledge llectual (Van Buren (Snowdon ognitive epts are estern

f Dynamic creation andwledge mation


30

To understand the true nature of knowledge and the knowledge creating process, they advise us to consider tacit and explicit knowledge as complementary, and that both types are essential to knowledge creation.

Nonaka, Toyama and Konno define knowledge creation as a continuous, self‐transcending process where the boundaries of the old self are crossed and one enters into a new self by acquiring a new context. But they also stress that it is equally important to realise that knowledge is created through interactions between individuals and their environments, and that one should also speak of a transition between self and other.

Table 1. The SECI Process (Nonaka & Takeuchi 1995).

From/To Tacit Explicit

Tacit Socialisation

Creates sympathised knowledge through the sharing of experiences, and the development of mental models and technical skills. Language unnecessary. ⇒

Externalisation

Creates conceptual knowledge through knowledge articulation using language. Dialogue and collective reflection needed.

⇓

Explicit Internalisation ⇑

Created operational knowledge through learning by doing. Explicit knowledge like manuals and verbal stories helpful.

⇐ Combination

Creates systemic knowledge through the systemising of ideas. May involve many media, and can lead to new knowledge through adding, combining & categorising.

Nonaka and Takeuchi have presented the SECI model in their book Knowledge Creating

Company, and Nonaka et al. has represented it several times elsewhere. Nonaka, Toyama and Konno believe that organisations create knowledge by means of interactions between tacit and explicit knowledge. They have divided these interactions or conversions into four modes and call them the SECI process. The four modes of knowledge conversion are:

1) Socialisation. By this they mean a situation where individuals/groups share tacit

knowledge. As tacit knowledge in their eyes can only be shared through shared experience, the key is to construct arenas or possibilities for shared experience to take place. Apprenticeship is one common way of achieving this, but firms will also need to take advantage of the tacit knowledge embedded in customers and suppliers.

2) Externalisation. By this they mean a conversion process from tacit to explicit knowledge. A good example of this is concept creation in product development, and they suggest that the successful use of metaphor, analogy and models are key points for succeeding in this.

3) Combination is sharing of explicit knowledge. This can be supported by the creative use of computerised communication networks and large‐scale databases.

4) Internalisation is the process where knowledge is converted from explicit to tacit knowledge, often accomplished by ‘on‐the‐job‐training’ or ‘learning‐by‐doing’.

The knowledge spiral is a dynamic process of organisation‐wide knowledge creation.


31

Figure 5. The knowledge spiral (Nonaka & Takeuchi 1995: 72)

If we look at Nonaka & Takeuchi's later (1995) use of the pluralist explicit/tacit interaction, the outcome is still strictly managerial – for learning, growth and control in the organisational context:

Table 2. Managerial interpretation of knowledge (derived from Nonaka & Takeuchi 1995)

Individual Social

Explicit Conscious Objectified

Implicit Automatic Collective

Even though SECI is very popular and has been referenced in the focal company, this research claims that we need a more pluralistic epistemology for a knowledge‐based strategy of organisation.

The SECI model has been used in many further conceptualisations. According to Beijersee, the core of the knowledge management process is the set of instruments with which learning is being stimulated and knowledge is being managed. These instruments are classified according to the kind of interaction between tacit and explicit knowledge: i.e. socialisation, externalisation, combination and internalisation; and the four main aspects of the knowledge value chain: determining the knowledge gap (between necessary and available knowledge), developing and/or buying knowledge, sharing knowledge and evaluating (the use of) knowledge. The following table visualises this.

Systems

and consthe instrsee BeijeWhere is

knowledknowledparticipaNonaka,philosopdefined does not

Toyamo western needs tointeract taken int

that are moderat

that takithis reseeither phthem wi

individuaAbductivkind of brelationsorganisaorganisaorganisa

Th1.8

on informorganisa

between inf

Table 3

From sultants, queruments usederse 1999, pps the memor

Accordge creating dge‐creating ate’, and sug Toyama andpher Kitaro Nas ‘a shared t necessarily

Even tand Konno, view is base

o be replacedand, as a coto use althou

Nonakindispensabting factors o

It is paing a snapshoearch it will ahysical or virth living, dyn

This mal support fuve thinking isbasis of the nship betweenational memoation) has goations.

he missing li

In thismation systeational data,

formation an

3. Examples o

this model, estions regard in knowledp. 108 – 109ry? Memory ding to Nonaprocess. Knoprocess is neggest that bad Konno, ba Nishida and fcontext in w

y mean a phythough this rthat much red on the assd by creatingnsequence, tugh it has beka, Toyama able to creatinof the knowlaramount toot of these aalso be questrtual – that cnamic knowlmodel excelleunction, but s a key procenew knowledn explicit andory. Also, thet a sociologic

nks

s summary wems. The ma information

nd knowledg

of knowledge

Beijersee derding (1) the dge managem). This reseais the basis aaka, Toyamaowledge neeecessarily co, which rougis based on aurther develwhich knowleysical space aresearch devesearch on ksumption thag arenas whetheir shared een introducand Konno dg value for tedge creatino note the dyat any given ttioned whethan support tedge assets.ently demonrather demaess in SECI, adge and inford tacit knowe communitycal and histo

we can see win gap in then and knowle

ge

32

management

rives four castrategy of tment; and (4rch, howevealso for knowa and Konno,ds a contextontext‐specifghly means pa concept thoped by Shimedge is shareand that the velops an alteknowledge cat individualsere people wcontext caned several tidefine knowlehe organisatng process. ynamic naturtime is not eher organisathe SECI proc. strates that ands the supnd this inferrmation inteledge will bey of practice orical perspe

what is missine current statedge – as ma

t instruments

ategories of ithe organisa4) the outputer, questionswledge creat the SECI prot to be createfic in terms oplace, offers shat was originmizu. Ba in Ned, created akey concepternative moreation focus are the prim

with different evolve. Howmes. edge assets tion. Assets a

re of such asenough to cretions need tcess, and wh

knowledge mport of co‐oprence logical erpretation ine remodelled(discussed ictive, and is

ng from knowte of the art anagement o

(Beijersee 19

mportant qution; (2) orgat of the entirthe adaptabtion. ocess is just oed, and theyof who particsuch a contenally proposeNonaka, Toyaand utilized.’ t to understadel, we assuses too heavmary drivingt experienceswever, the ba

as organisatact as inputs

sets, and it iseate knowledo create roohether organ

managemenperative worand cognitivn this book. Hd. This researn case III secstill alive in

wledge manais that a genof organisatio

999, p. 105)

uestions for anisational mre process. (Fbility of such

one part of t state that ‘tcipate and hoext. Accordined by the Japama and KonThey stress

anding ba is ime, like Nonvily on individg force of cres and backgra concept ha

ion‐specific rs, outputs an

s important dge support oms – a type isations nee

t never can brk at various ve process wHowever, thrch will also ction 3.2, Extcontempora

agement andneral model fonal memory

managers matters; (3) For the list, a model:

the the ow they ng to panese nno’s text is that ba interaction. naka, duals. This eativity. This rounds can as not been

resources d

to realise systems. In of ba, d to provide

be an levels.

will be one e model an tended ary

d research for extendedy – is

d


33

missing. The smaller gaps are discussed in the following subsections. All of these are a result from constant literature surveys and field observations.

1.8.1 Do we know our content – what is the difference between data, information and knowledge?

Peter Drucker, a consultant already mentioned and one of the first people to write about the idea of the 'knowledge society' and the 'knowledge economy' (Drucker, 1969), disputes the notion that knowledge can be managed. At the Delphi Group's Collaborative Commerce Summit, Kotzer (2001) reports Drucker as follows:

…Drucker… scoffs at the notion of knowledge management. 'You can't manage knowledge,' he says. 'Knowledge is between two ears, and only between two ears.' To that extent, Drucker says it's really about what individual workers do with the knowledge they have. When employees leave a company, he says, their knowledge goes with them, no matter how much they've shared.

This is one extreme. In this sense, the subset of knowledge that can be managed with systems has not been discovered yet. The problem with McDermott’s and Drucker’s critique is that they see knowledge as what they want and only related to an individual. This research will examine knowledge more explicitly and yet organisationally. The species of information, and classification of data, information and knowledge will be investigated to determine whether there is a difference between knowledge derived from data management, information management or knowledge management system.

1.8.2 Where is the memory we have lost?

Organisational memory has been forgotten. Memory is a primary missing link. Most of the knowledge management literature discusses abstract information tools no matter where the information is stored, but not the means and content. What makes all knowledge and information concrete is the storage related to it, or the concrete memory which makes the memories physical. The crucial component presented in this research is organisational memory, and information systems that can be fit to that model.

Memory is crucial also in individual knowledge work. It is the basis for communication. What thoughts are formulated and communicated in real time is highly dependent on the actual retrieval of the memory content. But first, the history of memory itself should be presented.

The literature of organisational management – and the memory inside it – dates back several decades. The studies can be categorised in multiple disciplines:


34

Table 4. Narrowing road of organisational memory (based on Ackerman & Stein 1996a)

Theoretical Orientation

Authors Themes

Management Science

(Cyert and March 1963)

Memory as contained in procedures

Communication (Krippendorff 1975b)

Memory as 1) communication process 2) organisational structure 3) by-product of encoding/decoding. Grounded in (Ashby's 1956) notion of memory as construct based on observer's point of view.

Organisational Learning

(Argyris and Schon 1978) and (Hedberg 1981)

Memory as a consequence of learning. Memory required to store learning, but memory may be obstacle to change.

Systems Theory (Miller 1978) Memory as second stage of learning process. Human Information Processing perspective.

Decision- making and information management

(Morgan & Root 1979)

Memory as means to increase information exchange. May be designed.

Organisational behaviour

(Weick 1979)

Memory co-produces personality of the firm based on people's interpretations of their environments. Treat memory as a ‘pest’. Flexibility vs. stability.

Political Theory (Covington 1985)

Memory as information about government agencies contained in people and agency files. Memory development measured by frequency government officials use agency files or consult their predecessors.

Economics (Nelson & Winter 1982)

Memory as routine behaviour

Organisation Theory

(Smith 1982)

Memory as collective experience

Organisation and Information Theory

(Stein 1989, 1995)

Framework and empirical study of memory. Measurement of memory and expertise using network analysis techniques.

Talking about organisational management theory, Ackerman & Stein (1996b) analyse

different social theoretic positions bringing different questions to bear on KM, like Rational actor models, economic models (March, Simon), theories of organisations as bureaucracies (Blau 1955), Political conflict theories (Orlikowski 1992), Sense‐making theories (Orr 1996). Decision making has long roots as an area of computer systems; one example is IBIS application. The elements of decision making are all neatly handled by a conversational model IBIS developed in the early 1970s (Kuntz & Rittel 1972; Conklin & Begeman 1989). Even today IBIS classifies all of the points in any creative conversation into four simple elements: questions, ideas, pros and cons. (Conklin 1999.) The history is good to know, even though this thesis concentrates more on the current status of KM research.

Organisational memory research, however, also has some limitations, mainly in the individual and social management of meanings, before and after capturing information. The results of this thesis should fill those gaps, as it models the organisational memory systems, taking into account the memory mechanism – both short and long term – as well as the content of the memory. The withering paradigm of memory may flourish again.


35

1.8.3 SECI is oversimplified, ba has got cultural boundaries

The SECI model and even the ba concept have been used as a reference in the focal company in different projects over the years. Even though it is a fascinating model, its adaptation has not worked as expected. The problem with SECI is that it is based on ‘beliefs’, and mainly on creating something.

In their theory of organisational knowledge creation, Nonaka & Takeuchi (1995: 58) adopt the traditional definition of knowledge as ‘justified true belief’, but they highlight the nature of knowledge as justified belief, while traditional Western epistemology has focused on truthfulness as the essential attribute of knowledge. Beliefs are not the only type of knowledge, and the knowledge that is created can not be managed. The main enhancement for this research is to add the suitable information and knowledge species to a new model.

Also strict tacit‐explicit division does not seem to be enough. In the SECI model, externalisation easily becomes a substitute for real understanding. Transformation from tacit to explicit knowledge is not done by inventing new acronyms. For example, concept creation in product development is often done by combining letters, and not a by the successful use of metaphor, analogy or models. An externalisation is not a bunch of words without deep discussion and documentation cycles: there is a difference between information overload and learning, including understanding and memorising.

Cognitively‐oriented constructivist theories emphasise exploration and discovery on the part of each learner as explaining the learning process. Knowledge is still very much a symbolic, mental representation in the mind of the individual. But a collaborative system needs a socially‐oriented approach (if only context is engineering, like in case III).

Knowledge resides in human beings – not in the media. However, in sociological constructivism the context is part of the knowledge. The main problem with SECI is that the cycle or process of the SECI model is constructivist (Meehan 1999), while the structure is positivist (Yolles 2000). Explicit knowledge is seen as originally objective, positivist, while tacit knowledge is subjective, constructivist. The KM system based on object‐oriented analysis can utilise the concepts from the qualitative information species. Then the constructivist objective substance is the same as the positivist explicit information of the SECI model. In the new model we’ll make the difference between information as presentation or communication.

Explicit knowledge is based on positivist philosophy, while in the real world knowledge is always constructed in the present moment. Most knowledge can not be articulated. The construction is invoked in a particular situation and some information related knowledge comes to mind when we need it to answer a question or solve a problem. The positivism is a different scale of construction, compared with positivist presentation (independent on human memory and communication).

This research will also clarify whether the ba concept is valid when we concentrate more on systems, and how it differs from communities of practice, organisational memory and other Western concepts. Because the extended organisation is in question, we will investigate whether ba is organisational or cultural, and hence working in collaboration with external organisations.

1.8.4 Extended enterprise in a real actor‐network without explicit‐tacit dichotomy

One crucial missing component from current IT research is the sociotechnical context of an organisation. Knowledge management literature has not given results on the mode of operation between enterprises. It has been on either an organisational or social level – but mostly individual. The information system model for an extended enterprise has not yet been discovered.

The actor‐network theory has nowadays become concrete on information and communication technology. These systems are the core for collaboration that requires special networking skills, with less emphasis on individual achievement and more on teamwork between organisations. IT tools are facilities providing access to all information and databases, and knowledge is constructed on communication. The crucial question is why would employees be more willing to ‘share


36

knowledge’? Is the value of knowledge, like the value of information, being deflated? Also the explicit‐tacit dichotomy is not being practiced.

1.8.5 Processes should end – all knowledge is not the same

As we could see from section 1.2, Definitions of knowledge management, numbers of knowledge management process phases exist. These include e.g. knowledge identification, acquisition, generation, validation, capture, diffusion, embodiment, realisation, utilisation/application (Johnston and Blumentritt 1998) or acquisition, refinement, storage/retrieval, distribution and presentation (Meyer and Zack 1996; Zack 1999). They have two common missing links: First, these processes do not end, and as this research points out, the importance of archiving and deletion of information are of major importance. Second, there is a lack of categorisation of knowledge, which dictates what is meaningful to the process. One main problem is that most definitions of knowledge cover everything. We have tried to avoid this in the following chapter, by presenting knowledge as a subset of information, independent from communication. All the species and paradigms are crucial to enable context‐specific understanding.

CONTENTS

Part I – The model building tools



38

The definitions of knowledge management are empty loops, recursively telling us that knowledge management is management of knowledge. We should utilise practical distinctions of data, information and knowledge as well as the information classification that philosophy has generated.

The roots of knowledge management are diverse. It will not help us to simply verify that human issues and system issues are separated from each other. There must be real core how the information systems can be utilised, on one hand. Knowledge management systems are, on the other hand, a history of disappointments. Whether the current third wave is mature enough to address these problems is still unclear.

The meaning of data is underestimated, because data as content of the system have got a meaning. How it differs on the individual and organisational levels will be modelled here.

SECI and ba originate from Japan, but they can be adapted to the western philosophy with beneficial results. Universally thinking, knowledge of an organisation – even of an extended organisation – has the same sociological basis, no matter where we are. But the missing component, that will now be utilised, is memory. Organisational memory is not a separate system, and the role of all information systems should be emphasised. The functions of organisational memory, however, are not complete.

The extended enterprise has got generic processes, but processes are differentiated depending on the data, information and knowledge. This study distinguishes between knowledge and information types. Also the endpoint to all processes will be added – the value of information is a critical factor, and all information utilised in a process should have a lifecycle with an endpoint.

The actor‐network theory is in practice the network model of knowledge management. This is based on systems between information and knowledge – where knowledge content presented in the system is communicated.

Content management is the new layer between human knowledge and information systems. How this concept fits to slow human evolution versus rapidly changing technology will be examined with grounded theory: The systems are studied from a human viewpoint. Not only the method but also the theory of this research is presented in this chapter.

Systems

1 CAS

and Anseresearchinteractiqualitati1994), mmanagemproposethis met

the contgroundetheir con

engineerobserve case stud

framewoThis bottnormativinstructi

G1.1

researchbeen woare a sorfindings theory mconsistee.g. Haig

Rigorousadditionfor usingbeen dopeople k

criteria: managemobservatbeen revknowledoffer to

purposephenomthe systemanager

between inf

SE STUDY M

The stelm Strauss h case. Methionism, pragve in the sen

meaning the iment and ICT extensions thodology wi

Grountent are well ed theory wantent – which

Informring worldvieat most, usedies, while t

The reork theory crtom‐up methve model baons, built on

Grounded the

Grounh strategy. Inorking with thrt of samplesfrom the res

method is an ntly argued fg 1995).

The bas grounded tal perspectivg grounded tne yet, and wknow about t

More how informament of knotion and quevealed. Thesdge managemknowledge m

Glasers. Managemenon. Thus fems. An apprs, and admi

formation an

ETHOD BASE

trategy behin(originally 19ods are mainmatism, ethnse that it doimprovemenT model (or to the existinill result in a nded theory suited as sus adapted toh are not intmation systeew. In prior rers only – nohe case studesearch desigritically. The hod brings tosed on critiqn grounded t

eory princip

nded theory n this researche same subs from deepesults of the walternative ffor the induc

asic principletheory can ofves to those theory in thiswhere a newthe surroundexactly, thisation systemwledge. By uestionnaires e basic sociament. On themanagementr and Strauss

ment of knowfar the researoach in whinistrators, b

nd knowledg

ED ON GROU

nd this resea971), but thenly qualitativno‐methodooes not claimnt of existingstrategy) areng theories, new model was selectedbjects to be o these case erviewed, bums have genresearch, grot systems. Gdy method wgn is based oexisting reseo light the reques gathereheory, with t

les

was originalch, the ethnobject as the cer work. Thewider contexfor deductivctive discove

es and methffer managealready beins research. Itw viewpoint iding world ans study uses ams with commusing a combover an exteal studies refe other handt. s encourage

wledge and inarch on the sch informati

but the subje

ge

39

UNDED THEO

arch is groune method is cve, derived frology and hem to deliver sg theories. Ine the objectsbut – as the or a theory bd because instudied withstudies, becut investigatnerally been ounded theoGrounded theas used to seon the instruearch is baseeal situation d from empthe logic of t

ly meant forographic concase studies i method is uxt, and indicae verificationery of theory

ods of grounment researng offered byt fits this kindis needed. Wnd how theya grounded tmunication tbination of inended periodlect the attitd, the system

researchersnformation sysubject has cion systems ct is the kno

Part I – Th

ORY PRINCIP

ded theory, case study, arom interprermeneutics. statistical but this researcs of interest, scientific apbased on theformation syh this methoause the subed. studied on tory–based caeory was useeparate and mentalist thed either on ehow people irical researcthe case stud

r sociologicalditions are min the focal cused here to ated also in tn of a theoryy grounded in

nded theory rch a unique y other methd of study ar

With this mety understandtheory case sechnologies ndividual andd of time, thetudes to infoms themselve

to use grouystems shouoncentratedare studied fowledge of pe

e model buil

PLES

based on thand adjustingetive paradigThe case stut theoreticalch, the theorand empiric

pproach is use case studieystems, theirod. However,bject of the r

he basis of aase studies hed here for acompare the

heory, and vieengineering and systemsch. The modedy method.

l research bamet because company for compile a ththe case study. Glaser andn systematic

suit manageunderstandihods. There rea, in whichhod it is posd it. study approacould be utid focus groupe details of thrmation systes are studied

nded theoryuld be studied either on thfrom the vieeople, has no

lding tools

e work by Bag is done for gms such as sudy method generalisatiies behind kcal findings aed in a new es. r users and e, the originalresearch is sy

a mathematicave been usell common pe cases. ews the exisor humanists act, as wellel is a collect

ased on ethn the research years. The cheory based dies. The grod Strauss havally analysed

ement researing that provare also oth not much resible to find

ach to discovilised for thep interviews,he interviewtems suppord to see wha

y for new discd like an intehe human isswpoint of thot been don

arney Glaserevery symbolic is ions (Yin nowledge re used to context –

especially method of ystems and

cal or ed to parts of the

ting tic sciences. l as a tion of

nographic her has case studies on the unded ve d data (see

rch arenas. vides er reasons esearch has out what

ver relevant e , participant

wees have rting at they can

ciplinary eractive sues, or on he users, e before.

r



40

Two distinct approaches associated with the original authors, Glaser and Strauss, are recognised. A key question over which researchers still debate is the extent to which a priori theory should be applied in a grounded theory study. Some claim that a clear mindset is important in order to avoid interpreting in accordance with existing theories (Glaser and Strauss 1971), whereas others (Miles 1979; Eisenhardt 1989; Yin, 1994) claim that a priori theory is important for positioning the emergent theory and stimulating creative analysis. Grounded theory, if applied in its true sense, has a scope and potential for the study of knowledge management and information systems given its emphasis on context, theoretical emergence, and the social construction of realities.

In this case the researcher could as well be the object of the study. In general, the researcher should be familiar with the environment. The researcher must know the language, vocabulary and the symbols of the target groups. The meanings that the researcher observes in the field must be reported in a scientific language. And for objectivity, the researcher must consider all findings meaningful, although his experiences might be contradictory. With grounded theory there is a dilemma: maintaining objectivity versus developing sensitivity (Strauss & Corbin 1998: 43). In this research, experience is not used as data, but it gives a context to the data (see e.g. Strauss & Corbin 1998: 59).

Proponents of the inductive grounded theory approach argue that explanations of social phenomena are relatively worthless unless they are grounded in observation and practice (Glaser and Strauss, 1971). Opponents of the deductive approach also argue that scientific, quantitative methods create artificial situations and therefore affect the outcome of the research. Therefore, an ethnographic‐like approach was chosen, condensed in case studies. It is based on observing people in real situations with the researcher exerting as little influence as possible.

Since its introduction, grounded theory has been progressively developed in a way that is consistent with its original formulation, and it is the most comprehensive qualitative research methodology available. Also, this grounded theory based case study method is initially qualitative with an aim that suits this research: To build a practical model or a theory from research material based on systematic case studies. The model or theory could be substantive, i.e. from a narrow sector, or formal, i.e. possible to generalise. The method claims simultaneous systematic and creative thinking. With this method it is also possible to compose new conceptual systems and theoretical, explanatory structures. The research model and the rules of the method aim at inductive model building by comparative analysis based on empirical material (Glaser & Strauss 1971: 79 – 92; Strauss & Corbin 1998: 12 – 14).

However, this ethnographic participation was complemented by case studies using different methods. For observations and practice, three case studies were analysed. They were used to determine the key criteria (Eisenhardt, 1989), in addition to those supported by the literature, for this final report. Literature and articles do not replace data, but used for comparison with data (Strauss & Corbin 1998: 44).

Eisenhardt (1989) describes how the process of grounded theory starts with no overriding theoretical framework, but rather key criteria are developed as the study of the organisation grows. The methods used in this study, as recommended by Eisenhardt, included the literature, ethnography, semi‐structured interviews and archive material. However, the phenomena from the data formulated the first analytic ideas.

The purpose of qualitative research is to define new concepts and their relationships by conceptual ordering, and to describe what phenomena are like. Qualitative research seldom reaches the theoretical level, where the phenomena could be explained, forecasted or controlled. However, Glaser & Strauss claim that with the grounded theory method it is possible to explain and forecast phenomena. There is no fundamental inconsistency between the purpose and possibilities of qualitative and quantitative research. In grounded theory, the explanation and forecasting is not based on a causal relationship, as in traditional quantitative research, but on the object of the research itself (Strauss & Corbin 1998: 19 – 21; Glaser & Strauss 1971: 31 – 43).

Grounded theory is typically presented as a qualitative research with no statistical or other quantitative procedures. However, with some methods, such as the case study method used in this research, it is meaningful to include different data in the systemic research. Interviews and field observations can be supported by all data derived from the cases. This case study makes a distinction

Systems

betweenelaborat

Ca1.2

analysis enter thof reseaHoweve(Strauss

1995). Tstudied i

the diffesuch as pothers, sprocess scheme Corbin 1

comparadesign oqualitatiBoyatzis

such as treality. Itfor the cresearchstudy usthe procrole of li

interviewbased toEthnogra

and prioresearchthat it is (See Stra

comparaon contidescribestrategy,multipleon the reresearch

multiple

between inf

n data and ption of quant

ase study on

Grounis the interpe picture. In rch is interacr, the resear& Corbin 19

A lineThe method iteratively th

A specerent layers oproblem defsuch as withi(see Eisenhathat systema1998: 25).

Thereative methodof case study ve data by M (1998).

The dethis. As Glaset generally pconfusion surh, this thesis ses literaturecess of builditerature.

The litws begun witoo strictly onaphic study w

The loor knowledgeher uses this not the reseauss & Corbi

Initialative methodnuous comped the design, developme case analysesults and oph design base

The ca levels of an

formation an

henomena (stitative data

n grounded t

nded theory play betweenthe beginninctive in the srcher can also998: 12 – 14)ar‐thinking mdoes not comhe whole procial process oof this researinition and cin‐case analyardt 1989: 53atically integ

are theoriesd for developresearch by

Miles and Hu

evelopment er and Strauspermits the drrounding thutilises the me together wing a model f

terature wasth the forma the questiowas used befogic of groune of the reseaexperience –earcher’s pern 1998: 47). ly Glaser andd and the proparison of dan of case student of a typolis. In these rpen questioned on case stase study apalysis within

nd knowledg

see e.g. Haigis possible (S

theory

can be usedn researcher ng the reseasense that tho return to p. methodologyme before thocess is moreof building arch. It is baseconstruct valysis and repl32). Theorisingrates variou

s for buildingping the grou Yin (1981, 1berman (198

of a model oss (1971) argdevelopmenthe distinctionmore clarifieith case studfrom cases, e

s studied befal questionnannaires. Thefore, during nded theory iarcher are p– without purception that d Strauss (19ocess of builta and a moddy research. logy of case ecurring casns. Validity atudies, as staproach base a single stud

ge

41

g 1995). EvenStrauss & Co

for iterativeand data. Thrcher knowshe clues derivprevious find

y of groundehe problem ie systems the model or a ed on groundidation are sication logic,ng is the act us concepts t

g case studieunded theory1994) and co84), togethe

or a theory isgue, the tie tt of a testablns among qued case baseddies, Eisenhaespecially re

fore, during, aires, but as e interviews cand after this more deduresent and autting the ext matters, bu

971) and lateding a groundel beginninIn general, cstudy designe studies, rend reliabilityated by Eisened on grounddy (Yin: 1994

Part I – Th

n with the baorbin 1971: 1

e research likhe process iss what questved from onedings when th

ed theory hasn this researeoretic than theory fromded theory, bsimilar to hyp, are unique of constructthough state

es. The literay by Glaser adification of r with additi

s a central aco actual datae, relevant aalitative datd method of rdt’s methodgarding the

and after thopenness wcontinued afe interviewsuctive than inactive duringperience itseut how the re

er Strauss (19nded theory g with data ccase studies ns, and descrplication logy are taken innhardt (1989ded theory ca4). Case stud

e model buil

asic grounde185 – 220).

ke this. As Sts repeated wion is to be ae phase direhe results of

s been criticirch. Becauselinear. case study rbut some feapothesis‐testto the inducting from datments of rel

ture includeand Strauss (procedures onal notes b

ctivity in an oa is the connand valid moa, inductive Eisenhardt (d is also suitacentral induc

he layer‐specas the aim, dfter the forms. nductive, be the researchelf into the desearch part

987) have wrfrom case stcollection. Yhave been dription the regic is adjustednto account 9: 534). an employ andies typically

lding tools

d theory a th

rauss and Cowhen new vieaddressed. Tct the next pf later phases

ised (see e.g different ca

research wasatures of theting researchctive, case‐orta an explanationship (St

s not only th(1971), but afor analysin

by Kvale (199

organisationnection with odel, but as alogic and cas(1989). Becaable becausective process

cific case studdiscussions wmal questionn

cause the exh. However, data. Anotheticipants see

ritten about tudy researchin (1981, 199defined as a replication logd to the situin this exper

n embeddedcombine da

heoretical

orbin write, ewpoints The process phases. s require it

. Haig ses are

s used for e process, h, while riented atory trauss &

he also the g 96) and

nal research empirical a solution se study use this e it clarifies s and the

dies. The were not naires.

xperiences the r principle is the world

the h. It relies 94) has research gic for ation based rimental

d design, i.e. ta

s



42

collection methods such as archives, interviews, questionnaires and observations. The evidence may be qualitative, quantitative or both. (Eisenhardt 1989: 534.) This research utilises multiple levels of analysis and the coding is based on case studies. Interpretive procedures are an integral part of the research process. In grounded theory – or qualitative research in general, and in this research as well – some of the data may be quantified, as with a census or background information about the subjects or persons studied, but the bulk of the analysis is interpretative (see e.g. Strauss & Corbin 1998: 11).

The definition of model building from case study research based on grounded theory is summarised in Table 5 (based on Eisenhardt 1989: 533). The process of the case study research is systematic analysis as in the original grounded theory. However, the phases of analysis (see e.g. Strauss & Corbin 1998: 101 – 161) are split into more detailed levels because of the cases. This case study approach based on grounded theory is the best possible way to get the users’ real experiences of the systems observed, even though the actual subject is the systems which can not be interviewed. The documents and literature on the systems can not be included without case study method. On the other hand, the critique from the users is best available through grounded theory. The feedback helps to create normative instructions for model creation.

Table 5. Process of building a model from case study research (based on Eisenhardt 1989: 533)

Step Activity Reason Getting

started Definition of research

question Possibly a priori constructs

Focuses efforts Provides better grounding of construct

measures Selecting

cases Neither theory nor

hypotheses Specified population Theoretical, not random

sampling

Retains theoretical flexibility Constrains extraneous variation and

sharpens external validity Focuses effort on theoretically useful

cases – i.e. those that replicate or extend theory by filling conceptual categories

Crafting instruments and protocols

Multiple data collection methods

Qualitative and quantitative data combined

Multiple investigators

Strengthens grounding of a model by triangulation of evidence

Synergistic view of evidence Fosters divergent perspectives and

strengthens grounding Entering the

field Overlap data collection and

analysis, including field notes

Flexible and opportunistic data collection methods

Speeds analyses and reveals helpful adjustments to data collection

Allows investigators to take advantage of emergent themes and unique case features

Analysing data Within-case analysis Cross-case pattern search

using divergent techniques

Gains familiarity with data and preliminary model generation

Forces investigators to look beyond initial impressions and see evidence through multiple lenses

Shaping hypotheses

Iterative tabulation of evidence for each construct

Replication, not sampling, logic across cases

Search evidence for ‘why’ behind relationships

Sharpens construct definition validity and measurability

Confirms, extends and sharpens model Builds internal validity

Enfolding literature

Comparison with conflicting literature

Comparison with similar literature

Builds internal validity, raises theoretical level and sharpens possibility to generalise, improves construct definition, and raises theoretical level

Reaching closure

Theoretical saturation when possible

Ends process when marginal improvement becomes small



43

In this chapter, the words model and theory are used in parallel. There are solid reasons for this: the case‐based study is building a model more than a theory. However, the data of the case studies are the basis of the model. Generating this model from data means that most hypotheses and concepts not only come from the data, but are systematically worked out in relation to the data during the course of the research. Thus, generating this model involves a process of research, even though in other studies the source of certain ideas or even models can come from sources other than data. (Glaser & Strauss 1972: 6). In any case, this research has generated a model as valid as a theory.

At the beginning of the actual research several potentially important constructs were identified from literature and experience. But as there was not much existing literature the study began with new experimental thoughts under consideration and a new hypothesis to test – with the aim of finding an answer to the research questions. According to the grounded theory method, the researcher selects time, place and cases to study, as well as the people to interview. The number of people is not defined statistically but based on the theoretical importance. The actual interviewees’ sampling can be open, relational or variational, and selective (Glaser & Strauss. 1971: 62 – 71; Strauss & Corbin 1998: 201 – 205). In this study the selection of the cases was based on the relevancy for and quality of the subject. The selection between model and theory was not predefined, but the case studies used as a starters for observations, were more like evaluative research. Even though the case studies and especially the incremental discovery between them utilised grounded theory, the result is rather called a model than a theory.

The theoretical principles guide the selection of the cases in grounded theory. When investigating a new subject that has not been researched before with the same scientific paradigm, the researcher selects cases from which the data are supposed to be found. The existence and availability of phenomena direct the sampling. The sampling is cumulative. The simultaneous analysis of data guides the gathering while the content of the categories develops, becomes richer in nuances and more receptive to represent the phenomenon, and finally there is a saturation, after which no new information for a model will be accumulated (see e.g. Glaser & Strauss 1971: 62 – 71; Strauss & Corbin 1998: 201 – 205). Selecting cases in practice during this research is discussed more in section 1.3, Research design – the cases. The fundamental principle is that random selection is neither necessary, nor even preferable. Given the limited cases, the process of interest is ‘transparently observable’ (Pettigrew 1988).

Crafting instruments and protocols was applied in this research to make triangulation possible by multiple data collection methods. This provided stronger substantiation of constructs and hypotheses. As noted, some quantitative examples have a large amount of qualitative data. Although the terms qualitative and case study are often used interchangeably (e.g. Yin 1981) case study research can involve qualitative data only, quantitative only, or both. The combination of data types can be highly synergistic.

Glaser and Strauss (1971) argue for frequent overlap of data analysis with data collection – such as joint collection, coding and analysis of data. This was done to a certain extent in the third case study. Entering the field in the third case is utilising multiple investigators, but only to verify the researcher’s results against complementary insights. That gave confidence in the researcher’s observations.

Field notes were written during this research, but observation and analysis were separated from one another (see e.g. Van Maanen 1998) in all three cases. All the impressions were written down because it was often difficult to know what would or would not be useful in the future. However, the notes were constantly called into question, one example being ‘how does this case differ from the last?’

When repeating the cases, it was questioned whether it is legitimate to alter and add data collection methods during the research? For model‐building research, it is, because researcher is trying to understand each case individually and as deeply as is feasible. Consequently, a new data collection opportunity appeared or when a new line of thinking emerged during the research, the investigator took advantage of it by altering the data collection when such an alteration was likely to better ground a model or a theory, or to provide new theoretical insight. This flexibility was not unsystematic, but controlled – the uniqueness of each specific case was utilised.



44

Analysing the data is the most difficult and the least codified part of the process, but it is the core of building a case study based on grounded theory. When reading reports and quotes from a piece of research, one can not ordinarily follow how a researcher got from thousands of pages of field notes to the final conclusions (Miles and Huberman 1984: 16). In this research the volume of data was not excessive despite the open‐ended questions. This was mainly because of the descriptions of the cases. The write‐ups of each site helped the analysis process (see e.g. Gersick 1988). In this research the investigator became very familiar with each case as a stand‐alone entity, and followed their development long after the questionnaires were analysed. The analytic blinders of the research tools were avoided: in grounded theory, it is crucial that the analysis techniques help and only help the analysts to recognise the properties of what is in the data (Strauss & Corbin 1998: 96).

A cross‐case search for patterns is a crucial part of case analysis based on grounded theory. Reliability increases when conclusions are not based on limited data (see e.g. Kahneman & Tversky 1973 or Nisbett & Ross 1980). In cross‐case comparison the data is examined in many divergent ways. It is possible to select categories or dimensions and then look for similarities coupled with intergroup differences. Dimensions can be suggested by the research problem or by existing literature, or the researcher can simply choose some dimensions. In this research ‘layers’ of the cases are used, but also divisions of information systems and information species. The similarities and differences between the cases are explained. Third strategy is the division between data sources, since different data collection methods are used.

Hypotheses shaping is done by an iterative process comparing systematically the emergent frame with the evidence from each case in order to assess how well or poorly it fits the case data. The researcher constantly compares model and data, iterating toward a model or a theory that closely fits the data. Shaping the hypotheses is done first by sharpening the constructs, i.e. refining the definition of the construct and measurements of the construct in each case. The construct – as well as its definition and measurement in this kind of grounded research – often emerges from the analysis process itself, rather than being specified a priori. When attempting to construct validity, the researcher must collapse multiple indicators into a single construct measure, because the indicators may vary across cases, and qualitative evidence is difficult to collapse.

Hypotheses shaping is also done by verifying that emergent relationships between constructs fit the evidence in each case. The key difference between case research and the traditional testing verification process is that each hypothesis is examined for each case by logic of replication, not for the aggregate cases. The first examination and interpretation in this research was microanalysis, like in grounded theory — but not necessarily in a case study (Strauss & Corbin 1998: 58).

The literature which ‘enfolds’ the object of the research, must also be analysed. The comparison of the emergent concepts, model, theory or hypotheses with the extant literature involves asking what is similar, what is contradictory, and why. There are mainly two reasons for this examination: First, if the conflicting findings are ignored, confidence in the research is reduced. Second, conflicting opinions in the literature represent an opportunity by forcing the researcher into a more creative, frame breaking mode of thinking. Therefore, this research includes comments mainly about contradictory findings in the sparse literature. Literature discussing similar findings is important as well. It ties together underlying similarities in phenomena normally not associated with each other.

In trying to reach closure, two important questions emerge: when to stop adding cases, and when to stop iterating between model and data (Eisenhardt 1989: 545). As Glaser and Strauss have written, theoretical saturation is simply the point at which incremental learning is minimal because the researchers are observing phenomena seen before (Glaser and Strauss 1971). Also, the iteration process stops when the incremental improvement to model is minimal. The final product may be a model, concepts, a conceptual framework, propositions or possibly mid‐range theory. On the downside, the final product may be disappointing if the research simply replicates prior theory, or there may be no clear patterns within the data. In this research, conceptualising led to classification (see e.g. Strauss & Corbin 1998: 103).

Comparison with other literature is quite common in all research methods. Eisenhardt has extended theoretical sampling and saturation together with overlapped coding, data collection and analysis (Glaser and Straus 1971) by case study design, replication logic and concern for internal validity

Systems

(Yin 1994here haswork con1984; Ki(Van Mafrom casreferenc

testing, instrumeclosely r1987), wtheory. Pproperti

that are these meconstrucknowled

cases is ijuxtaposcontradiSecondlyand hypovalid, be

can yieldtestable,Perhaps

researchfulfilled extensioparticulainadequother orstudy inf

research(about adependshypothethe data

Re1.3

systems whole re

between inf

4) as well as s differencesncentrated ork & Miller 1aanen 1988).ses, rather thces from grou

The pbut includesentation, croelated to eth

where the ricProcedurallyes and dime

The caconsistent, rethodologicactivist views dge.

This reits likelihoodsition of contictions forcesy, the emergotheses thatecause the pr

Case‐bd an overly c, novel and ea grand theo

Buildih or to providin this reseaon (see e.g. Karly appropriate because with commoformation sy

Whenh process is t ‘good theors upon empireses testing, aa.

esearch desi

The fube used for esearch and

formation an

the tools of s from previoon other top1986), case st Glaser and han explaininunded theorrocess descr a priori specoss‐case analhnography (Vh, complex dy, axial codingnsions (Strauase studies hreproducibleal notions arare discusse

esearch procd of generatitradictory evs the researcgent model ot can be provrocess is tiedbased modeomplex theoempirically vory requires ng a model fde a fresh perch. Theory Kuhn 1970). Hiate if little isthey have liton sense. In ystems from n evaluating to begin to dry’, see e.g. Prical issues, sa strong theo

ign – the cas

undamental managementhe case stu

nd knowledg

tabular dispous work. It hics such as qtudy design Strauss (197ng the meanry together wribed by Eisecification of lysis and sevVan Maanendescription og is used to russ & Corbinhere are close, possible tore not to be ud more durin

cess is iterating novel thevidence (Camcher to refraor theory is liven true. Thed to the evidel building haory. It may alalid, but are an accumulafrom case stuerspective tois normally dHowever, mos known abottle empiricathis researchthe viewpoithe researchevelop a parPfeffer 1982)strength of mory‐building

ses

question in tnt of knowledies. Concer

ge

45

play of evidenhas focused oualitative da(Yin 1981, 1971) originallys of how to dwith case stunhardt also cconstructs, peral uses of n 1988) or quof the specifirelate categon 1998: 124 –e to groundeo generalise, understood ing the analy

ive. One streeory‐like resumeron & Quinme perceptikely to be tee model, theence. s some riskslso result in nnot theoriesation of multudy researcho an already rdeveloped thodel or theoout the phenoal substantiah the subjectnt of organish method, wersimonious, t). The assessmethod and tstudy has a

this researchdge?’ Semanrning the ent

Part I – Th

nce (Miles anon model buata analysis (994; McClinty focused modo it. Therefody methodscontributes npopulation sliterature. Hualitative anac cases is as ories to subc– 127). ed theory: thand meet otin a positivissis of the lim

ength of grouults. Creativenn 1988). Thons into a neestable with ory or other

s too. The intnarrow and is about orgatiple theory h is most appresearched through increry building fromenon, if ction, or if tht itself is newsational knowe must defintestable andsment of mothe evidencegood (althou

h was ‘how cntics are somtire study, th

e model buil

nd Hubermauilding from ce.g. Miles 19tock et al. 19ore on defenore, this rese. new ideas. Itpecification,owever, thisalysis for themuch an aimcategories alo

he scientific mther acceptet sense. The

mits between

unded theorye insight oftee process ofew gestalt (Bconstructs th result is like

tensive use oidiosyncraticnisation in abuilding andpropriate in ttopic. Both omental emprom case stucurrent perspe perspectivw, and the newledge manae the criteria logically cohdel‐ or theore grounding ugh not nece

an informatimewhat diffee grounded

lding tools

n 1984). Thecases, while 979; Miles &979), and ethding buildingearch utilised

t is not just h flexible s research is social scientm as buildingong the lines

method giveed canons, alpositivist an

n information

y building froen arises fromf reconciling Bartunek 198hat can be mely to be emp

of empirical ec theories thaany grand send testing studthe early stagof the conditiirical testingudy researchpectives seemves conflict wew perspectagement. a. The point herent modery‐building rethe theory. Aessarily perfe

ion managemrent compartheory kind

e process previous Huberman hnography g theory d later

hypothesis

more tist (Strauss g a positivist s of their

es results lthough nd n and

om three m the these 88). measured pirically

evidence at are nse. dies. ges of ions are g and is m with each ive is to

of this el or theory esearch As in ect) fit with

ment ring the of question



46

means ‘is it possible at all that knowledge can be managed with information systems’5, whereas the case studies started the interviews and deeper observations by studying ‘in which way management is supposed to be possible’. The question was approached in the first case by a survey of existing data and information systems for knowledge management – at the focal company, and ongoing surveys of the market. In the second case one generic knowledge management application category was studied more deeply, and projected against the requirements of the knowledge workers. This was done with interviews supported by basic statistics. In the third case the human knowledge requirements versus a tailored system were approached independently of the existing applications.

The final test to judge the quality of the grounded theory‐based case study aims at examining the reliability of the study. The aim is to ensure that another researcher, doing a study on the same subject again, and following the same procedures, arrives at the same findings and conclusions (Yin 1989). This theoretical option must reckon with other sources of information. The knowledge management literature and conferences were followed iteratively before, during and after the case studies. To name them by mnemonics, the three cases of the research were modelled as follows:

• DATA layer. Data management systems: Managed individual knowledge versus all existing systems: – finding out what kind of applications there are.

• INFORMATION layer. One general KM approach with two information systems compared: Managed knowledge versus features of two general systems: looking more deeply at how the systems matched the various needs of different organisational units, on publicly available platforms (more specifically, documentation management)

• KNOWLEDGE layer. Knowledge management independent from systems: Managed knowledge versus one tailored system: independent from the existing systems, independent from the organisation.

The reported case studies are knowledge management researches within a European multinational company and its collaborators, already referred to as the focal company.

Before, in parallel with, and after the case studies the researcher participated in many lecture series at universities, such as:

• Database Symposium, Helsinki (1997) • Lotusphere Europe. Berlin (1998) • Tietotekniikka. Nyt, University of Helsinki (2000) • Dokumenttien hallinta, University of Helsinki (2001) • Tiedon johtaminen, Helsinki University of Technology (2002) and • Dokumenttien hallinta ja tietojohtaminen, Management Events (2006 ‐ 2009) • ITIL Foundation, Tieturi Helsinki (2009)

The papers published, presented and participated at conference proceedings include:

• Organisational memory information systems (APROS 2000) • Integroitujen dokumentaation hallintajärjestelmien ominaisuuksia.TKK

tietojohtaminen, (2001)

5 One construct of this research was to study further the results from the literature review, as some referenced articles claim that information systems can not be used for knowledge management – either from social constructivism’s (e.g. Lang 2001), or IT point of view (e.g. Wilson 2002). As research is always comparison (between systems, between theory and observations, or between new and existing results), also findings on other knowledge management aspects using method described in Table 5 during case studies are versatile.



47

• Information systems for management of knowledge. Telebalt 2002 • Knowledge management and Information Systems – Finding a Sociotechnical

Golden Mean. Key note presentation for ROCKET Workshop in ICE 2003 • Sociological context for Extended Organisational Knowledge. The Proceedings

of the 9th International Conference of Concurrent Enterprising 2003 • Managing Knowledge in Networked Product Creation Organisations. Presented

in Vienna (2004) and published in P.Cunningham & M.Cunningham (eds.) eAdoption and the Knowledge Economy

• Practical Knowledge and Collaboration in Engineering. Presented in Karslruhe (2004) and published in G.Zülch, H.S.Jagdev & P.Stock (eds.) Integrating Human Aspects in Production Management

• Support for Knowledge and Innovations in Software Development – Community within Company: Inner Source Environment. Presented in WebSphere conference, Barcelona (2007)

• Dokumenttien hallinta ja tietojohtaminen. Keynote speech, Espoo (2007) The case studies were more like field studies than system evaluations. One characteristic

of a grounded theorist – the sensitivity to the words and actions of respondents (Strauss & Corbin 1998: 7) – was present in every study. In all cases the systems were compared with organisational human knowledge – and the resulting frame – systems between information and knowledge. That gave us the following table:

Table 6. Comparison of cases

Variant CASE I CASE II CASE III

Level Data Information Knowledge

Scope From individuals and groups to organisational units

From organisational units to focal company, including processes

Extended enterprise, processes between companies

Number of systems Many systems Two systems Unique system

Phases of each case Phase 1: Many Systems

Phase 2: Users as individuals from many organisational units

Phase 1: Users as organisational unit processes and roles

Phase 2: Two Systems

Phase 1: Users from many organisations

Phase 2: One system

The matrix helped to get deeper results: the interplay between conditions, the responses

of actors and the consequences that resulted from each case. The matrix was a ‘conceptual guide’ (Strauss & Corbin 1998: 193). Case I concerned mostly data, while case III was mostly about knowledge, but they were used for comparison with the information layer (case II). However, all the possible comparisons were made between data, information and knowledge.

Each case had its own two phases (systems vs. users’ perspective) but the phases also had different sub‐phases, such as data collection, analysis and interpretation, and each sub‐phase entailed choices and decisions concerning the usefulness of various alternative procedures, whether these were qualitative or quantitative, but also decision on which qualitative and which quantitative method to use (see e.g. Strauss & Corbin 1998: 29). The phases that opened discussions with users of the systems on each case were basically evaluative research. Cases included constructive parts, and the methods presented in this chapter were used to combine all evaluations, iterative interviews and participatory



48

observations. The slightly modified iteration of ambiguous subjects was the main role of replication. The decisions, how the interplaying methods were selected, are explained in each chapter according to the case. The interviews and other data gathering methods are explained in the sections presenting each case. In this final report the way how citations are presented is varying from cases I to III on purpose.



49

2 INFORMATION AND KNOWLEDGE SPECIES

This chapter presents a philosophical and practical background of information and knowledge from the viewpoint of knowledge management and, especially, information systems for KM. Information and communication technologies (ICT) have become the symbols of our age. The argument, that post‐industrial organisations and societies live by information, is provocative. Something more than information is valuable. Is the gap between ‘knowledge’ processed with ICT‐based KM systems and human knowledge insurmountable, or can the information subspecies be analysed and selected so that information and knowledge complement each other? Is there something in between?

There is also a risk when different information species and even knowledge are treated in similar manners. This research characterises the observed distinctions between human knowledge, data processed with information system and the information species generic for both. Data is the basis of information systems, the aim being human knowledge: but systems and users share information. It is the common factor between systems and users. As the information and knowledge concepts have been misinterpreted, information species are analysed more deeply. The systematic classification of information, data and knowledge could be benefitted by philosophy. The result would be a pragmatic appliance of knowledge. In current practice, knowledge and information are either totally distinguished, or managed in analogy with each other. This should be reconsidered.

Knowledge differs from information and data. The social construction of knowledge, however, is also different from the cognitive model of knowledge management6. Plato’s ‘justified true belief’ is the first known definition of knowledge. Other philosophers from Aristotle, Descartes, Locke, Kant, Marx, Hegel, Dewey, Husserl, Sartre, Wittgenstein, Polanyi, Heidegger, Merlau‐Ponty, James, Habermas, Popper and Tsoukas to Nonaka and Takeuchi, have debated this concept through doctrines, such as continental rationalism, British empiricism, German philosophy or modern doctrines of twentieth‐century philosophers. In general, Western philosophy has adopted Plato’s definition, although it is imperfect in logic (Kakabadse et al 2003). ‘Belief’ is regarded as a way of reducing doubt and uncertainty. Even Nonaka & Takeuchi referred to Plato’s definition. This research approaches the concept with wider and more practical categorisations of information and knowledge species.

We began with a description of the current narrow and short philosophical history of knowledge management concept. In the background, rationality is the viewpoint for decision‐making, and it gives a little attention to knowledge. Other kinds of background and information species should be emphasised when dealing with ICT – even with engineering information.

The traditional data, information and knowledge classification is presented together with comments from the philosophy of information. Systems and users share information, which is why it is more important than data as the basis for information systems. Human knowledge, however, is the final aim. Information is the common factor between systems and users, but not categorised in everyday practice. Here the different species of information and paradigms of science are analysed more deeply.

ICT consists of information processing and communication technologies. The same main streams can be found in philosophy: information and communication theory. Information theory classifies physical, syntactic and semantic information based on probability. A semiotic model is a link to communication‐based information categories. Communication theory gives us a variable amount of information categories. Some of these are relevant for modern ICT, but not included in traditional information management. Notions about these categories mainly concern pragmatic information for engineering in the case studies.

Pragmatic information in the communication context is divided into epistemic, doxastic and modal information species. Information overload is discussed with the critical factor of pragmatic

6 Many knowledge management models belong to this ‘cognitive’ category by Swan and Newell (2000). The models are based on positivistic view of modern civilisation, assuming that knowledge lies with individuals, treating knowledge as an ‘asset’ that needs to be managed, and consisting of formal KM processes. Examples from prominent cognitive models are mentioned in the Introduction and chapter 1.7, SECI and ba.

Systems

informatinformatultimatebe categand retrknowled

analysis framewopaper.

Pa2.1

no univeturn of tdissemindefined & Sensip

was the perspectbe reme

rationaliparadigmmanagem

and orga(conceptmodel isjustificat

philosopthe natuparadigmb) post‐s

data as fPositivisextende

probabil‘engineethings. Oengineerimplemeoptimisaobjective

economladen wiand post

between inf

tion: noveltytion systemsely processedgorised. Tacitieval are notdge.

Informis an engineork for study

aradigms fro

How dersally accephe millenniunation of knoas ‘informatper, 1998).

Experfirst to claimtives argue tembered whe

It is alstic interprems of philosoment.

The manisational rets, objects, ps also concertion, causatio

The paphy) concernure of knowlems used herestructuralism

For pofacts. Positivm is a simpled to behavio

Post‐pistically. Theering view’ thObservers haring processeented. Peoplation becomee.

The cric, ethnic, anith respect tot‐structuralis

formation an

y. Data and ts, decision md in the humt knowledge t independen

mation and cering disciplying and mod

om philosoph

does knowleted definitioum (see e.g. Towledge to bion that is re

ience as a som that ‘knowhat proceduen talking abso doubtful etation of huophy and spe

main concerneality and hoprepositionalrned with theon, doubt anaradigms fros beliefs aboedge and thee are 1) positm, and 4) conositivistic ontistic episteme belief in scourism in psypositivistic one epistemolohe observersave consciouses convertine can be betes relevant.

ritical theorynd other facto the worldvsm. Critical t

nd knowledg

he informatiaking and enan framewoincludes notnt from cont

communicatiine based ondifying the w

hy and scien

dge manageon for knowleTsui, 2000). benefit an orelevant, actio

ource of knowledge itself ires for enqubout ‘knowlethat businesman behavioecies of infor

of a philosoow one focusl) and the soe relationshind revocabiliom philosophout the form e relationshiptivism, 2) ponstructivism.tology realitymology is basience in Wesychology andntology findsogy admits ths can have thsness that (ing informatiotter or worseMind is biase

y is based ontors shape review of an enheorists hav

ge

50

ion derived fngineering. Brk which inct only the facext. Ultimate

on theory livn this theoryworld is used

nce

ement then dedge manageMost definitganisation aonable, and b

wledge was s power’. In uiry are also tedge sharing’ss‐oriented kour and decirmation to fi

ophical modeses on objecturce of knowp of knowledty (see e.g. Khy and sciencand nature p between tst‐positivism

y can be apped on objectstern industrd naïve systes an objectivhat only an aheir own persn extension ton acquired ae at this engied machine,

the ontologeality. The epnquirer. Enque undertake

Part I – Th

from it coverBut data, infoludes tacit kcts, but relately context i

ve side by sid of the naturfor some ex

differ from inement. The stions share thnd its individbased at leas

first definedthis traditiontechniques o’. knowledge msion‐makingnd out what

el is how onetives (valueswledge (percdge to other Kakabadse etce can be disof reality, anhose who knm, 3) critical t

prehended, ativity, a possrial history. Tmic thinkingve reality thapproximate spective, whto simple beas observationeering proc reality is act

gy that realitypistemology uiry is value n a critique o

e model buil

r special cateormation andnowledge. Etionships bets important

de with systere of systemsamples abou

nformation msituation hashe perspectiduals. Typicast partially o

by Francis Bn, the criticaof power. Th

management . We must tat is meaningf

e gathers infos, abstractionception, memnotions suct al. 2003). stinguished bnd epistemolnow and knotheory with a

and there aresibility of findThe results ag. t is appreheimage of reahich can influehaviourism)on from ‘outcess, and at tually out th

y is virtual, ais subjectivisdetermined of positivist

lding tools

egories very d knowledgeven tacit knotween the fafor all huma

ems theory. Ss. This analysut engineerin

managements changed litve of collectally, knowledon experience

Bacon (1605)l and post‐mat is obvious

is based moake an overaful for knowl

ormation abns, minds), tymory, reasonh as certaint

by ontology,logy, which cwing. The foa) postmode

e observer inding universare mechanis

nded imperfality is possibuence the wa is seen to beside’ into infleast a fuzzy ere, and kno

and that sociast. Findings in both postknowledge,

relevant to e are owledge canacts. Search n

Systems sis ng in this

t? There is tle from the ion and ge is e’ (Leonard

). He also modernist s and should

ostly on a all view at edge

out social ype n). The ty, belief

which (in concerns our main ernism and

ndependent al truths. stic science

fectly and ble. As an ay they see e a set of formation owledge is

al, political, are value tmodernism favouring

n

Systems

post‐strucreation1991). Hentity, brestrictio

knowledpower. Koperatin

ontologyspecificaand epissituation

worldvieor interaworldviechange a

organisaorganisashows ththreat oftechnolo

Th2.2

as the ancreated philosopother arepistemophilosop

philosopknowledobjectivethe ratio

Humber

7 ‘Lebensepistemo

between inf

ucturalism a and politics

Habermas’s (1but is the proons and repr

Foucadge constantKnowledge dng domain of

This sty of this reseally constructstemological n. Subjectivis

Thereew. The interaction occursews, which aas the institu

The pation, where ational capachat KM doesf this actor‐cogy belongs i

he rational lo

Rationnswer to all osome interephy of artificieas as a comology, logic aphical view fr

Informphy. The ratiodge, intelligeely, and optionalistic pers

1. Chpr

2. Fi3. A

w Autop

to Maturana

welt’ also origological enquir

formation an

nd post‐mod, values, ideo1972; 1984) oduct of intenression of theault (1980) asly induces thdefines an anf power (Towtudy has comearch finally rted realities;knowledge st epistemolo are no obseraction of difs in a framewre relative toutional realithilosophy‐bathe environcity, it has a v not need tocentred parainto holistic

oad of KM

nalism has itour questionsting but limial intelligencmputational oand ontologyrom knowledmation and konalistic persnce, behaviomal solutionspective in th

haracterise roperties. nd general pply the rul

what should

poiesis is alsoa, applied in

ginates from bries.

nd knowledg

dernist theorology and cocritical theontional and se establishedsserts that thhe effect of pnalytical spacwnley 1993). mpared levelrelates to co; ontology ofis created in ogy relates tervers, only vfferent worldwork, ‘lifewoo the instituties change.ased paradigment is condvery low dep be technolodigm is that systems as w

s strengths,ns about knomited researcce. On the otor informatioy, or philosodge managemknowledge mspective hasour and cognns to problemhree steps (W

the situatio

rules that ales logicallybe done.

o a cross‐disccomputer sc

biology (and c

ge

51

ries by exposontrol (Kouzmry argues thsometimes ud order and ihe exercise opower. It is nce and provid ls from positonstructivismf constructivinteraction o created finviewers. Viewdviews occurrld’7. Knowletions that on

gm of knowleductive to anpendency onogy intensiveit has too rawell.

but it has itswledge manch niches in pther hand, inon‐theoretic phy of langument is too tmanagement strongly inflnition in genems deduced oWinograd & F

on in terms

apply to situy to the situa

ciplinary concience by Wi

cultural Protes

Part I – Th

sing the connmin 1980; Alvat knowledgunconscious ideology reqof power crenot possible fdes the basis

tivism to conm, according tism states thbetween enndings. ws, like behars through a edge is basedne is attached

edge managen open, quali technology.e and should ationalistic a

s traps too. Tnagement. Inphilosophy, information smethodolog

uage, sciencetight and narliterature is luenced the eral. It assumobjectively. WFlores 1986)

of identifia

uations in teation of con

ncept from thnograd and

stantism), and

e model buil

nection betwvesson 1991ge does not eactivity. To bquires engagiates knowlefor knowledgs of action an

structivism ito which thehat reality is quirers and p

aviours are dsemantic cod on experied to in a give

ement is cruty dialogue. . The philosonot be techview of the

The main pronformation syncluding comsystem sciengy; those aree and mind. rrow – or toorooted solelunderstandimes that theWinograd an:

able objects

erms of objncern, draw

he biologists Flores. Yet th

d has develop

lding tools

ween knowle1; Alvesson aexist as somebe free from ing in self‐refdge, and conge not to engnd intervent

in different cere exist bothrelative pheparticipants

erived from ommunicatioence throughen society, an

cial for any lBecause of i

ophy of infornology driveworld. Howe

oblem is thatystem sciencmputer ethicce is absorbeeas include However, tho dichotomicly in rationaling of inform world can bnd Flores sum

s with well‐d

ects and prwing conclus

Francisco Vahe autopoies

ed into a fund

edge nd Wilmott e abstract the flection. nversely, gender ion or the

context. The h local and nomenon, of a

a on process, h nd they

earning its strategic mation en. The only ever,

, it is touted ce has cs and the ed within

he current cal. istic

mation, be described mmarised

defined

operties. sions about

arela and sis and

dament for

Systems

systems especialrepresenand elecprocessihuman binformatnot an apractice,past expsolution intuitive

western seeking organisadecisioncoupled advantagsystems.

In2.3

procedutesting tsimple. Wrationali

philosopexplicit, Knowledexperienpropositof know

filtering such as sthe procconstrucCAD dat

Informatenquiry,perceptupriori). A

in a bioloobjectivesomethi

8 Besides understa

between inf

originate froly for enginent concepts actronic memong in analogbeing as a prtion processbstract activ, problem soperiences, anis and what ly (see e.g. v

The raphilosophy and serving ational cultur making is. Rinformationges of dynam.

nformation b

A ratiores that can o show howWe must anasm; in some

The inphy. Current even thoughdge by acquances, which ltional knowleledge is ofte

What are effectivesorting, comcessing of anyctivist, no maa to a simila

Tradittion and kno to derive knual experiencAt individual

In a boogical contexe knowledgeng personal

distinction bending depend

formation an

om biology, teering, it is eaas ordered scory as analogy with a ratiooblem solveing metaphovity that takeolving requirend understanmakes us w

von Foerster ationalistic ilof knowledgrational and re will confirRarely simplen managememically chang

between dat

onalistic parabe exteriori

w managers oalyse data, in cases mananformation cconstructedh it can utilisaintance is baead to a furtedge tells usn the basis othen is explie as long as tparing, or viy interpretatatter how trir design wortionally, the owledge are tnowledge froce (a posterilevel this coook Autopoixt. The idea e the universand private

etween informds on the cont

nd knowledg

they are philasy to buy pachemata, cogies. But infonalistic mor – a data pror for humanes a problemes interactionding of the pant to solve 1970). lusion becomge managemcommonly um how compe, rarely ratint and cogniging human

a and knowl

adigm impliesed by obseror engineers nformation agers and engoncept was knowledge e some dataased on expether distincti that the earof decisions.icit? Even infthey are procsualising dattion. The intevial it is. For rk. How are tmain distinctthe results oom informatiiori) but suchuld nowadayesis and Cogis that althose appears sythat can not

mation vs. igntext. Objectivi

ge

52

losophically aradigms anmmunicatioformation sysdel are inaderocessing sysn knowledge ‐description n, knowledgproblem situthe problem

mes extensivent presumeunderstood oplex, muddy,onal, more litive psycholknowledge a

ledge

es order, a syrvation. Resethink and mand knowledgineers only also presentcontains infoa and informaerience, but ion betweenrth goes arou

formation sycessing datata, but their cerpretation oexample, anthese categotion has beeof the humanion means thh knowledgeys be called kgnition, Humough objectivystematic ant be transferr

orance, the Gty and subjec

Part I – Th

different. Fod frameworkns and speecstems that cequate. For estem based ois a failure. Pas input and

ge, expertise,uation. Whatm, are all answ

ve when talkes that organobjectives. E, dissensual,ike life (see eogy form a sand the still m

ystematic arearchers canake decisionge separatelthink they mted as an oppormation, buation that arwe can also direct and iund the sun

ystem functio. They workcapabilities aof information engineer caories linked toen made betwn interpretathat much knoe can be usedknowledge tberto Maturve knowledgd predictablred. That wh

Greeks alreadyctivity are pres

e model buil

or knowledgeks such as a sch patterns aonsider humexample, expon input, proProblem solvd produces a, conceptual constitutes wered by hu

ing about ornisations areEven a cursor contingent, e.g. Hannabuslotted link. memory bas

rangement on use methodns. The worldly to avoid thmake decisioposite to ignut also bits ore assumed trecognise thndirect knowand one plus

ons, such as when applieare rather limon could alwan examine aogether? ween data anion of data. owledge is bd to understaransfer. rana describee seems pose, knowledghich one belie

y remarked hosented furthe

lding tools

e managemesystem, whicas scripts, anman knowledpert systemsocessing and ving in the rea solution as isation, learna problem, wuman cogniti

ganisations: rationally ory study of and partisanus 2001). TheWe must reced on inform

of conceptuadologies fromd, however, ihe traps of pns. orance in Grof ignorance; to be objectivhings withouwledge. Simps one is two.

search, retried to appropmited when ways be consia situation a

nd informatiIn a construcbased on senand new thin

es knowledgssible and thrge as an expeeves to be tr

ow subjectivityr in results.

ent, and ch nd human dge s see the output. Theeal world is output. In ning from what its on

the rdered,

n much e tightly cognise the mation

al m pre‐post s not that seudo‐

reek it is not ve.8 t sensory ply, . This kind

eval and riate tasks, applied to idered nd compare

ion. ctivist sory or ngs (a

ge transfer rough erience is ransferable,

y of



53

objective knowledge, must always be created by the listener: he or she must be prepared to understand (Maturana 1980.) The knowledge transfer and understanding are more complicated at organisational level.

Understanding as the final form of knowledge is subjective, but the formulation of understanding depends on context, including media. The spoken word and other human interaction will transfer different knowledge compared with written documents. Knowledge of the document does not mean understanding of the content. No KM system ensures understanding. It is crucial to update knowledge constantly, and not only documents or KM systems. Research into the field of existing KM systems indicates that early efforts to document lessons learnt or best practices – filling reams and reams of web pages with details – has proved of marginal benefit (Highsmith 2000). However, this could depend on whose benefit is measured. Even more crucial question is what information species is it meaningful to present in or communicate by systems.

2.3.1 Information species based on probability

As the data, information and knowledge are separated, the middle layer remains crucial. Rationalistic philosophy of information is based on probability and syntactic communication. The probability interpretation of information gives three categories (Virtanen 1989):

1. semantic, 2. syntactic and 3. physical information.

These categories of quantitative interpretation could be extended and presented as follows (All categories are adapted from Hintikka 1982; Niiniluoto 1980; Virtanen 1989):

Figure 6. Information species based on probability

Physical information and syntax

Physical information is the orientation degree of systems, the opposite of entropy. In theory, it is the common denominator that can bring matter, energy and time into a single, unified framework of analysis. All matter‐energy transformations are change of state information that can be represented with data. Animate and inanimate objects – information condensations of matter‐energy (e.g. DNA, atom, galaxy) include the more information the more complicated they are. They are

Probability

SyntacticCommunication Semantic

Presentation in language

-novelty- content

- relative information

PhysicalExistence

ArtefactsNature



54

information condensations of matter‐energy. Things can be informative in the sense that humans derive information from them. For example, a stump of a three contains information about its age as well as about the climate from those years (Hjørland 1997).

This kind of information is not too far from all organisations: In organisational memory all objects store some amount of information in organisational memory bins. Actually, meaning can be generated from nearly all physical information – Or, written with negation: we are unable to say confidently of anything that it could not be information (Buckland 1991). Physical information can be further classified as natural and man‐made artefacts.

Syntactic information is attached to communication in any channel where messages are sent and received using some notation system. The amount of syntactic information depends on the rarity of each notation string. The theory of syntactic information is very close to the statistical‐mathematical information theory based on the works of Claude Shannon and Norbert Wiener. Engineering information can also be very close to syntactic and physical information. However, when someone creates or utilises syntactic information, there is always an interpretation of that information – even with completely automatic systems. The interpretation of syntactic information is a new kind of information class (Virtanen 1989).

Semantics and semiotics

Semantics is basically the study of the relationship between what an object represents, and the object itself. Semantic information is attached to declarative sentences about states of affairs that have a linguistic meaning. This term was presented first by R. Carnap and Y. Bar‐Hillel in the 1950s, but K. Popper already launched the idea during the 1930s. The generic idea is that information is the same as eliminating ambivalence. The probability of a sentence is inversely proportional to its information. That is because information is the amount of ambivalence, which disappears when we know that the sentence is true. For example, the sentence ‘This airbus must be ready next June, July or August’ is less informative than the sentence ‘This airbus must be ready next July’, because the state of the situation is more rare in the latter sentence. (Virtanen 1989).

Even though semantics belongs to information species based on probability, semiotics is presented here also, because a semantic interpretation can be represented with semiotic triangle. The semiotic triangle model includes three elements:

1. a perception of something that exists in the physical world, 2. an object or concept to which the perception is said to refer, 3. a thought, image or concept that is formed in the mind as a result of the

perception and which relates to the object. It is this relationship, present among the elements of the many signs that human beings

constantly encounter, that forms the basis for the patterns of meaning that develop in human communication. Models involving the use of signs have been developed by a number of semiotic theoreticians (see e.g. Pearce and Eco).



55

Word

Significatio(meaning,sign)

Form Referent(Object)

Convention

Experience

Perception

Concept

Figure 7. Enhanced semiotic triangle (based on Niiniluoto 1980: 118)

This approach (presented in Figure 7) is slightly wider than the basic referent‐signifier‐

signifient triangle of semiotics, where the ‘concept’ covers both the word and its form. In practice the semiotic triangle means that documents contain information only in a metaphorical way. A document may have meaningful marks, but the meaning is something attributed to the marks and is not a physical property of the marks. The meaning of the marks can change, even if the marks do not change (Buckland 1991).

The representation problem is the same with knowledge as it is with information. It can be represented by means of facts, rules, manuals, or best‐practices descriptions, but these representations differ fundamentally from the knowledge that they are to represent. As is the case when converting data to information, the knowledge representation has to be interpreted in order to regenerate the knowledge, to put the knowledge into context and to apply it when acting in a situation (Lueg 2001: 152).

The semiotic triangle can be used in practice, as presented later with pragmatic information. However, traditional information management approaches to information seeking and knowledge sharing do not highly enough value the role of language. In distributed systems, such as global organisations, where people rarely meet each other, language is the tool by which people communicate. Communicating interests and information needs to others does not require explicit and formal descriptions, but copes with vague and, in a computational sense, incomplete information and even tacit knowledge. Language plays an important role in local and global workplace communication. (Ehrlich & Cash 1994). This is especially concerning extended enterprises.

There are many subspecies in semantic information (see e.g. Niiniluoto 1986: 72). There is a set of relative information, which includes: added information, relativated information, shifting information, surface and deeper information (Hintikka 1982: 255). Analysing the species of information is not futile for engineering – it helps to recognise what is relevant and what is not. Based on the qualitative or substantial side of information we understand the differences of objects. For example, knowing the difference between GSM transceiver station and UMTS transceiver station has nothing to do with probability.

In practice humans are social beings and our understanding of features of information, such as quality, relevance, importance, and timeliness, depends on context, culture and education. While probability gives a mathematical information system interpretation, the communication theory leads us to a more human viewpoint.



56

2.3.2 Information species based on qualitative interpretation

The practical classification of information is much more valuable for organisations. This classification will be used in our model, because it combines management science, philosophy and communication. (See e.g. Virtanen, Niiniluoto, Hintikka etc.) The classification is quite perpetual, and it was combined when Finnish philosophy results were measured with something other than publicity.

Qualitative interpretation consists of 1) communication, 2) presentation and 3) information processing interpretations that give us species of pragmatic (expressive and knowledge‐related), epistemic, doxastic, modal, data‐derived and meta‐information.

Qualitative interpretation of information as communication, as presentation or as information processing gives seven species (based on Virtanen 1989).

1. Pragmatic information with the subspecies of expressive and knowledge‐related information

2. Epistemic information 3. Doxastic information 4. Modal information 5. Data 6. Data‐derived or ‘datalogical’ information with the subspecies of knowledge

based on information systems. 7. Meta‐information, which can be divided into system‐specific meta‐information

and data‐specific meta‐information, or meta‐data. All propositional knowledge is derived from causal connections. Causalities are

retrospective (there has been this causality before) or prospective (this causality can be seen coming). Linked with truth‐values and propositional knowledge there is a justification condition. For instance, organisational partners believe that the system will enable virtual team working between its geographically separated offices. This demonstrates the causal function of knowledge at work with justification.

Pragmatic information (presented in Figure 8) is built into proposition and justification. Belief is regarded as a way of reducing doubt and uncertainty. Knowledge management is married with pragmatism: managers and engineers know what they know and they know what is what. Knowledge management literature often uses pragmatic knowledge as an umbrella covering all sorts of information listed here. But pragmatic information should be analysed more deeply. Basic components of philosophy are applied in the semiotic triangle, for example classes, objects and relations presented later in Figure 9. An example of classes, objects and relations. It is an example of concepts that can be derived from linguistic philosophy for object‐oriented analysis, design and modelling methods of information systems.



57

Figure 8. Information species based on communication

As can be seen, the truth value is not required for all forms of information or knowledge.

‘Belief’ is one object in the class of expressive information. ‘Justified true belief’ is different from belief. The Nonaka & Takeuchi model is too much based on the belief without any evidence, while the Western philosophy accentuates the justification. This model includes both expressive and knowledge‐related information. Both kinds of beliefs also appear in the following Presentation model, because pragmatic information is communication, and different from the presentation of epistemic knowledge, not to mention other beliefs. That significant difference will be adapted in our results.

Figure 9. An example of classes, objects and relations

Qualitative interpretation Pragmatic information

communication

ExpressiveTruth value

not necessary

-assumptions, moods, intuitions, beliefs

-absolute values, norms- questions, orders,

exclamations, requests…

Knowledge-relatedTruth value necessary

-singular,- general,

- explanatory,-instrumental,- evaluative

Value-novelty-utility

- exchange

Transmission-verbal

- non-verbal

Language user

Expression: 1) Terms

2) Predicates a) singular

b) plural 3) Sentences

Expression

Significatio (meaning, concept)

Referent: 1) Object 2a) Class

2b) Relation 3) Truth-value

Pragmatism

Syntax

Intension

Extension

Semantics



58

Pragmatic information is related to the significance of information to the person receiving it in a particular situation. Pragmatic information does not depend on semantic information. For example, it is more valuable for an engineer to receive local or personal instruction in a particular situation than for him to receive an abstract about the theory of engineering, even though it had received a Nobel Prize.

Most KM systems suffer from information overload lacking utility value. Semantic information exists together with pragmatic information, which is always relevant to its social context: person, group, class, ‘tribe’ or community9, organisation, society or culture. The pragmatic meaning of information is most basically explained in the theory of semiotics, described already. However, semiotics does not concentrate on the significance or probability of information. Linguistics, psychology, philosophy and many social sciences have their own theories about pragmatic information. In economics, especially, information is a resource. The semantic interpretation with some measurable factors could be the approach for information systems of knowledge management.

The qualitative interpretation of pragmatic information has at least three subcategories: communication, presentation and processing. Outside of statistical interpretation, language includes potential, cultural information (see e.g. Wiio, Lyons). For engineering, pragmatic information is the most important category. However, when converting it as data to the information system, it must be noted that pragmatic information includes together with the actual sentence, the state of the surrounding world excluded from the sentence: e.g. when the message or other string is ‘(This component is delivered by) an affiliated company’, the quantity of information depends on the number of alternatives, i.e., other deliverers. Pragmatic information will always lead to qualitative interpretation, which is not included in the data. For example, the definition of an affiliated company depends on the other alternatives: subcontractor, other department, this organisation, outside supplier; who are the objects in the class of subcontractors….

For information systems, one aspect of pragmatic information emerges above others: novelty or newness. It is a crucial component when measuring the value of pragmatic information. Three kinds of novelty can be separated. First, it is the value of pragmatic information: the amount of how meaningful or surprising the information is for the recipient. Second, it is the usefulness, the utility value of information. For most pragmatic information this aspect is clear‐cut: it is the usability of that particular information in one’s own work. Third, for information sharing, it is the exchange value. It is the interpreted value — how much others (co‐workers) respect that information — in proportion to the value. For information service systems, the risk is to deliver irrelevant data that everybody knows. Most of the characteristics of pragmatic information describe the value of information, e.g. relevancy, accuracy, reliability, validity, readability and topicality.

The communication of human knowledge is wider than the exchange of pragmatic information. Expressive information covers assumptions, intuitions, beliefs, moods and non‐linguistic expressions, such as sounds, pictures and artefacts. Expressive information is obvious for practical knowledge. Examples of expressive sentences give some background: ‘this working method is good’, or ‘you must do this next’. Also, questions, exclamations, advice, requests or orders with no truth‐value to expressive information, but knowledge‐related expressive information always has a truth‐value. An expression can be a communicated as verbal, aural, visual, based on sense of touch, or based on artefacts. Presentation is also important for information systems – both philosophically and technically.

9 Community of Practice (CoP) is an informal, self‐organised collaboration of people, who share common practices, interests or aims — within or between organisations. When the CoP proves useful to its members over time, they may formalise its status by adopting a group name and a regular system of interchange through enabling tools. Other types of KM communities include Communities of Interest and Communities of Purpose (CEN 2004: 9).



59

Figure 10. Information species based on presentation

Epistemic information is now the one and only form of platonic knowledge – and it is the only category where ‘justified true belief’ is the right definition. Common knowledge and scientific information are both epistemic. It has a truth‐value, but the knowing of that value is not required. The truth‐value of scientific information is heterogeneous; there are, for example, realist and instrumentalist views of science, as well as relativists and objectivists.

Epistemic information can be classified in many ways. Information theory and philosophy of science give the following categories: Singular information refers to individual units of information, whereas general information remains in a wider context, such as the rules and theories of the natural sciences, or systematic social sciences. Explanations give reasons or causes for state of affairs or events and their occurrences. Instrumental knowledge is information about what kinds of means should be used for achieving aims. For example, engineers use the knowledge of applied sciences. However, engineering knowledge should be rationalised by scientific knowledge. Evaluative information gives the value of information, based on a particular hierarchy, for example, an economic, aesthetic or a moral scale of values. Pragmatic information can be classified in the same way. As mentioned, pragmatic information is a repetition of epistemic, doxastic and modal information species in a communication context.

Truth‐value and evidence are requirements for epistemic information. When there is a truth‐value, but no evidence, we are talking about doxastic information. Doxastic pertains to belief – and alternatively also to states sufficiently like beliefs (thoughts, judgments, opinions, desires, wishes). We may know that something is so, and we may also believe that something is so. Believing may be based on facts, or opinion, or both, and may be true, or erroneous, or both. However, this is one critical distinction between information and knowledge: Knowledge in practice entails some degree of personal belief, while information does not. The limit is fuzzier, when it comes to memory.

Something may be true but we may not believe it, and we may believe something that is not true. Beliefs, suppositions and hypotheses, have not necessarily got a truth‐value. This kind of expressive information is multiple and it is named as modal information. For example, deontic modalities reflect a relation to the phenomenon as subject (obligatory or prohibitive relation). The modal information category has absolute values or norms, commands, questions. In information systems these can be represented in many ways – not only by sentences.

The most obvious information in the systems is data and data‐derived or ‘datalogical’ information, which has been quite separate from truth‐values. So far, information systems have been the main reason for this category, presented in Figure 11. Information species based on processing. We



60

talk about information management, no matter what kind of data we are processing. Raw data, which are processed, are converted to data‐derived information. Knowledge in systems can be one category of data‐derived information, and it is relevant to applications studied in such research. Knowledge is traditionally divided into proceduralistic and declarative knowledge. In an information system, procedural information is a series of instructions following a specific algorithm or heuristics, whereas declarative information is a passive, non‐algorithmic description of the world in a symbolic and explicit format. All data is finally numerical (with the current architecture, binary).

Figure 11. Information species based on processing

The information species based on information processing lack infrastructure. The history of artificial intelligence shows that binary‐based processors can not emulate human cognition. On the other hand, why should they? The heuristic data‐logical species here can refer to any knowledge derived from data, not classical artificial intelligence systems.

Knowledge technology was a subpart in this research, but artificial intelligence systems were not evaluated – neither the 'old‐fashioned' artificial intelligence with its symbolic presentation of knowledge nor the technology of modern artificial intelligence with neural networks and other subsymbolic numeric processing. The applications that use those technologies were studied from literature, and the case studies covered more practical technologies, including knowledge bases.

In artificial intelligence, knowledge bases are generated for consumption by so‐called knowledge‐based expert systems, where the computer uses rule inference to answer user questions. Although knowledge acquisition for computer inference is still important, most recent knowledge management developments make knowledge available for direct human consumption or develop software that processes that knowledge. This research has concentrated more on the human side of the systems instead of technology. It has focused on how organisations can utilise the systems.

Cognitive paradigm of knowledge management accepts information systems as they are. The economic value of knowledge and continuous effort to derive benefits from information management both contribute to the cognitive paradigm (Swan and Newell 2000). However, the paradigm is deeply embedded in positivistic science as the tool for understanding a mechanical universe driven by single cause‐effect relationships (Skolimowksi 1994). This research claims that information systems can be seen through the constructivist worldview, even though they are based on positivist paradigm.

Qualitative interpretation Processing

Metadata Data-logicalData-derived

AlgorithmicNumeric

HeuristicSymbolic

Knowledge System

Procedural,declarative

DataNot processed

Media Data

Systems

organisaspecific iwhile dapragmat

informatsome pamost critforthcom

the samepatternsoccurs. Tboth for informatknowledpractical

Pr2.4

practicalin practithe evercommuncompletbuilt on holds thThe tool

are comknowledsectionsother, al

knowleddevelopelike are gare not tfall into t

describeeasily shthe procknowled

theoretiput to pr

between inf

Meta‐ational systeminformation.ata‐specific intic meta‐info

This intion containsart be dealt wtical categorming chapter

The he procedures leads to theThe notion omemory antion systemsdge. The knowl examples; p

ractical know

Even tl knowledge ce, and thusryday work onities of pracely unrecognthe concept at one can ns and techno

Technmonly considdge managem: The creatiolthough they

Practidge managemed to create generally prethe main subtwo categor

Conkled formal knohared in traincess of creatidge.

To obtcal and practroductive us

formation an

‐data and ‐inm strategy. I System infonformation iormation abonformation cs many subswith informary for all infor. uman memo. People note interpretatof meaning, od informatios, such as thewledge definplatonic know

wledge can n

though the pconcept app can not be cof organisatioctice have a vnised to largof knowledgot separate ology are stilnology and todered as enament – compon of knowley are often tical knowledgment. Knowla competitivesented as thbjects of this ies: cultural in divided knowledge as tning courses.ing formal kn

tain the infotical knowlee. The enabl

nd knowledg

nformation mt can be furtormation is as solely dataout the skillsclassificationpecies not mtion systemsrmation spe

ory is based oice patterns tion of new inor the makinon technologe question ofnition ‘justifiwledge is on

not be passe

prior mentiopears in knowcompletely igons. This canvariety of relely institutioge defined bknowledge fll more practools, togetheablers of knopared with pdge and the ed togetherge is used asedge is seenve advantagehe consequestudy. Thereand technicanowledge (19that which is. Informal knnowledge. H

ormation reqdge are needlers in the fo

ge

61

managementther divided bout the mea about data.of a person, is not comp

mentioned hes, or engineecies in pract

on categorisain the environformation ig sense out ogies. Many isf belief and ued true belienly one speci

ed by

ned categoriwledge managnored. Mosn be the worklations to theonalised (Way Heron (199from practicetical. er with the cowledge manractical inforsharing of kin practice. s a common as a ‘corpore. 'Business ences. Althoue are obstacal (see e.g. C996) using ths found in bonowledge waere knowled

uired, and toded. Thus knollowing pictu

Part I – Th

t is nowadayinto system‐edia where th. Important f, such as drawplete, but releere: e.g. memering, which ice is memo

ation. All infonment, andin the contexof one's infossues in humunderstandinef’ is recursives.

isation will bagement litest practical ink of the come organisatiogner 2000). 96) and None. This is just

ulture and lenagement strrmation manknowledge. T

concept in mrate asset’, wresults', 'orgugh they motles to an effeConklin 1999)he terms formooks, manualas described dge is mainly

o assess the nowledge canure describe

e model buil

s a significan‐specific methe data is stofor knowledgwing or drivievant to engmory, sensesis the case sry. This will b

ormation catd when the pxt of previouormation, is aman knowledgng, and explive, and could

be tested in cerature — annformation omunity of pron they are wThe practicaaka and Taket one side of

eadership of rategy (see Fnagement– cThey can be d

many busineswhich can beganisational etivate rationective organ). mal and infos and documas the know practical, su

value of thatn be conceiv this kind of

lding tools

nt componenta‐informatioored and retge manageming. gineering. Qus, and sketchubject. Howebe presented

tegorisation perception ofus knowledgean importantge are ignoreicit versus tad be describe

case studies,nd, as the naor knowledgeractice. Modworking in, fral knowledgeeuchi (1995)f practical kn

those organFigure 12). Ccan be dividedistinguished

ss strategies managed aneffectivenessal managemisational me

ormal knowlements, and wwledge that isuch as engine

t informatioved of as infoutility value.

nt of on and data‐trieved, ment is also

ualitative hing could in ever, the d in its own

is based on f these e, meaning t aspect ed in cit ed with

the me implies, e is used in els of rom e model is ) which owledge.

nisations, Current ed into two d from each

for nd s' and the

ment, they mory, that

edge. He which can be s applied in eering

n, both ormation .

‐



62

The way we interpret information in practice is based on memories. Memories,

knowledge from the past for the future, include semantic (general and theoretical) knowledge as well as episodic (context‐specific) knowledge (Ackerman & Stein 1996a). Knowledge is defined as a collection of concrete experiences, or a set of abstract conceptualisations. Concrete experiences consist of stories, feelings, data and opinions about what has been observed. The abstract models can be scientific; they contain laws, theorems, and procedures accepted as valid knowledge; or judgements; they can contain workable knowledge in the form of policy rules, probabilities, and heuristics (Wijnhoven 1998, p. 30.). However, there is a distinction between organisational knowledge and experience, and it is analogous to the declarative and production memory in Anderson's ACT model (see Anderson 1983). To simplify, organisational knowledge can be compared to declarative memory in that it represents ‘book knowledge’ that can be explicitly recorded and is often stored in organisational artefacts. Organisational experience representing tacit knowledge can be compared to production memory in that it is the result of understanding, learning and performance: it is a ‘skill’ acquired with time (Burstein et al. 1998). Implicit knowledge and memories will be redefined in the results.

Common interpretation of knowledge requires subjective validity over a time period. ‘Personal knowledge’ is knowledge found in a tacit form by an individual. When someone extracts a theoretical counterpart of this knowledge, this part may be found in the same defined and articulated form by several individuals, and can not be characterised as ‘personal’ anymore. However, personal knowledge may include parts of such theoretical knowledge together with tacit knowledge on how to use and evaluate it. This is the foundation for Polanyi’s term ‘practical knowledge’. Practical knowledge depends upon rules and the personal knowledge concerning their use and limitations. Parts of practical knowledge are often articulated for the sake of communication, reflection and creation of new knowledge. Polanyi claims that reflection may elevate tacit knowledge from its state of tool function for action to a state of refinement of the knowledge itself. Thus ‘knowledge for reflection’, as a sort of meta‐knowledge, is different from practical knowledge. ‘Knowledge for reflection’ is a kind of knowledge aimed at theory building on the phenomenon only known in tacit knowledge. Whether the reflection itself is practical is another question.

Practical knowledge is categorised as ‘skill’ or ‘ability’, ‘know‐how10’ and ‘competence’ depending on the feedback. Competence corresponds well with Schön’s (1983) understanding of professionals at work, with their 'reflection‐in‐action'. Professional competence allows the individual to give feedback to the social system by changing its rules. The professional collective may thus learn from its members. With the support of Schön, the term ‘professional knowledge’ is used here to cover competence. Professional action is then understood as requiring the competence level of knowledge.

The dimension consisting of ‘skill’, ‘know‐how’ and ‘competence’ seems to have an increasing level of advancement in common with Dreyfus & Dreyfus’ five stages of skill acquisition (1986, p. 50). The different levels are labelled ‘novice’, ‘advanced beginner’, ‘competent’, ‘proficient’ 10 Know‐how is nowadays in some publications translated into Finnish as taitotieto (~knowledge of skills) instead of tietotaito, as suggested already in the 1980s by Terho Itkonen.

Figure 12. Example from common knowledge management strategy enablers



63

and ‘expert’. However this model characterises the process or routines of professional action rather than it describes the fundamental conception of the specific property knowledge as perceived by practitioners or scholars. The five stages thus may be understood as levels of ability to practice ‘competence’ or ‘know‐how with reflective feedback’.

Another dimension that cuts across these three types of practical knowledge is characterised by the type of rules upon which actions are based. Polanyi points at the difference between changes of the rules of operation and changes of the operation conditions. When the rules are independent from contingencies, the knowledge may be applied in a highly standardised way. On the other hand, when the rules are subject to conditional changes, it is not possible to centralise the way the rule is applied. In the first case the skill, know‐how or competence may be found within a static set of conditions, altered only once in a while by meta‐competence of reflective knowledge. In the second case, dealing with unstable contexts in system development would require specific methods and technical properties that require a more decentralised rule structure with fewer possibilities to build a standardised rule set.

In the cases of this research, it is important to be able to identify the different categories of knowledge – no matter how many sets of categories there are. This is a prerequisite when putting together information systems to better support knowledge use and maintenance. Moving from theory to practice, knowledge becomes more ‘foundationless’, partial, constructed and pragmatic. Brown & Duguid (2000) give practical examples of the distinction between knowledge and information in their book The social life of information. Abridged:

First, knowledge usually entails a person. That is, where people treat information as independent and more‐or‐less self‐sufficient, they seem more inclined to associate knowledge with someone. Just a simple linguistic example: in general people ask ‘where is that information?’ but it sounds weird to ask ‘where is that knowledge?’ as if knowledge normally lay around waiting to be picked up. It seems more reasonable to ask ‘who knows what’.

Second, given this personal attachment, knowledge appears harder to detach than information. People treat information as a self‐contained substance. It is something that people pick up, possess, pass around, put in a database, lose, find, write down, accumulate, count, compare, and so forth. Knowledge, by contrast, does not take as kindly to ideas of shipping, receiving, and quantification. It is hard to pick up and hard to transfer. You might expect, for example, someone to send you or point you to the information they have, but not to the knowledge they have.

Third, one reason knowledge may be so hard to give and receive is that knowledge seems to require more by way of assimilation. Knowledge is something we digest rather than merely hold. It entails the person's understanding and some degree of commitment. While one person often has conflicting information, he or she will not usually have conflicting knowledge. And while it seems quite reasonable to say ‘I’ve got the information, but I don’t understand it’ it sounds less reasonable to say ‘I know, but I don’t understand’ or ‘I have the knowledge, but I can’t see what it means.’

Alongside these quoted ideas, Brown and Duguid make a distinction between know‐how and know‐what: ‘The organisational knowledge that constitutes ‘core‐competency’ is more than ‘know‐what’ explicit knowledge which may be shared by several. A core competency requires the more elusive ‘know‐how’ – the particular ability to put know‐what into practice.’ (Brown & Duguid 2000, p.91).

Brown and Duguid are stuck on English words. One point they are missing is the object of the two verbs: to know vs. to inform. There are linguistic differences in most languages – in Finnish there is not even a verb derived from the noun ‘tietämys’ (mostly translated as knowledge), but all kinds of ambiguous meanings for ‘tieto’ (knowledge, information, data…).

Practical knowledge should also be drawn from constructivist history: from Kuhns (1970) analysis of the history of science, Habermas’s (1972) critical theory, Foucauldian (1980) power and

Systems

knowledPracticalcontext –thought practicalexperien

Ta2.5

2) are ata strategrequiressuch as ‘actual quargumenawarenedoes not

tacit knoincludesmight be

personaknowledembeddtherefora knowle

well knoknowled

this cate

Table 7. T

KNOW

DON’T

modern as such, systems,knowled

idea thatconscioumention

between inf

dge discoursel knowledge,– based on msystem shoul knowledge nce. How clo

acit and expl

The pttempts to thgic overview s questions a‘What do weuestion is ‘Wnt that ‘We cess of things t exist.

Changowledge con elements the skills, intuit

Explicl, subjective dge may be eed in the mire the challenedge manage

Ikujiroown. Not onlydge gaps as w

Table

egorisation w

Types and Ga

W

T KNOW

Polanyinterpretatiowhile this re, it is necessadge.

Althout certain cogusness goes bned, Polanyi h

formation an

e, Derridian (, as well as smost interpreuld govern bis the highesse to each o

licit knowled

hilosophical hink more deto build a nend reflectione know that wWhat do we kcan know mowe may not

ges in dynamcept covers hat are signiftions or visioit knowledgeand difficulteasily codifiend and depenge is to makement systeo Nonaka andy individualswell (see .

7. Types andwill be renew

ps of Knowle

HAVE Explicit KnowKnowledge yTacit KnowleKnowledge y

yi presentedon of tacit knesearch focuary to step b

ugh the exprgnitive proceback at leasthas founded

nd knowledg

(1976; 1978)cientific trutetations. Theelief or invesst form of knther is pragm

dge are both

models preseeply about hew model, an in organisawe don’t knoknow that weore than we be able to t

mic knowledgnon‐articulatficant for chaons evolving e may be expt to transmit d into a knowendent on thke appropriam (see e.g. Sd Hirotaka Ts, but organis

d Gaps of Knwed.

dge

wledge you know yoedge you don’t kn

d the conceptnowledge coses on practback a little to

ression tacit esses and beht as far as He the modern

ge

64

) de‐constructh, is merely e continual cstigation. As nowledge, asmatic commu

h dynamic an

sented in thehow an organd create a vational practiow about oue do not knocan tell and ell’. A sharp

ge are often ted informatanges and induring one opressed and – or sometimwledge manhe context. Pate tacit knowSwan 2001). akeuchi havesations have

owledge). A

ou have

ow you have

t before Nononcept. Sinceical knowledo dissect the

knowledge whaviours are elmholtz's won understand

Part I – Th

ction and Lyothe construcchange in reaa culminatios it remains ounication an

nd construct

e Informationnisation thinvaluable framice. Extendedr collaboratoow?’ This mowe can tell ndivision betw

rapid, radication. Hiddennovations inor more perscommunicatmes even toagement syseople are crewledge expli e made the t these two ty

s implicit kno

DON’T HKnowledKnowled

e UnknowKnowled

naka and Take he has beendge that can e origins and

was introducundergirdedork in the 19ding of tacit k

e model buil

otard’s post‐ction of langality shows ton, Heron (1open to the ld practical k

ted

n and knowlnks and acts.mework of ud organisatioors?’ (See e.gdel is built onothing withween tacit a

al, discontinu but existing organisatioson's work lifted, while ta recognise. Wstem, tacit kneators and uicit and poss

terms tacit aypes of know

owledge cou

HAVE dge Gaps dge you knowwn Gaps dge you don’t

keuchi, and in primarily inbe explicatemeanings of

ced by Michad by operatio9th century (Rknowledge.

lding tools

‐modern conuage in a lochat no single996) asserts lenses offereknowledge w

edge species Those modenderstandinon may poseg. Murray 20n Polanyi’s (out relying und explicit kn

uous and nong tacit knowlens. Those elefe. acit knowledgWhile some enowledge is users of knowible to be tra

and explicit kwledge. The

uld become a

w you don’t h

t know you d

s an originatnterested in d to informaf the tacit as

ael Polanyi (1ons inaccessReber 1995:

nditions. calised e a‐priori that ed by new will be seen.

s (Chapter els provide g, that e questions 000.) The 1966, p. 16)upon our nowledge

n‐linear. Theedge often ements

ge is explicit strongly wledge and ansferred to

knowledge re are

articulated

have

don’t have

tor of the knowledge ation spect of

1958), the ible to 15). As



65

Polanyi regards knowledge as a process in which the individual and the culture interact. Personal knowledge here plays the role of linking the individual to the culture. In this context, Polanyi wants to focus the tacit dimension of knowledge on the place where intersubjectivity or tradition merges with subjectivity. Polanyi’s theory shows that the totality of knowledge is impossible to fit into language. The dimension that makes this impossible is tacit knowledge — the experience and evaluation upon which an individual is dependent, in addition to eventually explicitly formulated theoretical knowledge, on her or his action. The theoretical dimension may not exist without its tacit counterpart, but the contrary is normal.

Polanyi does not share Wittgenstein’s idea about types of knowledge that are principally impossible to articulate. For him it is a question of the function of knowledge. Tacit knowledge is applied without being articulated. This implies that it is not explicitly in focus for communication in certain activities. In another context, the tacit dimension of knowledge might be articulated. When practical work is done in a group of experienced colleagues, most activities are not spoken of. What each worker does is result of experience and earlier division of tasks. However, when an individual worker wants to teach an inquiring newcomer how to do the work, he will articulate (some of the) earlier non‐articulated knowledge. This puts forward a different perspective on knowledge, which makes it possible to analyse actual practice or prescribed practice as it happens or as it is described in a text. What knowledge dimension is articulated and what is tacit, thus points to the focus of the practice or the description.

As mentioned, Nonaka and Takeuchi (1995) have made the accessibility of knowledge popular by plain division of those two categories, tacit and explicit, in their famous book The Knowledge‐Creating Company ‐ How Japanese Companies Create the Dynamics of Innovation. The knowledge types are also defined as subjective and objective. From a philosophical point of view, as Locke and Leibnitz have defined it, objective elements of knowledge have impact on process, and subjective knowledge is where interactions and sense‐making (modes of inquiry as Hegel or Kant have defined it) between people occur (Malhotra 2000).

Explicit knowledge can be expressed in language and so it can be processed, transmitted and stored relatively easily. Implicit knowledge can be somewhat articulated. Tacit knowledge, on the other hand, is personal and context‐specific and therefore hard to formalise and communicate. This type of knowledge is deeply rooted in action, tools and procedures. Subjective insight and beliefs also fall into this category. According to Nonaka and Takeuchi (1995), tacit knowledge and explicit knowledge are not totally separate, but mutually complementary entities. They interact with and interchange into each other in the creative activities of human beings.

In this research, it is not information as tacit or explicit that is in focus, but rather the use of the wide range of knowledge and its management by information systems, to obtain organisational operations and extended co‐operation. Thus, all experience‐based knowledge, and competence of evaluation and quality assessment that typically have both explicit and tacit aspects are in focus. In any case, an interpretation that tacit knowledge is applied without being articulated is too strict. There are many kinds of tacit knowledge, and the borderline with explicit always varies.

The function of knowledge depends on knowledge storing and transfer. There is a big difference between transferring explicit and tacit knowledge. Tacit knowledge can not be transferred by extracting it from people’s heads and putting it on paper or into storage. Moving the people with the knowledge around can transfer tacit knowledge. The reason is that tacit knowledge is not only the facts but also the relationships among the facts – that is, how people might combine certain facts to deal with a specific situation. For example, a briefly outlined best practice combined with a mechanism for the transferee and the transferor to have face‐to‐face conversations has proved much more effective at tacit knowledge transfer than sending off documentation (See e.g. Dixon 2000).

Tacit knowledge, as well as explicit, has subcategories. Social information (along with physical and other information) can be split into individual and organisational (group) information. Individual tacit knowledge consists of language, identity, membership practice and work practice. Tacit organisational or group knowledge has three practices: work, identity and membership (Linde 2001, pp. 160‐162). These tacit subcategories are quite self‐evident, and will not be discussed here, because this research concentrates more on information systems. As a result, however, this research will show that

Systems

the tacitterms of

informatmake thstatic. Inbrowsingtalking t153.) Tac

Se2.6

knowledknowledHoweve

collectiofilters docontext.

Informatinterestslooking finterestsquestion

problembasis of tInformatCommon

two recenetworkrepresenmovemeinterpretDavenpobeneficia

differentorganisarelevant

Table 8. T

AbstracConcret

between inf

t vs. explicit df knowledge.

The tation searchineir informatnformation ng a documeno colleaguescit, implicit a

earch and re

In rathdge managemdge‐bases anr, the proble

It is dion, the resulto not work e

The cotion systemss or informatfor by automs. (Lueg 2001n.

In pram. The relevathe documetion becomen search eng

Howeent approachked users. Contation of coent from oneted as a partort & Prusakal especially

Also int ways. As seational memot together wi

Types of Know

ct Techte Event

formation an

dualism is no. acit knowledng. Informatiion needs exeeds are chant collection,s. There is a gand meta‐inf

etrieval in the

her well‐defiment performd other expeem arises whifficult to dests are often iven though a

ontext of infos for managetion needs. Smatically mat1: 152 ‐ 153)

ctice, informnce of a docnt. The releves relevant wgines and socever, active ches that enabollaborative ontent. Sociae item to anoticular kind o1997; Nardi in collaboran the search earching is baory categorieith the mean

wledge in Org

Descriptnical, Scientifits, People, Inp

nd knowledg

ot enough. M

ge relationshion needs sexplicit. Furtheanging: they , thinking abgap betweenformation wi

e context of

ined situatiom well and pert systems, hen such systscribe user inrrelevant. Sea correct des

ormation is aement of knoSuch profilestching user in. Should the

mation is typiument to anvance depenwhen it has bcial networksollaborative ble the indivfiltering reliel navigation other influenof social navi& O´Day 199tion by extraand retrievaased on memes are presenning of the ac

ganisational M

tive ic knowledgeputs, Outputs

ge

66

Moreover, th

hip and conteem to be at ermore, infoare differenout informatn actual userill be redefin

f information

ns representprovide valuainformationtems are connformation nearch enginescription of d

accurate alsoowledge requs enable to pnterests withprofiles be a

ically neededn informationds on the siteen shared as should not filtering (Ma

vidual to benes on meta‐duses data genced by the aigation. Both99). Collaboranet and intral context knmory, and mnted here (Sctual objects

Memory

PoliciRu

Part I – Th

e tacit and e

text must be least partly tormation neet while searction found, ar interests anned.

n species and

tation‐basedable support retrieval sysnsidered conneeds. Whenes return largdocuments i

o for systemuire a profilerovide a useh documentsavailable to p

d to do somen need can ntuational aspamong all paspy professialtz & Ehrlichefit from thedata about oenerated by activity of othh are a part orative filterinranet. nowledge typemory is basee Ackermans.

Prescripes, Values, Etules, Norms, R

e model buil

explicit are bo

noted also wtacit. Users aeds are highlching for a spassociating, rnd their repre

d meaning

d approachesto their userstems and intext‐indepenn retrieving dge quantitiess supplied. T

atic searchine which formr with the ins potentially public search

ething, for inot be determpects, and is arties involveional activitieh 1995) and e informationobjects and islarge numbehers. Collaboof informationg and social

pes can be clsed on classifn & Stein 199

ptive hics, StrategieRoles, Tasks

lding tools

oth construc

when talkingare often uny dynamic rapecific documrememberinesentation. (

s to informatrs. Some exanformation findent generadocuments fs of bulk. InfoThe reason is

ng and retriemally describenformation hsatisfying thh engines, is

stance, to somined merelsocially deteed (Harper 19es. social navigan‐seeking acs not based oers of users aorative filterion ecology (snavigation c

assified in mfication, the 96a). These w

es

cted in

g about able to ather than ment, g, and even (Lueg 2001:

tion and amples are iltering. al solutions.rom a ormation s the

val. es the user's e or she is hese a critical

olve a y on the ermined. 999).

ation are ctivities of on and denotes ing can be see e.g. could be

many

will be

Systems

data typtypes an

experts unetworkendogencategoriillusion ocloser to

U2.7

considerbetweencomponunderstainteractithere is

between inf

The abes exist on and formats. D

1. Me

Data oexampsent tmail mfor co

2. Stru

Traditmost parts

3. Sem

Documdecaddistribwell a

4. Uns

The ussegmetrainin

5. Tem

Organstructstruct

Even t

use categoriking and collanous part of sation of infoof all‐embraco communica

nderstandin

Enginered as humann informationents in the sanding comeion or informpreliminary

formation an

bstraction lea more concrData types ca

ta‐data

on the assocple, e‐mail mo announce messages to ordinating in

uctured data

tional file andorganisationalong with th

mi‐structured

ment managdes planned abute, store as transaction

structured da

se of unstrucent in a docung materials

mporal data

nisational wotured (changtured (chang

though this csation like aaboration. It the organisaormation specing ‘knowleation of cate

ng retrieved i

eering or anyn understandn and underssoftware indues from a commation and counderstandi

nd knowledg

evel is relativrete level, anan be classifi

iations amonmessages whthe results othe project, nterdepende

a (records)

d record typnal processeshe knowledg

d data (docu

ement systea paperless ond process pnal data are

ata (audio an

ctured data iument, strea.

ork is often tres to the listes to a docu

classificationbove in the cis not techn

ational culturecies and thedge sharing’gorised cont

information

y other KM sding, as descstanding maustry, has nembination ofonversation ng and hidde

ge

67

e, and classind the searched in the fol

ng work activich are assocof a vote. Theworkgroup aent activities

e of data is ns. For exampge associated

ments)

ems are analyoffice, but arpaper‐based an integral p

nd video)

in organisatiming video o

ransient andt of parts neement) or un

n is just an excontext of seical in the firre, yet basede generic dis’ is criticised tent.

system does cribed in an ay also explaiever lived upf explicit andamong peopen framewor

Part I – Th

fication is ofh engine will lowing mann

vities and thciated with ae meta‐dataand activity (.

now integratple, a user acd with the pa

ysed in the cre still spenddocuments part of an org

ons is growiover an intra

d has a start aeded to manstructured (e

xample, it is earch, knowlrst place. ‘Knd on the prevscipline how in the result

not convey article by Higin why reuse to expectatd tacit knowleple who haverks. Furtherm

e model buil

ften even moneed to accner:

e process ana project and here is the g(i.e. the vote

ted into moscessing a sysarts.

case study. Cding part of t(cf. Skapinkeganisation's

ng. For examanet, or using

and an end. nufacture a nedits made t

interesting, tedge sharingnowledge shaviously descrcommunicatts of this rese

knowledge, ghsmith (200e of, for examions. Based oedge, such ae certain knomore, as the

lding tools

ore technicalommodate t

nd context ofd a workgrouglue that linke). Meta‐data

t organisatiostem might n

ompanies haheir annual ber 1991). Docmemory.

mple, placingg audio to pr

This data manew product)to a video rec

that many reg and other karing’ must bribed critical tion is happeearch. In the

if knowledge01). The diffemple, objectson Polanyi's s documentaowledge. Howcomplexity o

l. Different the various

f work. For up and are ks the e‐a is critical

ons and need a list of

ave for budget to cuments as

a video resent

ay be ), semi‐cording).

elevant kinds of be

ening. The e model it is

e is erence s or work, ation and wever, of the

f



68

knowledge to be transferred increases, the ability of documentation or KM systems alone to convey that knowledge decreases, and the need for interaction with knowledgeable people increases dramatically. In a complex situation, documentation by itself may provide only a fraction of the understanding required.

If we examine a sample group from an engineering research and design (R & D) software project, it is normally a team of programmers and other experts. For an extreme programming team – a small group (typically fewer than 10) collocated – understanding is generated from minimal documentation in the form of intense interaction. When several teams are involved and distributed, and customers are many and varied, story cards or other methods alone will probably not be sufficient. However, the critical question in this latter situation is not what kind of documentation should be generated, but rather how the project manager can ensure adequate interaction among all the participants. A secondary question is how much documentation will be necessary to document the information exchanged in those interactions.

In a mobile internetworking era, when engineering and other projects require bringing together diverse information, diverse skills, and diverse organisational entities, focusing on the issue of understanding — the right mix of documentation and conversation – is critical. However, as understanding and wisdom are generally personal concepts, diverse in every culture, and quoted in secondary references of modern KM literature, they are included in the knowledge category here, and a new layer between information and knowledge will be formulated.



69

3 ORGANISATIONAL MEMORY

In the previous chapter, the knowledge categories and information species were separately presented. Prior to this research, there has not been a strong paradigm based on the fact that all information is stored somewhere in the physical world. The concept, which covers information, as well as the media where the information is stored, is memory. Even though work is no longer fundamentally a physical activity, the memory‐based viewpoint on information is relevant because of the physical existence of information – regardless of whether it is human, computerised or in whatever format.

Information systems are trustworthy in terms of storing and sharing of content. Even though human memory works individually in different ways, organisational memory (OM) combines systems and individual memories. The undeniable benefit is information reuse. Another benefit is that archiving and current information can be distinguished, and concerning new knowledge, communication systems can be used to capture dynamic information. Thus, memory does not have to be overloaded and can be of optimal size. Memory, however, is not used much as the metaphor for all information systems.

Organisational memory has been researched for some years, but a common understanding of the ‘memory’ needs of organisations and users remains speculative and unclear (See Mandviwalla et al. 1998, p. 4). A common division nowadays is between contents of memory and processes of memory. The contents of organisational memory are declarative (data, information and knowledge – system content), procedural (e.g. routines — supported by and mnemonic forms and automation in the system) and systemic (e.g. culture, social structures). Basic processes of the organisational knowledge base include acquisition, retention, maintenance, search and retrieval (Ackerman & Stein 1996a).

Knowledge and information can both be part of organisational memory. Information can consist of data relating to knowledge and information (for instance its location, validity or author), termed knowledge‐information and meta‐information. Alternatively, knowledge about information and knowledge about knowledge can be defined as information knowledge and meta‐knowledge, respectively (Wijnhoven 1998, p. 31). In this research we talk about information systems, whether they deal with knowledge, information, knowledge‐information, or information knowledge.

Organisational memory refers to stored information in an organisation, which can be brought to bear on present decisions. Organisational memory is not centrally stored, but distributed across different retention facilities. Storage is the retention of experiences and information to memory, as it is important for the organisation to retain its experiences and information for future use. Walsh and Ungson (1991) suggested the existence of five storage bins or retention facilities that compose the structure of memory within organisations and one source outside of the organisation. Five structures include individual, culture, transformation, structures, and ecology. In this research, an extended organisation uses one outside source: external archives. According to their organisational memory structure, all bins facilitate organisational memory (Walsh & Ungson; 1991; Cho 1999). Also, the most critical bin in terms of this research – the systems – is used as the seventh category.

Retrieval is the recovery of previous experiences that are stored in memory. It is important for organisations to enhance their ability to retrieve their previous experiences whenever they need them. Huber (1995) suggested computer‐based organisational memory, arguing that it would help automated capturing and sophisticated retrieval of information, and result in complete and precise computer‐resident organisational memories (Cho 1999).

The effects of computers on organisations have been studied for a long time (see e.g. Attewell & Rule 1984) and many articles have been written on organisational memory. The methods for researching organisational learning and memory in practice are usually: case study, simulation, experiment, and interview. Organisational memory systems (OMS) are a slightly different from other information systems.

Systems

an intellorganisa

the disciorganisastructureconsequenable Oempirica

new formgenetic adocumestored it

experimmemoryReview; IntelligeOrganisa

Di3.1

evolutio

3.1.1

planningexistencskills anda strong memorieineffectiresultedorganisaon systeshould nmethods

Organisacapitalismanagem

between inf

Organigent assistaational memo

o Colleo Mino Explo Inteo Acti

Both tpline is baseational memoe of OM; (2) uences of OMOM to evolveal research (S

What ms of hypertalgorithms, ant and knowtems, as disc

How dental? The cy is based maStein & Zwant Systems, ational Comp

isciplines of

Theren? Does the

Evolution or

The fig and manage more fromd knowledgemovement tes. The essenve an undire from organiational knowms, the lattenot be seen as.

Self‐oational memation. This cment: inform

formation an

nisational ment to the useory in indust

ection and simization ofloiting user gration intoive presenta

the contentsed on the defory researchparsing the

M. One way fe as a researcSee Morrisoncomes to sytext, groupwand other cowledge basesussed in thedo these theurrent situatainly on fundss in InformaHypertext Fuputing and E

organisation

are two basdefinition of

design?

rst basic queement. Acco

m evolution te, whose reletoward morence of busineected parts oisational evoledge, experer, design prias the only d

rganising sysory systems orresponds wmation techn

nd knowledg

emory can noer. The followtrial practice

systematic of up‐front kfeedback foo existing wation of rele

s and means finition by W agenda as eOM acquisitfor researchech domain isn & Weiser 1ystems as theware, intelligeomputing tec. This kind of previous chories and tection will be edamental artation Systemunctionality alectronic Com

nal memory

sic questionsf organisatio

estion is to dording to Wehan from deevance is diffe intentionaess re‐engineof organisatioolution. The 1rtise and inteinciple seemiscipline, as

stems might normally unwith Thomasnology can ea

ge

70

ot simply be wing require(Kühn & Abe

organisatioknowledge eor maintenawork environevant inform

of organisatWalsh and Unencompassinion, retentioers to address by theory t1998). e memory bient agents, nchnologies tof retrieval reapter on infochniques fit explored in ticles (like Wams Research)and Organisammerce, tog

s. Is organisatonal memory

efine, whetheick (1979), oesign, becausficult to assel designs of oeering or corons and their1990s also shellect (Wijnhms more relevis the case in

be differentnderscore thes Davenport'asily capture

Part I – Th

a passive infements are cecker 1997,

on of informengineeringance and evnment mation.

tional memongson (1991)ng three sequon, and retriess Walsh andesting throu

n, there are neural netwoo provide releespects the mormation antogether? Arhis chapter.alsh & Ungso and currentational Memgether with v

tional memoy cover the m

her organisatorganisationase organisatiss in advancorganisationre competenr memories ahowed a (re)oven 1998, pvant – but non systems en

t from tradite need for co's demand foe, transform,

e model buil

formation syrucial for thepp. 935 ‐ 936

mation from g volution

ry are includ. Walsh and uential phaseeval processed Ungson's regh hypothes

exciting posorks, advanceevant retrievmeaning relatd knowledgere all of thesThe concepton in Academt research fromory Workshvarious web

ory designedmedia withou

tional memoal memoriesons need a ve. In the 199s and their oncies, for exaand to clean)discovery ofp. 30). Becauot necessarilngineering an

ional informomputer‐assor a hybrid so and distribu

lding tools

ystem, but me success of 6):

various sou

ded in this reUngson defies: (1) assesses; and (3) aesearch agensis generatio

sibilities for ed search tecval access in tionships ame species. se technologt of organisamy of Managom HICSS, IEops, and thepages and se

, or is it a resut the conten

ory evolves w (retentions)variety of inf90s, howeverorganisationaample, was to up the disorf the importause this reseay more natund agile desig

ation systemsisted knowleolution for kute large amo

must act as

urces

esearch, and ined an sing the ssessing the nda and n and

the use of chniques, large data,

mong the

ies still tional gement EEE e Journal of eminars.

sult of nt?

without ) come into ormation, r, there was al o remove rder that ance of arch is moreral. It gn

ms. edge nowledge ounts of



71

highly structured knowledge; but for tacit, hard‐to‐formalise knowledge that must be interpreted in a broader context and combined with other types of information, humans are the ‘recommended tool’ (Davenport 1998). Already year 1890 Alfred Marshall stated that capital consists in a great part of knowledge and organisation, and knowledge is our most powerful capital of production (Marshall 1965, p 115). However, this research does not deal with capitalisation or materialisation of human knowledge.

There are many articles about managing non‐formal knowledge. The authors of Toward a technology for organisational memories (Abecker et al. 1998) have developed and fielded a three‐layered model for processing knowledge and for meeting the growing need for enterprise‐wide knowledge management. Their article shows how their organisational memory serves as an intelligent assistant and deals with both formal and non‐formal knowledge elements in a task‐oriented fashion.

As organisational memory in general, also evolution vs. design distinction covers both systems and their content. Tuomi (2000) argues that Wiener’s description of living systems focuses on regulation mechanisms, whereas Maturana and Varela focus on production mechanisms. His 5‐A model of knowledge creation encapsulates Anticipation, Appropriation, Articulation, Accumulation and Action. Knowledge (as a content) is produced from a breakdown in human adaptation to the world.

The relationships of individual and organisational level cognitive processes and knowledge structures have been discussed. Evolution and design require the existence of goal‐directed collective behaviour. In an organisation there is purposeful, collective behaviour, even if individual participant or subgroups may have inconsistent aims. It is only through shared cognitions or knowledge structures that organisational participants are able to achieve coordinated action. Thus, if there is collective, coordinated behaviour, it can also be expected that there are corresponding cognitive processes and collective knowledge structures (See e.g. Jelinek & Litterer 1994, p. 3 and pp. 14 – 15) enabling the implicit or explicit coordination of this behaviour (See Pirttilä 1997, p. 68). This observation of shared cognition has influenced the interpretation of the case studies.

The idea and the assumption that organisational level cognitive processes and knowledge structures are likely to exist is accepted as a result, as this creates a way of explaining the otherwise inexplicable facts. In various approaches to organisational knowledge structures, collective schemata are considered to be above all a result of social activity, i.e. interaction with other organisational participants. Indeed, organisational level cognitions are many times referred to as ‘social cognitions’ (see e.g. Walsh 1995, p. 305 or von Krogh et al. 1994, p. 59). This aspect has also been noted in studies of the nature of expert work, which argue that expertise is not solely an individual quality, but is formed in social interaction and that a new solution or knowledge is legitimated as expertise in the social community (See Engeström 1992, pp. 3 – 28 or Sveiby 1990, pp. 61 – 62) when a consensus on new knowledge is achieved. Intra‐organisational communication is expected to be the mechanism through which organisational level knowledge structures are formed and developed from individual level schemata.

It is obvious that in order for an individual participant's, or one subgroup's, knowledge to evolve into organisational knowledge it must somehow be ‘shared’, i.e. a sufficient number of organisational members have to have the opportunity to become acquainted with this new knowledge. As von Krogh et al. point out, knowledge at one point in time does not connect with new knowledge at a later point in time, unless there are sufficient ‘knowledge connections’ available, i.e. the potential for individuals to convey messages about their observations with sufficient means of communication (see von Krogh et al. 1994, p. 61). Also, for organisational knowledge and the subsequent collective schema to develop it is necessary that a sufficient number of organisational participants reach some kind of agreement on interpretations of their individual and collective experiences, i.e. achieve a consensus (See Pirttilä 1997, p. 69 ‐ 70). Systems can be designed, but communication based content is evolving, and it can not be just ‘shared’, as claimed in the results.

3.1.2 Organisational memory mechanism and content

There are many definitions of organisational memory and OM information systems. The two major memory design parameters are content (influencing effectiveness) and mechanism



72

(influencing efficiency). However, the two most referenced definitions emphasise content and mechanism differently.

Walsh and Ungson, applying an organisational memory information processing approach, refer to OM ‘in its most basic sense’ as the stored information (from an organisation's history) that can be brought to bear on present decisions. This information is stored as a consequence of implementing decisions to which they refer, by individual recollections, and through shared interpretations… information can be considered as decisional stimuli and responses that are preserved in particular storage bins and that have behavioural consequences retrieved (Walsh and Ungson 1997, p. 181).

In addition to the content of organisational memory, the definition of Walsh and Ungson also refers to organisational memory retention media (bins, in their terms) and processes that handle memory content, mostly for decisional reasons. Memory media (for retention, processing, and transfer) and organisational processes together are the memory means in an organisation (Wijnhoven 1998, p. 31). The bins are introduced in subsection 3.2.2, Bins of organisational memories, while retention is analysed more deeply in subsection 3.2.1, Organisational Memory Processes. Conklin differentiated OM from the collective memory, and defined OM as a by‐product of organisational processes (Burstein et al. 1998).

The other most referenced authors, Stein and Zwass (1995, p. 89), define organisational memory as ‘the means by which knowledge from the past is brought to bear on present activities, thus resulting in higher or lower levels of organisational effectiveness’. However, there are some similarities in these definitions.

Both (Walsh & Ungson, and Stein & Zwass) definitions incorporate three ‘building block’ concepts (Mandviwalla et al. 1998, p. 2):

• sequentiality – the idea that information has a time‐line (an age) so that it can

be distinguished from current and new information • storage and retrieval – including the capture, recording, representation, and

retrieval of information • action – the use of the information to perform a relevant action such as

making a decision In addition, the above (and other) definitions also seek to qualitatively differentiate

information – often as data versus knowledge. If the qualitative differentiation aspect is ignored, it is interesting to contemplate organisational memory system (OMS) design from the perspective of the above three building blocks. From the design viewpoint, ‘sequentiality’ is a property of data; ‘storage and retrieval’ are features, while ‘action’ is a use. A system designer may then interpret an OMS as a database management system (DBMS) that focuses on temporal data and includes a specific context of usage (such as decision making) (Mandviwalla et al. 1998, p. 2).

The data versus knowledge distinction is not all. Walsh and Ungson (1991, pp. 177 – 212) emphasise in their OM definition the distinction between information and memory, as these concepts can be mistakenly interchanged in the context of acquisition and retrieval. The difference between information and memory lies in their temporal qualities and also in their uses in organisations (Walsh and Ungson, 1991, p. 181).

One could further argue that the focus on temporal data is limiting. The term ‘temporal data’ implies the need for a special data type – it is unclear if it is important to explicitly represent information in a sequential or time‐stamped manner – it may be sufficient to simply store it for use later. Similarly, it can be argued that the phrase ‘context of use’ provides little value – to be useful, all applications need a target audience and usage context. All we are left with then is that an OMS is a DBMS application, but that phrase can be turned around: current transactional DBMS are only one of many types of OMS. In the book Data Management — An Organisational Perspective (1995) Watson takes this same approach (Mandviwalla et al. 1998, p. 2).

Traditionally the core underlying technologies needed for an OMS include: 1) a DBMS, 2) a database that can represent more than transactional data, 3) an application than runs on top of the DBMS (Madviwalla et al. 1998, pp. 2 – 3). In the case studies of this research, there are similar levels. For

Systems

exampleapplicatifocuses

organisa

Know‐hoMemory

Relationbetweenthe direcretrievincontext

In this reknowledin under

between inf

e, Lotus Noteion. A systemon temporal

Furtheational memo

1. Ope

ow and knowy.

2. Meskills,

3. Meand u

The flos between mn memory tyction of memng informatioof storage an

esearch the cdge deals witrstanding, wi

Figure 13. R

formation an

es and Documm designer ml data and iner, the OM cories consist

erational me

1.1 ope

1.2 con

w‐why are bo

ta‐memory: assets capab

mory informsing operatio

ow diagram memory typeypes and opemory contenton is not enond context o

contents areth the organiill be classifie

Relations betw

nd knowledg

mentum havmay simply includes a specan be divide of:

emory:

Know‐how:erational act

Know‐why:nceptual and

oth knowledg

Knowledge bilities and in

mation: Knowonal memory

relating to tes, operationerational actit flow (Wijnhough, if it is nof use (during

included in isational meed in subsect

ween memory(Wi

ge

73

e DBMS in thnterpret an Ocific contexted into three

: Knowledge ivities.

: Knowledge background

ge and infor

and informanformation.

wledge and iny and meta‐m

he concepts nal activities vities, and phoven 1998, not understog the retriev

the memorymory contention 3.3.1, U

y types, operaijnhoven 1998

Part I – Th

he backgrouOMS as a datt of usage, su parts. Accor

and informa

and informad understand

mation for o

ation about t

nformation, wmemory.

is introduceand problemproblem percp. 33). As claod correctlyal) do not m

y definition, ants. Knowledser goals and

ational activit8, p. 33)

e model buil

nd, as well atabase manauch as decisiording to Wijn

ation that ca

ation that givding of know

operational a

he value and

which are im

ed in the follom perceptionceptions, witaimed in the. Context is tatch, interpr

as the wholedge maturity,d media.

ties and probl

lding tools

as case III tailgement syston making. nhoven (199

n be applied

ves the theow‐how.

activities in O

d quality of e

mportant for

owing Figurens. It illustrath an arrow se end of Sectthe key hereretation is an

e manageme, that is anot

lem perceptio

lored tem that

8),

d in

retical,

Operational

existing

retrieving

e 13. tes relationssignifying ion 2, e. When n issue.

ent of ther factor

ons

Systems

M3.2

organisaexample

3.2.1 O

can be id(input), 2

acquisitidata warrecords

mechaniexperieninformat

useful if knowledinformatelementsystems)

values (slevels. Thoweverinformat The stru

Table 9. R

RetentionSchema

Scripts

Social Sys

Physical sInformat

organisaschema.concept participaLitterer,

between inf

Means for org

The mational memoes in this sect

Organisation

As witdentified (se2) retention

Thereon means hirehouses anin this resea

Knowisms (Wijnhonces of othertion.

Acquithe purposedge and infortion) must exts; and an ind) may be use

Retensee e.g. Kripphere are alsor, do not alwtion thus rec

ctures of ret

Retention Str

n Mechanism

stems

systems ion systems

The le

ation is adde A schema oor stimuli, itants use to in1994 p. 16).

formation an

ganisational

means for orgory processetion.

nal Memory

th database e e.g. Stein &(storage), 3) are two typiring of peopd data mininrch. ledge and inoven 1998, prs (also called

sition is relate of the searcrmation gap xist that givedex that linkseful (Wijnhovntion is the mpendorff 197o many typesways mean thceived (Stein

tention can b

uctures (base

m IndividuaCognitionschema Personal

‐‐

ProductsFiles, dat

evels of the sd here as ther a knowledgts attributes nterpret the . This seems

nd knowledg

memory

ganisational es on the syst

Processes

managemen& Zwass 199) search (retrpes of acquisple as well asng (Ackerman

formation acp. 33). They ad vicarious le

ted to searchch is known, an organisates a storage ls retrieval keven 1998, p. most importa75a or Dougls of storage hat people w & Zwass 199

be divided as

ed on Ackerm

al level n as a persona

routines

s, ‘icons’ tabases, etc.

structures are new level. ge structure and the relaenvironmento be, howe

ge

74

memory contem level are

nt systems, fo5 or McFaddrieval) and 4ition: humans learning. Ren & Stein 199

cquisition anare 1) learninearning) and

h in the knowwhich meantion has. Alsolocation to keys to retent33). ant and recogas 1987) andmedia. The t

will be motiva95).

s in the follow

an & Stein 19

Groual Group

schemGroup

FormnetwoWorkGroupDecis

e as the caseAs Pirttilä (1is an organis

ationships amnt and generaever, a decide

Part I – Th

nsist of procee mnemonic

our operatioden & Hoffer) maintenann‐based and ecords‐based96a). All con

nd input inclung from direcd 3) learning

wledge and ins that thereo, a memoryknowledge antion locations

gnised featud operates attypes of memated to actua

wing table:

996a)

p level p thinking as ama p roles

al and informorks kspaces pware, Groupsion Support S

e studies of t1997) has notsed, pre‐exismong those aate appropriedly behavio

e model buil

esses and binc functions an

ons for an orgr 1991). Theyce (update).records‐basd means acqtent is includ

ude three kinct experiencevia transacti

information e is meta‐mey directory (tnd informatis (e.g. perso

ure of memot individual, mory retentially use the k

Oa shared Cu

scStPrce

mal social FonePh

p Systems

Dico

this researchted, there arsting knowleattributes thiate behavioourist and inf

lding tools

ns, i.e. mediand described

ganisational y are 1) acqu ed. Human‐buisition of filded on top o

nds of learnie, 2) learningons of know

acquisition pmory contentype of memon classes ans, books, in

ry. It reflectsgroup and oon and disseknowledge a

rganisational ulture as a shachema tandard Operarocedures, ‘riteremonies’ ormal and infoetworks hysical infrastistributed DBsommunication

h, but extendre many defidge system aat organisatiurs (see Jelinformation pr

a. The d with

memory uisition

based les, DBs, of the

ng g from

wledge and

process. It is nt on the ory nd nformation

s the group organisation emination, nd

level ared

ating tuals &

ormal social

ructure s, n networks

ded nitions for a about a ional nek & rocessing‐



75

oriented view. Kiesler and Sproull (1992, p. 554) note that reasoning by analogy and similarity is very common in interpretation, whether normatively appropriate or not.

According to Walsh (1995, p. 282), a schema is always connected to a specific information domain, i.e. the knowledge structure is always connected to particular expertise or knowledge content and is not general in nature. Furthermore, Jelinek & Litterer (1994, p. 17) describe schema as a relatively stable construct, if not completely static, whereas perceived reality, i.e. context, is constantly evolving as events unfold.

Organisational memory retention can serve many objectives besides that of having a location in which to put knowledge and information. One of these objectives is access control by which authorisations are attached with authority comments (i.e. confidential, secret, firm property) and access codes on electronic databases. Another objective is to guarantee that valuable knowledge and expertise can be used in the future. The objective of memory indexing can be to ambiguously allocate responsibilities to organisational memory parts, by stating who the gatekeeper of particular memory locations is. Not allocating these responsibilities seems to be the surest way of losing the value of the knowledge and information involved (see e.g. Hamel & Prahalad 1994). Another problem related to knowledge retention is knowledge maintenance; especially when the knowledge is updated at one site while at another site people still use the old information (Wijnhoven 1998, p. 34).

Search and retrieval in the organisational memory context are included in the decision making and problem‐solving. Past knowledge is searched for and retrieved if 1) the decision maker values past knowledge, 2) the decision maker believes information exists, 3) the decision maker can find desired information and 4) the cost to find information is less than starting from scratch (Ackerman & Stein 1996a).

Retrieval or dispersion of memory may be the most used organisational memory process. In computer‐based knowledge systems, knowledge is traditionally retrieved via automatic processes. This enables the system to make machine inferences by combining several knowledge elements (e.g. production rules) via a defined formal logic stored in the inference engine of the expert system's shell (see e.g. Kerr 1991). The results of this reasoning can be used as direct input for actions of other machines. Mostly however, a knowledge‐based memory system is used only as an assistant to a human expert who must add his or her personal knowledge and insights to form a correct interpretation of the system's results (see e.g. Turban 1995). Often, controlled retrieval on the same subject happens at different locations within an organisation (Wijnhoven 1998, pp. 34 – 35). In the system D3E publishers’ toolkit, interpretation plays a significant role. The article and system example from Simon Buckingham Shum’s (1998) Organisational Memory and Interpretation – Roles for Hypertext Functionality – gives a practical example.

Another important difference between controlled and automated retrieval systems is the lack of opportunities for the latter, with the specific purpose of keeping others away from important knowledge and information (Wijnhoven 1998, p. 35) – by ‘sharing knowledge’. In KM literature it is often assumed that professionals and experts are always willing to ‘share their knowledge’ – and thus somewhat to inflate the value of their knowledge. The skill of ‘collaboration’ has been a requirement for decades (see e.g. Conklin 1996, p. 2). However, it should be questioned whether real experts or other professionals are really satisfied with ‘team work’ or other kinds of collaboration. Teams could often be just a burden for them.

Professional organisations try to improve knowledge sharing by a combination of rewards and punishments (see e.g. Levinthal & March 1993). However, in many professional organisations we see that an individual's interest in salary increases and career opportunities are stronger than in incentive systems. In professional organisations shared memories play a smaller role than shared facilities, (offices, computers, secretaries, marketing, finance etc.) which enable an individual to earn more as an organisation member than as an individual entrepreneur (see e.g. Quinn 1992). Shared knowledge and information facilities are part of the organisational memory infrastructure and can consist of a library, shared journals, topic selection services, rooms for knowledge exchange and so forth (Wijnhoven 1998, p. 35).



76

Filters are a critical question for search and retrieval, but also for the storing. If everything can not be stored, then who decides what to store? What is the mechanism and criteria? How is bias avoided?

Widely held assumptions about data imply that the more memories we store the harder it becomes to locate a specific memory item of interest. Therefore, OM conceptual models will need a built‐in filtering mechanism. As a rule of thumb, the retrieval is not working well, if it takes a longer time to find a document or a message than it takes to read it (Santos 25 April 2008). The retrieval kind of ‘knowledge sharing’ is also more of a sociological than a technical issue.

Maintenance means keeping information relevant, complete and up to date. Also, pointers (indices) to information must be kept up to date. According to Ackerman and Stein, human ‘components’ need maintenance as systems, and they list variables, such as: hiring and firing, socialisation, recurrent patterns of interaction and sanctification. (Ackerman & Stein 1996a.)

As a summary, the traditional processes of organisational memory are combined in the following figure.

Organisational

Knowledge Base R

eten

tion S

earch

Acquisition Retrieval

Maintenance

Knowledge acquisition and maintenance pose a serious challenge to OMs. They are the main

reasons why knowledge‐based systems have so often failed in industrial practice (Abecker et al. 1998). Short term memory may be one answer (Conklin 1999).

3.2.2 Bins of organisational memories

Walsh and Ungson argue that organisational memory is both an individual and organisational level construct. Information and knowledge that is exploitable from the organisation's point of view can be embedded at either level, i.e. within individuals or in systems and artefacts. When organisational memory is discussed, each of these levels must be considered. Walsh and Ungson state that organisational memory exists in five inner retention facilities or ‘storage bins’ as they call them. These five internal and one external bin are (see Walsh and Ungson 1991, pp. 183 – 187):

1. individuals, 2. culture, 3. transformations that occur in the organisation, 4. organisational structures, 5. actual physical structure or workplace ecology of an organisation, 6. external archives

Figure 14. Processes of organisational memory (Ackerman & Stein 1996a)

Systems

retrievedintuitivebehavioutransforof know

O3.3

applied torganisaconstantlong beesystems,availableconstantdecisionprimary

Decisioncapture and creaof the comeeting

Decisionfor the rprototyptesting t

managemthey globusinesson organof decisidecisionorganisamechaniecology,Miller &

technolobeen defat the orinformatis used in

provide resultingjustified,1988). It

between inf

Walshd from thesely. Accordingurs are basemations or cledge (and th

rganisationa

This rethe results inational memotly created nen able to mi, like expert e to support tly and inform support sysdriver for th

The ‘dn Support Sysinformal knoates shared uonversations managemen

Decisin support canetention andpe of a systemhe usability

Nowament. (Docubal, multinats processes. nisational peons (e.g., inds. Recent stuational memoism. Organis but it is the Nilakanta 19

Previoogical meansfined as ‘a frrganisationation systemsnstead of OM

OMS ca means by wg in increased, because tht is comprehe

formation an

h and Ungsone organisatiog to Walsh ad on previouculture. This herefore in t

al memory sy

esearch mainn the case story. The incrnew requiremitigate thosesystems, casorganisationmation systetems could bhese solutiondedicated’ Ostems is consowledge. Quunderstandins that led up nt. ion support hn be groupwd retrieval ofm built arouof the systemdays a very imentation mtional, or locInformation erformance. Sdividual or grudies on orgaory, its retenations retainmemory ret997). ous researchs under the tramework fol level’ (Steins are the coreMIS. can be definwhich knowld levels of efe effectiveneensive and a

nd knowledg

n (1991, pp.onal storage bnd Ungson, ous practices amakes organthis thesis).

ystems

nly studied tudies. Therereased compments for infe effects (seese‐based systnal memory em tools are be now replans, if not morM systems asidered orgaest Map helng in a knowto key decis

has usually bare or designf organisationd readily avm. (See Milleimportant ormanagementcal, are findintechnology Such decisioroup) and spanisational dntion, retrievn memory amtained by the

has been ocerm organisor a range of n & Zwass 19e, and there

ed further asledge from tffectiveness ess constructllows for con

ge

77

189 – 190) sbins either inone exampleand procedunisational me

heoretical oe are commoplexity and tuormation tece first implicatems and (gr(Ackerman &changing, buaced with bure enlighteneare for sometanisational mps to write dledge team. ions and pla

been a typicaned for indivnal memoryvailable off‐ter & Nilakantrganisationat is the seconng that informmanagemenns then havepecial considdecision suppval, and use, mong individe individuals

ccupied withational mempossible info995, p. 95). Bis no need t

s (Stein & Zwhe past is brfor the orgat is central tontingencies.

Part I – Th

state that infn a controllee of automatres that havemories mor

rganisationaon reasons fourbulence of chnology. Adations from Hroup) decisio& Stein 1996ut the subjecusiness intellied decisions?thing else. Fo

memory softwdown the keyUsers end uns. It is more

al subject of viduals. Milley among the the‐shelf softta 1997.) Thel memory synd case studymation technnt decisions te qualities therations muport have broas an effectiduals, culture that is most

the conceptmory informaormation sysBecause this o redefine th

wass 1995, prought to beanisation.’ Ano organisatioIt is consiste

e model buil

formation and manner ortic retrieval oe been sharere appropria

al memory syor all kinds ofthe post‐inddvanced infoHuber 1984).on support sy6a). The situact remains thigence soluti? or example, ware. But it iy issues and p with a ‘mae about discu

organisationer and Nilakaindividuals. Ttware wheneese policies aystem categoy in this resenology is a ptherefore hahat set them st be accordought forth tive support ae, transformat conducive t

tualisation oation systemsstems designresearch conhis concept.

. 95) ‘…a systar on presenn effectiveneonal science ent with prio

lding tools

nd knowledger automaticaoccurs whened and encodte in the ma

ystems (OMSf informationdustrial worldormation tec. Advanced ystems, haveation is evolvhe same. Forions, but wh

Quest Map bs simply a toideas duringap’ that showussion mapp

nal memory santa describeThey proposever possibleare quite staory is documearch.) Organpowerful enave a profounapart from oed to suppothe need to and coordinaations, structto systematis

of informatios (OMIS), whned to supponcentrates oThe abbrevi

tem that funnt activities, tess‐based mo(see e.g. Camr research (s

e can be ally and present ded in nagement

S), and n systems asd have hnology has information e long been ving r example, at is the

by Group ool to g meetings ws a history ping or

systems. e a model se building a e and ble. entation nisations, bebler of their nd impact other types rt these focus on ation ture, and sing (See

n hich has ort memory n systems, ation OMS

nctions to thus odel is meron see e.g.

e



78

Quinn & Rohrbaugh 1983) and agrees with Parsons’ (1959) landmark work in sociology. Effectiveness is also linked to memory functions. (Ackerman & Stein 1996a.) It is possible, however, that effectiveness is only a consequence of OMS and it should not be used exclusively in the definition. And in the context of corporate knowledge, the effectiveness and adaptation are valid constructs only within given goals (Tuomi 2000).

It used to be one risk that OMS realised only a portion of OM, as in the task‐specific approach in the early ‘90s, when the main focus was on groups, projects and topics (Ackerman & Stein 1996a). We must remember, as noted in the beginning of this chapter on organisational memory, that contents and means formulate organisational memory. Similar discipline remains in this study concerning information systems.

Many systems have been investigated. They are all listed in the Appendix 2. Systems can be classified as Stand‐alone vs. Add‐in OM systems, User focused, Retrieval focused, Task focused and Data focused OM systems. The case studies included a variety of applications, off‐the‐shelf systems and development tools, as the preceding research.

3.3.1 User goals and media

Figure 15, Concepts related to organisational memory (adapted from Wijnhoven 1988: 37 et al.) describes effectiveness and efficiency leading to fulfilment of organisational goals based on the contents and means. Lewin & Minton (1986) induct effectiveness from terms like adaptation, goal setting, integration and pattern maintenance. Effectiveness of organisational memory is about the potential of knowledge contents to contribute to organisational effectiveness. In this organisational memory system chapter, the user goals and media are critical factors.

Actual use and shape of organisational memory are highly dependent on user goals distinguished to three organisational levels: individual, intra‐organisational, and inter‐organisational. (Wijnhoven 1998, p. 41.) This study adds extended organisational goals between intra‐and inter‐organisational goals.

At the individual level, it is essential to determine who the user is and what the user wants. Individuals seem to differ on the type of knowledge and information they need and retain, the reasons being the type of job and the learning style of the individual. This results, for instance, in managers who are strongly data‐oriented and interested in abstract models versus managers who are more intuitive and apply tacit knowledge. The consequence is that in the first case, databases and model bases may be valuable memory contents media, whereas in the second case, the memory contents media are human (colleagues, intimates). Access rules will be defined depending on the extent to which the individual wants to share memory. In theory, tacit knowledge is extremely difficult to store. Traditionally, it can be indexed by information about who has knowledge about a certain subject and how they can be contacted — also in new research areas, with few publications and few results. With a catalogue of research topics and people, in principle, everybody can find out ‘who knows what’ about the state‐of‐the‐art of the research area and access new ideas and future options (Wijnhoven 1998, p. 42). Based on case III, however, it is doubted whether this is a working solution.

At the intra‐organisational level, coordination of interests happens in two separate categories: decision and agency. This distinction by Wijnhoven has a large impact on the use of memory contents: optimising decisions and handling conflict (Wijnhoven 1998, p. 42).

In the case of decisions, the knowledge and information for realising efficient transactions can be formalised and become part of the transaction processing system's software. For making the best choice in situations of multiple criteria and alternatives, software of different types can be used to model the situation and then be compared (see e.g. Epstein & Henderon 1987). To overcome agency problems, organisational memory systems can be used to understand the causes of the problems (historic roots), and electronic networks can be used to improve the communications required for understanding each other's opinions and for making bargains (Wijnhoven 1998, p. 42).

Wijnhoven (1998, p. 42.) states that decision information can often be well stored in databases and model bases, but agency‐related activities often contain information and knowledge of a political nature, kept mostly informal and personal. The sharing of this political information is an



79

important aspect of the coalition formation process in political systems, leading to shared anecdotes, stories, myths and development of a joint memory‐base (Walsh & Ungson 1991; Brown & Duguid 1991; Hedberg 1981).

One example studied (Information Systems Cybrarium) was between the intra‐organisational and inter‐organisational levels, and concerned organisational memory system for scientific communities. This example emphasised that memory is not necessarily the only function of such systems: also important are communication and inference. In this context, the term knowledge infrastructure may be more appropriate than organisational memory (Hars 1998). Case III application was somewhat similar meta‐system. In this research, however the communication appeared to be so present, that the result was not knowledge infrastructure but a memory management model and strategy.

The inter‐organisational level is about collaboration between organisations via, for instance, co‐makers and shared risks and resources. The related organisational memory means should then be chosen in relation to the joint project and to what memory content transfers are required. In an extended enterprise level, there is no risk that one partner might take over the technology and core competence of the other partner. Seamless partnerships in both levels require:

a) an overriding goal congruence and strategic compatibility, b) strong mutual trust, based on reliability, ability to co‐operate, and openness, c) an integrated information system that allows continuous real‐time monitoring

of all key variables and d) protected core competencies so that each party can continue to dominate its

own core competency to such an extent that the other party can not effectively bypass it to access the marketplace (See e.g. Quinn 1992).

Organisational memory means in inter‐organisational settings must be organised, but it

has often been a neglected topic, for example, in business intelligence (Tyson 1986; Gilad & Gilad 1986). Here all kinds of traditional design modes can be chosen, but because of the high information and knowledge intensity of these organisations, a variety of information technology organisational design parameters must be considered (see e.g. Galbraight 1996; Lucas 1996). As a result from this research, the whole extended enterprise architecture should begin with information architecture specification.

Wijnhoven (1998, p. 43) mentions three characteristics of memory contents that influence the appropriateness of memory media: 1) knowledge maturity, 2) memory indexability and 3) memory contents transferability. They are also adapted for this research model building.

Knowledge maturity can be defined as the ability to formalise knowledge (elements) without any loss in value. If knowledge can be retained, retrieved, processed, disseminated and maintained in a mechanical way, it has become universal, context‐free and fully described in formal logic. The maturity of technological knowledge can be rated on a scale ranging from 1 to 8 as follows:

1. complete ignorance 2. awareness but implicit knowledge 3. measuring but not yet a technology 4. control of the average 5. control of deviation from the average 6. know‐how, which explains the consequences of manipulations of variables 7. know‐why, containing advanced scientific models, by which processes can be

simulated and controlled 8. complete knowledge The following storage media are most appropriate at the different levels of knowledge

maturity, as presented in Table 10, which provides a set of hypotheses concerning the selection of retention media (See e.g. Bohn 1994, p. 68).

Systems

Table 10.

K

LocNatuSuita

Ease of

opportuunambigknowledmeans thconnectwhen it may helpmanage

memoryas intendreliable wtransfera

Several carrangem

Table 11.

Level of NGeneral

User‐goa

Media

for specicoherenWhereasstructureinterviewassessm

Co3.4

and syst

between inf

. Knowledge m

Knowledge fea

cation of knowure of problemability for autof transfer to an

Note. The se

nities of defiguously retaidge or informhat the indexion to the knis implicit orp in providinment of kno

The thy contents traded. Sometimway. Even wability can st

An ancriteria can bments, as de

. Levels of ana

Needs Analysis

l

Memoific business ce in a mems memory ares describe rwing personsents and not

onclusions fr

Organems, and he

formation an

maturity and

ature

wledge m solving omation nother site Adapted from

econd memoining memorined and retmation searchxes must be nowledge acc secret. In thng the answewledge into hird characteansferabilitymes it is extrwhen these setill be very loalysis of membe used at eaescribed in th

alysis, criteria

s MultilocaadaptatioinformatMemorypersonalarrangemMaturityknowled

ory contents areas, and gory contentsrchitectures relations ams and by asset described f

rom organisa

nisational mence this rese

nd knowledg

effectiveness

1 2 Worker's hTrial and e

NoneLow

m Bohn, 1994,

ory contents ry classes anrieved. Memhes have resdefined in scessed (Wijnhat case, the er. One obviometa‐systemeristic of memy, which concremely difficemantic probow because omory efficienach level, reshe following

a and results f

Cal, instantaneon, memory mtion processiny contents extel, secret, tacit,ment, strategiy, indexability ge

architecturegroup these ts class (a grodescribe relaong memoryessing concrefurther here.

ational mem

emory is a wearch. It is an

ge

80

s of organisat

3 heads error e

, p. 68, Table

characteristd elements i

mory indexabulted in retremantically nhoven 1998index should

ous questionms or not. mory contencerns the proult with certblems of effeof political anncy can be casulting in protable:

for organisati

Criteria ous access, dmaintenance ang capabilitiesent of being s, part of a parc and transfera

es describe ttypes of memoup of memoations amony elements. Tete activities.

mory

ide metaphon area of org

Part I – Th

ional memory

Maturity 4

Written andScientific me

2: Some Effec

tic is memoryin a precise wbility can be mievals that wunambiguou8: 43). Knowld refer to a p for this who

nts, which infobability of stain media toective messand competitiarried out atoposals for se

onal memory

issemination, and memory

shared, rtnership

ability of

he knowledgmory contenory elementsg knowledgeThese memos, but done a

or. However,ganisational t

e model buil

y media

score 5 6

d oral ethod

cts of Knowled

y indexabilityway, so that measured bywere differenus ways and edge is extreperson or orgole thesis wa

fluences medsent messageo transfer meage transfer aive reasons (t general, useeveral social

y means

Efficient iproject po

Social demconstraint

Selection

ge and informnts by optimis), and minime and informory elementsas informatio

, it is connectheory and a

lding tools

7 In DBs or sofTable look

High High

dge Stages

y, which refethey can be y the extent tnt than expecwith unambemely difficuganisation uas whether to

dia appropries reaching temory conteare resolved(Wijnhoven 1er‐goal or m and technic

Results nfrastructure ortfolio

mands and strts

of media

Wijnhove

mation that asing the intemising the exation classess were descrion‐systems n

ted to adminan area withi

8 ftware k‐up

ers to the

to which cted. This iguous ult to index nit who o divide the

ateness, is the receiver ents in a , 1998: 44). edia levels. cal

and IT

rategic

en 1998, p. 44

are requiredernal xternal links. s, memory ibed by needs

nistration n

4



81

information systems design. CSCW systems were the first widely spread systems of organisational memory. This societal issue is a meeting point of organisational theory and systems work (within the social network perspective). There are some significant organisational constraints on the technical design space, as well as technical constraints on the organisational design space. Memory is not just information retrieval. All aspects of organisational memory were studied further, in regard to systems that are evolving human knowledge.

Desirable organisational memory strengthens the identity of the organisation, maintains strategic direction over time. It also helps newcomers and helps the organisation to learn. The most important knowledge of a company should be linked with its strategy, core competencies and business areas (Hamel & Prahalad 1994), and the value of knowledge resources can be assessed simply by their ability to generate strategically important products and services (see Tampoe 1994). The capacity to transmit information over time may be facilitated by the impact of information technology on organisational effectiveness. IT support for organisational memory can be made a principle of systems design (Ackerman & Stein 1996a).

In general, the two dimensions of organisational memory, contents and media, should be separated. All OM definitions refer to the media. From the viewpoint of information processing, OM is also the stored information from an organisation's history which affects present decision. Thus, Stein & Zwass ‘s definition of organisational effectiveness might be too narrow. The business perspective is not a firm basis from which to study OM. Even though the most important knowledge of a company should be linked with its strategy, core competencies and business areas, and the value of knowledge resources can be assessed simply by the ability to generate strategically important products and services, there is much more to explore.

Organisational memory is both an individual and organisational level construct. OMS – even though it is a system – can be studied on both levels as well. Information and knowledge can be retrieved from the bins in a controlled manner, automatically, or intuitively, and on both levels.

OM is partially about organisations' knowledge and information bases, the contents of the memory stores. Operational memory, meta‐memory and information about organisational memory as the main contents can be distinguished. M.J. Earl has presented three management strategies for information technology (Earl 1988, Wijnhoven 1998, p. 37). These top‐down, bottom‐up and inside‐out methods for information strategy formulation can be applied to organisational memory as contents design.

OM is also about memory means, i.e. organisational procedures and management processes (e.g. debriefings) and memory media (e.g. expert systems). In principle, tacitness, indexability and transferability are important indicators of the extent to which memory contents should be managed by different knowledge media.

The analysis of memory contents improves the effectiveness of organisational memory, whereas an analysis of means improves its efficiency. The categorisation of OM systems is evolving. The requirements and design issues vary along with types of knowledge and other variables, but normalisation and meta‐information play an important role in all respects.

The definitions of OMS are: 1) A framework for a range of possible information systems designed to support memory at the organisational level, 2) a system that functions to provide a means by which knowledge from the past is brought to bear on present activities. The latter is conducive to effectiveness. Besides effectiveness, OMS can support the retention of core competencies and learning. There is no learning without memory, and both of them require categorisation. The information species specified in Chapter 2 help to learn and memorise.

One risk used to be that OMS realised only a portion of OM, such as in the task‐specific approach during the early ’90s, when the main focus was on groups, projects and topics. Nowadays the risk is somewhat similar: OMS will concentrates solely on core competencies, organisational effectiveness and learning. Those should be secondary results, not the aims of OMS. Maintenance of strategic directions, help for newcomers and the strengthening of identity are not such directly measurable values.

It is not necessary for organisations to store more on free and automatic access media than they do now. A well‐thought‐out cost‐benefit analysis may help in choosing these types of storage



82

media. The existence of such types of memory retention and dissemination does not always mean that people will be motivated to actually use the knowledge and information thus received; organisation incentives must be built in, these include regular reviewing of output against what the results could have been when applying available memory contents.

On content side, OM system research has been limited to data, information, memory and knowledge. A more ambitious pursuit is to develop a complete methodology for sociotechnical organisational memory system development, i.e. design and evolution. Further research is required on the questions of how organisational memory means evolve in such a way that they will lead to new sociotechnical environments. A sociotechnical example of a meta‐system could be implicit or secret knowledge; a fuzzy ‘index’ could refer to persons or organisation units. Modern ‘knowledge sharing’ will be discussed in the results.

Another question concerns how organisational memory contents and organisational memory means develop in concert. If, for example, tacit knowledge becomes increasingly important in more innovative organisations, how can its contents be managed? Information scientists may feel uncomfortable managing intangible aspects of informal organisations, or over‐value the abilities of information science. One reason for the failure to capture informal knowledge is that Western culture now values results – the output of the work process – far more than the process itself. How can information management support the management of metaphors, analogies, ambiguity, redundancy and the interaction of personal and organisational memories?

3.4.1 Summary and missing links concerning this research

Organisational memory concerns the knowledge base of an organisation and the attendant processes that change and modify the base over time (Stein 1995). Nowadays the role of human knowledge is more emphasised. As a summary of the concepts defined, Figure 15 shows the lines denoting generic‐specific relations; the arrows refer to causal inference lines. Memory contents analysis improves organisational memory effectiveness, whereas memory means analysis improves efficiency.

Figure 15. Concepts related to organisational memory (adapted from Wijnhoven 1988: 37 et al.)



83

The empirical interviews raised the following questions on missing links, which are related to organisational renewal.

• Archiving and deletion of the memories is crucial to avoid information overload and redundancy. Retrieve the relevant (short term) and historical (long term) memories separately 11.

• The most important media of OM bins are information systems in general • When the systems are built, the content is learnt and thus memorised. • The processes require a lifecycle with an end. Some loose ends of OMS can be distinguished. A hybrid solution where a machine

amplifies human knowledge is becoming a self‐evident type of OMS. Still there are orthodox seers at both (man vs. machine) extremes. Also, the following three gaps remain and will be discussed further: 1) vanishing and transmuting of information or knowledge from organisational memory, 2) the naïve assumption that professionals always share their knowledge with no distortion and 3) that the futile new work phase for converting memory content to the systems does not affect learning. Maybe OM could also have disadvantages.

Based on initial empirical evidence, the memory aspect is important for organisational information management and a fundament of a new KM model. This research investigates systems for management of knowledge, in which organisational memory play a critical role. The investigations is directed towards knowledge and systems labelled otherwise (i.e. not as organisational memory systems or content). But the trend of OMS is towards the same direction. Other systems are designed for managing more specified content, but together they formulate the main parts of organisational memory. The process itself is becoming more central. Thus, OMS is becoming a dynamic organisational process rather than merely an artefact such as a repository or data store. Is more emphasis going to be placed on the individual worker instead of organisation? The main point is that OMS is needed for a learning organisation to support the knowledge management of individuals. To become organisational, communication is required. Also explicit (semantic and episodic) vs. implicit (motoric and priming) categorisation will function better if the information species are defined in the context of organisational memories as well, and finalised with the formulation of new concepts. This will be done also in results of this research.

11 In the history, even display systems were based on the idea of short term memory [Conklin 1999]

CONTENTS

Part II ‐ Initial case studies


Part II – Initial case studies

85

INITIAL CASE STUDIES BASED ON GROUNDED THEORY

Cases of this research were somewhat different, but comparable. The research procedures were followed in every case. In grounded theory it is possible to pick and choose from available analytic techniques, but the procedures of making comparisons, asking questions and sampling based on evolving theoretical concepts are essential features of the methodology (Strauss & Corbin 1998: 46).

In this kind of research, where cases are studied, there are two operations above all: asking questions and making comparisons. Asking questions is an analytic device used to open up the line of enquiry and direct theoretical sampling. Making theoretical comparisons is an analytic tool used to stimulate thinking about properties and dimensions of categories (Strauss & Corbin 1998: 73).

One arisen question besides the research question was whether the organisational memory is fundamental or not. The hypothesis is that knowledge must reside in a memory and therefore organisations need a social memory besides individual memory for knowledge management. That is concerning also knowledge ‘creation’ and ‘sharing’.

Yin (1989) stated that quality of a research can be measured with four factors: construct quality, internal validity, external validity and reliability. In general, the only way to guarantee the validity in qualitative research is to describe the research process in detail. The case studies investigated systems and their use even more detailed than printed here.

Systems marketed for knowledge management have always been diverse. They are not a single technology, but a collection of indexing, classifying and information‐retrieval techniques coupled with methodologies trying to achieve results for the user (Binney 2001: 2). The main technologies enable (cf. Purvis, Samabmurthy & Zmud 2001; Wood & Sheina 1999):

• Content and workflow management, which categorises knowledge and directs it to workers who can benefit from it

• Search functionality, allowing users to look for relevant knowledge

• Collaboration to share knowledge.

It has been argued that well‐designed and implemented knowledge management systems give enormous return from leveraging organisation’s knowledge resources (Davenport & Prusak 1998). Some findings are from areas such as customer‐relationship management, research and discovery, business processes, and group collaboration for design (Lawton 2001). This research was dealing mainly with business management and engineering processes, research, and collaboration. Comparison was derived from three cases – however the cases were each including two phases, so there were originally six case studies in this research. Empirical discoveries were made also after the case studies.

Even though Eisenhardt (1989) claimed that four cases are required to make generalisations, it is depending on the nature of the cases. An external dimension of validity by Yin (1989) is transferability, which deals with the problem of knowing whether the study’s findings would have a larger significance in the overall context. This main challenge of the qualitative research was met here, as the cases included so much intensive work, qualitative and quantitative sources, different groups of people, a long time scale, and two phases each. After the cases, ethnographic observation continued.

So there is comparison between independent cases. Moreover, there are in each case comparison between groups. The selection of comparison groups (cases in this research) for discovering model is their theoretical relevance for furthering the development of emerging categories (Glaser & Strauss 1972: 49). These cases and comparison groups were strictly relevant. Construct validity (Yin 1989) is reached with multiple sources of evidence in data collection, which makes findings and conclusions more convincing and accurate.

In making multiple comparisons, the researcher constantly focused on generating and generalising a model, not on the comparison of differences to verify or account for a fact. Generating from differences is not easy to manage with quantitative data, since verification from differences comes



86

very easily with quantitative data. Verifying and accounting for facts by differences are subsumed in the process of generating a model; they are not the product of quantitative research for this purpose (Glaser & Strauss 1971: 211). Therefore most of the data and observations in this research are qualitative.

How the cases were reached

The cases were selected on the logic of case study method (Yin etc.). Actual division between categories of data, information and knowledge is presented in Part I section 45, Research design – the cases and condensed in Table 6. Comparison of cases.

There are some common denominators that can not be ignored in the current world of the information systems. First, relational databases are still the core of KM applications. Relational database management system (RDBMS) was in the background in all three cases. Second, IBM has pushed forward knowledge management with Lotus Notes and Domino since they bought the products, and there were plenty of applications on top of those platforms. Collaboration is considered a crucial part of KM, and many collaboration applications were based on Notes. Third, organisational memory applications have not been a commercial success because of the artificial extra work. Therefore in the focal company and its collaborators, the extended organisational memory was more or less managed with other systems already in use.

This research first found out the systems from the market and the existing new systems used in the focal company. Following are examples about systems that were related to knowledge management, and were available for this study from the beginning. The community of practice application (CoP) of the focal company was used as an example in international seminars (e.g. Curley 21.9.1998).

As the DBs, also the collaboration systems were all the time in the background. In the beginning of this research they were divided into Intranet collaboration modules as follows:

1 Communities of practice (CoP) 1.1 Homepages (and a toolbox to customise them) 1.2 Calendars 1.3 Shared URL links, bookmarks 1.4 Components from team applications and CoP, like

1.4.1 Document Library 1.4.2 Discussion groups

2 Task library Application templates were considered for the customised applications, such as: 1) Team application with

a) Team collaboration and b) Document archive,

2) Web site builders' tools for a) Web pages, b) Event calendars and c) Document library,

3) News magazines (Internal magazines), 4) Document library (simpler than in the website builders' toolbox), 5) Discussion (browser enabled discussion) and 6) Internal releases.



87

These modules were from intranet collaboration side of knowledge management. All of these were already then close to current social networking applications. The human meta‐information systems, such as Skill management system and Core competence system were examined in the beginning. These applications were omitted from the deeper study, because of the plan to concentrate on actual content instead of human resources.

Other company wide systems related to knowledge management were studied in the background, such as 1) Project Management System, Software Solutions Projects, 2) Learning tools, divided into a) Learning centre (self‐study, collaborative learning, distributed classroom) and b) Distance learning pilots: schedule, media centre, course room, profiles and assessments and finally 3) Future journals. Other, more engineering specific candidate applications were 1) Action databases, 2) Requirement management systems, 3) Notification tasks and 4) Problem management applications.

Later it appeared that two major players in these system categories were IBM Lotus Notes and Oracle based applications, as the first was used as an application platform for intranet and the second was in the backbone of other applications. Quite rare applications were actually more than specific information management, and therefore the whole field of knowledge management systems was studied from the market in the beginning of case I. All kinds of knowledge management systems were constantly studied in large scale when working in the Research centre of the focal company. On corporate level, the applications were in global scale – not only in Finland. For example, the search engines of the focal company were managing more URLs than search engines of the Internet operators in Finland.

It is worth to remind that all systems marketed as such, are not necessarily knowledge management from the business point of view. To keep theoretical level high in the study, the solutions were described instead of technical details. A golden mean between the media (tools and systems) and the message (content; data, information, knowledge) was found.

Case II described documentation management ‐ a research project in the context of knowledge‐related systems, based on case I. This case study is a perfect example about managing explicit knowledge (written documents and other recorded information). Case III was dedicated KM system.

The Focal Company as an Extended Organisation

In this case study the actor is business enterprise of closely collaborating firms. One description of a firm is: ‘Organisations that know how to do things’ (Winter 1993). Increasingly, economists, strategy academics, and commentators agree that a business enterprise can best be seen as a coordinated collection of capabilities, somewhat bound by its own history, and limited in its effectiveness by its current cognitive and social skills. It is crucial to note that organisational knowledge is stored in organisational memory – either in a controlled or uncontrolled way.

The challenge enlarges when the organisation is extended. The main building block of collection of capabilities is knowledge, especially the knowledge that is mostly tacit and specific to the organisation. Although the new ideas have not totally displaced older ideas of firms as primarily information processors, productive machines, or bureaucratic structures, they have proved potent enough to act as a spur to real action in concurrent organisations.

Organisation itself can be seen as a system, and a system can be self‐organising. There is also a difference between system of subsystems vs. system of systems. Here, however, system is referring to (computerised) information management system, not an organisation alone.

In this research the term ‘focal company’ is used to cover the organisation in question. The focal company is an example from an extended organisation that needs knowledge management applications. The strategic needs and business environment demand knowledge management. For this there are plenty of reasons: The focal company is highly knowledge‐intensive, global organisation, focusing mostly on knowledge work, (for example R & D) and therefore should manage its rapidly growing knowledge organisation effectively.



88

The focal company is one of the world’s leading manufacturers of various devices and mobile, fixed broadband and other networks. The company was founded in 1865 but product range has changed in the course of time.

During this research, the focal company employed over 52,000 people and had 17 production locations in 9 countries; research and development in 14 countries; and sales to over 130 countries. The business strategy focused on innovation and design, experience, user‐friendliness, reliable solutions and customer satisfaction.

The focal company entered new communication products businesses and therefore needed to create new knowledge and renew itself faster (organisational learning, competence development), and also needed to understand the nature of ‘communications’ better, and how communication technologies shaped the society and knowledge based organisations.

Information systems were a very suitable research subject in the context of the focal company, as the software development represented a significant portion of the research and development activities undertaken in it. The focal company had gained also recognition of its exemplary role as a front line organisation to adapt KM practices (see e.g. Goh 2004). The SECI model of knowledge creation was generally the framework of reference for the early adopters of KM. With further findings of the contemporaries, and company’s own R & D, it was influential for the conceptualisation of related issues also in the focal company.

The writer of this thesis worked for years in the research centre of the focal company, where it was possible to study the subject. In Finland many doctoral dissertations from technology or economics were published (Cf. Lovio 1993; Kiianmaa 1996; Verkasalo 1997; Tuomi 2000; Ketola 2002; Sydänmaanlakka 2003; Luoma 2009, etc.) during those years. Even though the publications mentioned are dealing with the focal company, they are not used as references here because their subject is different.

The focal company has moved away from traditional product cycles and viewed the development of telecommunications technology as a continuous process. It was moving towards a situation where the whole product range is continually refreshed. Underlying this new strategy was innovation and transforming knowledge.

The organisation focused on learning to support its innovation efforts. For this it had three main drivers:

1. Competence management ‐ identifying and developing those key knowledge skills leading to continuous improvement by competencies.

2. Performance management ‐ daily leadership and a focus on business planning and execution leading to continuous improvement of performance.

3. Knowledge management ‐ individuals and groups converting tacit into explicit knowledge leading to the continuous application of new knowledge.

The focal company could be thought of as a ‘thinking’ knowledge organisation. Rather

than ‘jumping’ into knowledge management, the company spent time researching, thinking, reflecting and then visualizing what a successful 21st century company would look like, and how it would operate in a dynamic, integrated global environment.

No matter how well known the theoretical KM concepts were, and how ‘thinking’ the organisation seemed to be, specific at the focal company was the pragmatic approach taken to KM thus avoiding the problem of bringing burdening extras to the actual work processes and embedding KM into the processes, guiding methods and practices. This made the company ‘seaworthy’. Further insights into knowledge management were brought up later, taking into account the human component of KM and stakeholder involvement to the ongoing development of the KM environment. There may have been several factors contributing to focal company’s excellence in knowledge management (see e.g. Goh 2004) but the underlying scientific work with which the practices have been nurtured can not be without influence.

One compared KM system provider of the focal company had for over a decade seen the extended organisation as the right direction (see Figure 16).The same system provider used the same



89

concept in another presentation 2008 (e.g. Hackney 13.11.2008). These directions have been good for companies that remain socio‐technically competent, but when this paradigm shift is managed by finance, it could lead to disaster.

Business D irect ionsBusiness D irect ions

Recastingexternal

relationships

Organizationaltransformation

Businessprocessredesign

PerformanceTeam

IntegratedOrganization

ExtendedEnterprise

Inter-enterprisecomputing

Integratedsystems

Workgroupcomputing

Tapscott & Caston, Paradigm Shift

Increasing complexity as well as benefits

Increasing complexity as well as benefits

High-

Figure 16. First indications from extended enterprise (based on Tapscott & Caston 1992)

Extended in this context are all the elements, relations and influences that are touching the boundaries of the learning organisation. An extended organisation sees its links to members of its value chain in multidimensional ways. Processes and value chains are seen as extended beyond the enterprise into suppliers, customers, or other stakeholders. These features matched the focal company well.



90

1 CASE I – INDIVIDUAL KNOWLEDGE FROM DATA MANAGEMENT SYSTEMS

Automated data collection tools and mature database technology lead to huge amounts of data stored in databases, data warehouses and other information repositories. Manual inspection of that data is either tedious or just impossible. Exploratory, semi‐automatic data analysis on large data sets is needed. Human knowledge derived from the data is the final aim for all data warehousing, mining, archiving and data marts. The creation of new knowledge is most appropriate to data mining, which is still quite unfamiliar for average users, and therefore emphasised in this introduction. Even though operational needs are different in every‐day processes, the fundamental reason for all data management should be the final profitability of an enterprise.

The target is extraction of interest in (non‐trivial, implicit, previously unknown, potentially useful) information or pattern from data in large databases. There is usually a larger process called knowledge discovery in databases, that covers data pre‐ and post‐processing together with data mining and other processes. Some more rare terms are knowledge extraction and data archaeology. More wide terms in this context are knowledge management and business intelligence.

There are many views to data warehouses, mining and data marts, e.g. descriptive and predictive. The views are based on the data warehouses to be mined, on the knowledge to be discovered, on the techniques that are utilised or the applications that are adapted. The databases can be relational, transactional, object‐oriented or object‐relational databases that cover active, spatial or time‐series data. Data sources can also be text, XML, multi‐media, legacy, inductive, www or heterogeneous sources. For knowledge management, there is much wider range of sources.

The data sources include different kinds of knowledge. The process can be interactive. Data plying from one internal system to another can be modified during different stages. For an end user it is important to know the background of the data. There is 'noise' and incomplete data that must be dealt with. Also the challenges of performance and scalability arise from large data warehouses and multiple sources.

In an enterprise where there are big, quite independent organisational entities, it is important to know the differences between data warehousing and mining compared to operational database management. Data in the normal database management system is homogenous and focused on operational business needs, while data warehousing and mining are exploiting diverse data to generate knowledge. In operational databases there are query tools as part of large software suite, while data warehousing, mining and especially data marts consist of small‐scale software tools. Data warehousing and mining are dealing with diverse information while data management eliminates data redundancy. Data warehousing is reporting from historical data sets instead of operational data. Data warehousing and mining are flexible compared with dependence on systems development lifecycle as the means of designing data management systems.

However, there are some convergence between operational data stores and data warehouses and mining. In the beginning of this case study, data warehousing and mining were becoming more real‐time, creating knowledge for the future. The redundancy of data in different systems (not within single warehouses) was one result of this study, and will be one target even in the extended organisational strategy.

Systems

Sc1.1

The fieldscope of

April 200author, wresults fo

the briefand a quone and

researchshown ininvestigaexhaustiand C we

between inf

cope of case

Case Id – what the f the researc

This in01 to April 20were involveor the future

o Do Do Do Dcompo Kn

Therefing phase thuestionnaire two, and th

Table

Orga

Orga

Orga

Orga

IM +

TOTA

Expert

her derived cn the Table 1ations. The nive answers cere quite sim

formation an

I

began with applicationsh. nitial researc002, but coned. The purpe of applicati

ata Warehoata Mining,ata Archivinata Marts; pared with nowledge M

was also a bhe project wawas compilee results are

12. Number o

nisational un

nisation A

nisation B

nisation C

Corporate

AL

ts from varioconclusions a12. The expenumber of ancame from Omilar and the

nd knowledg

system perss for knowled

ch on data anntinued in theose of the stions and dat

ousing, , ng and

Managemen

briefing befoas agreed uped (see appee summarised

of experts res

nit Numbe

ous organisaabout the corts were alsonswers emphOrganisatione IM as well a

ge

91

spective becadge managem

nd knowledge backgrountudy was to fa. The scope

nt applicatio

ore, and the rpon, the enqendix 2). Thed here in this

ponding from

er of respond

tional units arporate situao interviewehasises Orga B, and Inforas corporate

Part II – In

ause the mament are – w

ge managemd for years. find out the e of applicati

ons.

result analysuiries on proe result analys document.

m different org

dents Aninv

24

4

4

9

41

answered toation. The nud and all thenisation A frrmation Manfunctions se

itial case stu

rket situatiowas investiga

ent applicatNo other recurrent situaons and data

sis phase afteoposed respoysis was done

ganisations

nswers lead tvestigations

o a questionnumber of acte developmeom the focanagement (IMerved them a

udies

n must be stated to reduc

ions took plasearchers, eation and to a was

er the projecondents were both after

to further

1

2

naire, from wtual respondnt observed l company, bM). Organisaall.

tudied first. ce the

ace from xcept the derive

ct. During re collected, the phase

17

2

3

5

27

which the ents is in further but more tions A, B

Systems

Co1.2

warehouwarehoumaking rmanagemcontain updated(O´Leary

individuais additioas a sma

creating the dataevaluatioprocessiuser is aanalysis

system dmajor buthe warestore theday, andthrougho

now’). Toperatiothey exp

level datorganisaare prohvirtual dconcept

subject aspecific dsubset oand depfor a com

warehouhierarchSnowflakAllowed

enterpriasked, todiscusse

between inf

oncepts

Case Iuses. A data wuses differ frrather than sment informmultiple yead on a transay 1998: 55).

Browsal. An alternonal knowledaller scale da

Knowa target dat

a mining taskon, 9) knowlng phase is tdjustable: mpower.

In pradesigned to susiness subjeehouse is stoe data differed another as out (e.g., mo

All datThese data aonal databasepect that it w

Data wta warehousations begin hibitive. Instata warehouconcerning t

A dataarea. The dadepartment of the entire artments thampany‐wide

A datauses and datical and temke schema. Tuser access

To seese data archogether withed separately

formation an

research bewarehouse aom traditionsimply efficie

mation systemars of transacction‐by‐tran

sing data or eative approadge or informta warehousledge discovta set, 3) datak, 6) choosingedge presento select the more function

ctice, data wsupport manects of the enored in a singently. For exmonth‐day‐yonth‐day‐yeata in the datre extractedes, data are will reflect accwarehousinge. These appby implemenead, an orgause. Therefothe usefulnea mart is a scata mart is usor business firm. The treat want to benterprise da warehousea marts cont

mporal data inThis architecis to select ‐e how data witecture, seeh sources andy.

nd knowledg

egan from onacts as a cennal transactioently capturims and the scction databansaction bas

establishing ach is to use mation hiddesing process.very process a pre‐procesg the data mntation, and frelevant datns and metho

warehouse isnagement denterprise, sugle, agreed‐uxample, one year. In the ar). a warehouse from, and reaccurate as ocurate valueg approachesproaches difnting an enteanisation ofteore, data maress of the teccaled‐down vsually designprocess. A cend toward guild local datdata warehoue can be constain functionn de‐normalture support‐ normally bywarehouses ae Figure 17. Id targets of t

ge

92

ne of the firsttral storage on databasesing transactiocope of this ses stored insis. Instead, t

queries genea process caen in the dat. has got the fssing (cleaninmining algoritfinally 10) usta with respeods per func

s a relational ecision‐makinuch as customupon format operationaldata wareho

e are accuratepresent ‘snof the momees as of the ms can range fffer in scale aerprise data en has a co‐orts are implechnologies toversion of a dned to suppocompany cangreater use ota marts focuuse. structed of mnally indepenised, multi‐dts (relationay queries. and data man this study,the data flow

Part II – In

t knowledgearea for an os in that theyon data. Theresearch vern the same dthe entire da

erally resultsalled knowledta. Data mini

following phng), 4) data rhms, 7) datase of discoveect to the cuction provide

or multidimng. A data wmer, vendor,even when u application ouse, the dat

te as of someapshots’ of, ent of accessmoment of acfrom simple and complexwarehouse boperating coemented moo solve busindata warehoort the uniqun have manyof the data mused on thei

multiple, relandent, functidimensional sl) on‐line ana

arts are situa, only the datw. Business r

itial case stu

e managemeorganisationy are designeerefore they ry well. Typicatabase. Datatabase is up

s only in knodge discoveryng, on the o

hases: 1) Learreduction or a mining (seaered knowledurrent task. Te the user wit

mensional datwarehouse is product, or underlying omay store thte will have a

e moment inoperational s. When useccess. (data mart) xity. In practbecause the llection of dast often, usuness problemouse that foce business ry data marts;mart often isr specific ne

ated data maonally depenstructures knalytical proce

ted in one tyta warehousrecords and d

udies

nt tools: dat's transactioed to supporsuit to the kcally, data wata warehousdated period

wledge for pry to determither hand, is

rning the doprojection, 5arch), 8) pattdge. Aim in tThe informatth greater fle

tabase manaoriented aroactivity. The

operational ahe date as yea consistent

n time (i.e., ndatabases.

ers access a u

to complex, ice howevercost and timata marts foually for prooms. cuses on a paequirements each focuse driven by eneds instead o

arts. Both dandent, aggrenown as Staressing (OLAP

ypical modelses and data data archivin

ta n data. Datart decision nowledge arehouses ses are not dically

particular ine if there s introduced

main, 2) 5) choosing tern he pre‐ion for the exibility and

agement ound the e data in applications ear‐month‐format

not ‘right In unit of data,

enterprise r, few me frame rming a of‐of‐

articular s of a ed on a nd users of waiting

ta egate, r or P ‐ ROLAP).

l of marts were ng were

a



93

Figure 17. Enterprise data architecture (based on Shannon 2001)

Because of its reduced scope, a data mart takes less time to build, costs less, and is less complex than an enterprise data warehouse. It is easier to agree on common data definitions for a single subject area than for an entire company. Therefore, a data mart is appropriate when a company needs to improve data access in a targeted area, such as the marketing department. However, the indiscriminate introduction of multiple data marts with no linkage to each other or to an enterprise data warehouse will cause problems. Because these data marts proliferate quickly, they may create the same problems that earlier file management systems did, including data redundancy, data isolation and data integrity problems.

Data mining provides a means of extracting previously unknown, predictive information from the base of accessible data in data warehouses. Data mining tools use sophisticated automated algorithms to discover hidden patterns, correlation, and relationships among organisational data. These tools are used to predict future trends and behaviours, allowing businesses to make proactive, knowledge‐driven decisions. The most commonly used techniques in data mining are: artificial neural networks, decision trees, genetic algorithms, nearest neighbour method and rule induction.

As general examples from knowledge‐related predictions, targeted marketing is a predictive problem. Data mining can use data on past promotional mailings to identify the targets most likely to maximise return on investment in future mailings. Other predictive problems include forecasting bankruptcy and other forms of default. An example of pattern discovery is the analysis of retail sales data to identify seemingly unrelated products that are often purchased together, such as milk and eggs at a supermarket. Another pattern discovery problem is detecting fraudulent credit card transactions.



94

Data mining identifies facts or conclusions based on sifting through the data to discover patterns or anomalies. Based on observations on systems, literature, and other sources (e.g. Klemettinen & Ahonen‐Myka 2000; Monash 2001; Rainer 2001) the main functions of data mining are:

classification – infers the defining characteristics of a certain group (e.g., customers who have been lost to competitors);

clustering – identifies groups of items that share a particular characteristic (clustering differs from classification in that no pre‐defining characteristic is given);

association – identifies relationships between events that occur at one time (e.g., the contents of a shopping basket);

sequencing – similar to association, except that the relationship exists over a period of time (e.g. repeat visits to a supermarket or use of a financial planning product);

forecasting – estimates future values based on patterns within large sets of data (e.g. demand forecasting).

There is a wide range of knowledge‐based applications for data mining (cf. APPENDIX 2:

Systems investigated). These areas include customer relationships (i.e., retention), cross‐selling and up‐selling, campaign management, market, channel, and pricing analysis and customer segmentation analysis.

Even though the applications are highly business‐oriented, there is a firm scientific basis for data mining. Based on the previously mentioned sources, theoretical foundations of data mining could be derived into six basic blocks:

1) Data reduction To reduce the data representation Trades accuracy for speed in response

2) Data compression To compress the given data by encoding in terms of bits, association rules, decision trees, clusters etc.

3) Pattern discovery To discover patterns occurring in the database, such as associations, classification models, sequential patterns, etc.

4) Probability theory To discover joint probability distributions of random variables

5) Microeconomic view A view of utility: the task of data mining is finding patterns that are interesting only to the extent in that they can be used in the decision‐making process of some enterprise

6) Inductive databases Problem of performing inductive logic on databases The task is to query the data and the theory (i.e. patterns) of the database Popular among many researchers in database systems

Data Model is simply a description, either in words or diagram form, of the way a

database is organised. They can be very wide but implicit. The data model defines the way data are conceptually structured. The three most common data models are hierarchical, network, and relational. Other types of data models include multidimensional, object‐oriented, object‐relational, and hypermedia models. Using these models, designers can build logical or conceptual views of data that can then be physically implemented into virtually any database with any database management systems

Systems

(DBMS). linked lisdata.

containsAn ODS in the oroperationormalisassurancapplicati

record mbusinessother hawhich thnot a co

applicatiof the inoperateswhere thhelps org

during threpresen

quantitadata. Knrange ofpages (bwarehouservers.

managemsystems and loosmeaningfrom datas ‘databinformat

Cu1.3

Informatthey werof data,

between inf

Hierarchicasts. Relation

The Os functionallycan provide rganisation tons and manased on the Oce, data miniion database

Data amanagements activities. (Rand, almost ahe cost savinllection of m

Documion of data mfrastructures with less sthey do, it is fganisations t

• find the• relate d• group d

Text m

he ten years nt areas of go

Documative data typowledge maf databases, both internal uses are likel

In genment applicamust tightlyse coupling (agful to keep kta warehousbases from wtion species,

urrent situat

The 'dtion in the syre not aligneinformation

formation an

l, network, anal and multi

Operational Dy independenthe end‐useo access infoagement rep

ODS server. Ting and the pes and feeds archiving is mt and archivinRecord in thias importantgs can be ac

miscellaneousment managmining to none of stored datructured inffrequently foto:

e ‘hidden’ codocuments adocuments b

mining (also wof knowledgood progressment managpical of data anagement aincluding daand externay to be virtu

neral, the datations can coy couple withand no coupknowledge msing, data mawhere applic which is par

tion of applic

data' in knowystems existeed or harmonand knowle

nd knowledg

and object‐ordimensional

Data Store (Ont, functionars with an inormation forporting. ODShis relationaproduction ato other orgmore genericng is to colleis context is t as the presehieved. An as informationement is then‐structuredata to extracformation. Docused on do

ontent of doacross previoby common t

web mining ge discoverys during the ement has awarehouses

applications ita warehousal), Lotus Notal warehous

ta warehouso‐operate —h databases apling, to list amanagementarts and dataations generrtially transfe

cations and

wledge manaed in many fnised. Knowldge content

ge

95

riented DBMl DBMSs rela

ODS) is an infally dependentegrated vier strategic an data are boal architecturand managemganisations. c term for stct a fund of a wider concervation is tharchive is a con. e subject of t or less strucct additional Documents raocument form

cuments, incous unnoticethemes.

in some exte in databaselast ten yeara common grs, knowledgen the contexses, work protes and otheses where th

sing, data mi in practice, and data waall possibilitiet applicationa mining apprate knowlederred to kno

content

gement appformats ‐ theedge manag. Practical in

Part II – In

MSs usually tiate data thro

finitely extenent, relationaew of enterpnd/or tacticath current anre is a staginment of data

oring the histraceable infcept than in he obliteratioontextually b

the case II, actured text fuseful informarely have stmat rather th

cluding addited divisions;

end) is one as. Also assocrs all over thround with de warehousext of data geocesses, newer groupwaree knowledge

ning, data arcouple with rehouses. Thes). As a resus and businelications. In dge’ but comowledge pres

plications weere were temgement seemformation re

itial case stu

e related daough informa

nsible, univeral, hierarchicrise data to el decision sund historicalg area that sa warehouse

story. The ratformation tothe databason of uselessbased organi

nd defined tiles. Data mmation. Texttrong internahan docume

tional useful

area which haciation rules e world. data mining. s are aimed nerate knowws articles, exe, and peoplee is dispersed

rchiving and each other. here are alsoult from this ess records athe results, p

mmunication sentation.

re not necesmplates for dmed to be anelated to bus

udies

ta together tation contain

rsal repositoal and aggreenable userspport, day‐to and is storesupports dats, data mart

tionale behino be used to e terminologs documentaic body of ev

here. Text mining takes at mining, howal infrastructent content. T

relationship

as progresseand data vis

Rather than more at qua

wledge from xternal databe. Thus, knowd across a nu

knowledge Ideally, data

o semi‐tight ccase study, iarchiving distpeople are nis used as a

ssarily so expifferent purp umbrella fosiness progra

through ned in the

ory that egate data. s anywhere o‐day ed a quality s, new

nd all promote gy.) On the ation, by vidence —

mining is the advantage wever, ture, and Text mining

ps;

ed well sualisation

the kind of alitative a wide bases, Web wledge umber of

a mining coupling t is tinguished not defined generic

plicit. poses, but or all kinds ams:



96

roadmaps, feasibility studies, plans, feature descriptions, manuals, customer documentation etc. Data were entered into systems mainly by assistants and other information workers. It was mixed quality, but there were efforts to harmonise the quality, like versioning and date control and solutions for problems of different 'storing locations'.

Knowledge Management applications covered all other applications in this context, as presented in Figure 18. Together with the high‐level classification of applications, another memorable detail in this table is that content (management) is nowadays in the same immature conception phase.

Figure 18. Content originates from Asset Management, but is diffusing (based on Binney 2001)

Analytical knowledge management, where data warehousing and data mining belong to, provides interpretations of — or creates new knowledge from — vast amounts or disparate sources of material. In analytical KM applications, large amounts of data or information are used to derive trends and patterns — making apparent that which is hidden due to the vastness of the source material and turning data into information, which, if acted on, can become knowledge. This is quite common interpretation of knowledge in business, and the incidence of interpretations lead to new concept presented in the common results.

Traditional analytical KM applications (such as management information systems, data warehousing and data marts) have analysed the data or information that is generated internally in organisations (often by transactional systems). They were joined by a range of business intelligence — or competitive applications which incorporate external sources of knowledge or information. It is remarkable that all of these analytical KM applications are still just a fraction of the whole scope.

There were some general needs and problems in organisational KM applications: Information was stored and shared in many ways. There were no high level common processes and guidelines, but a need to harmonise the existing and establishing new simple processes and guidelines. Then there was a need for accessing right information and knowledge when it was needed at the most for the right persons. Information existed, but the question was how to improve accessibility. One solution was to build bridges to the content and channel them to right target groups.



97

Knowledge and information are tied to organisation, but work is not. A trend was towards more team or interest group based information (compare with collaborating communities). As a matter of fact, the ‘team’ based organisation did not work as well as communities based on pure interest.

Organisation‐based intranet was difficult to search and browse. The searcher needed to know the organisation well in order to ensure successful search result. A more common meta‐knowledge is to know the right person, but the value depends on many secondary factors. As a memory model, building a category or taxonomy portal to access relevant information from a single source could improve real meta‐knowledge and thus business efficiency.

Information security should be based on an access control framework. Security and access control management is needed to support the information channels in the intranet. Role based access for employees who belong to a certain interest group would be ideal. It is also essential for modern partnering mode, like extended organisation in this case — but extended organisation can go even further: Intra‐ and extranet can look very different to different collaborators, based on their rights and roles.

Customer and competitor information should be channelled to the right persons in the organisation. Business, operator and competitive intelligence need to be carefully looked into, to distribute key information from the area‐organisation, standardisation committees and account teams to right persons in production. That will help to make key changes to e.g. roadmaps, features etc.

Global and unified business communication tools and technologies were utilised more and more, but still the same tools were used for different purposes – that caused confusion. A requirement found was more centralised approach globally to support increasing information needs of business programs.

1.3.1 Phase 1 – System perspective

Phase 1 market situation of data warehouses and marts, data mining, data archiving and knowledge management applications are summarised in this subsection. When referring to focal company, organisational units are distinguished based on process: Company A, Company B and Company C represent the product creation process, while information management and other corporate‐wide units are representing the delivery and business support processes.

In modern companies for data warehouse, data mining, data archiving and KM there are many application domains: from retail industry to biomedical data analysis. For the focal company the most important application domains are telecommunication industry and financial data analysis. The processes, however, expressed themselves as dominant entities and therefore process is a basic building block of the final model.

For the financial data analysis data were usually of high quality: relatively complete and reliable. For multidimensional data analysis and data mining there must be multidimensional data warehouses designed and constructed. Examples are statistical information (max, min, total, average, trend, etc.), or debt and revenue changes by month, by region, by sector or by other factors. Financial data can be used for marketing purposes also: e.g. classification and clustering of customers.

Telecommunication industry was rapidly expanding and highly competitive and had a great demand for data warehouses, data mining, data archiving and KM applications and data. Applications help to understand the business and identify telecommunication patterns. They help to make better use of resources, to improve the quality of service and catch fraudulent activities. Also the telecommunication data analysis is intrinsically multidimensional with calling times, durations, location of callers, types of calls etc.

Financial business data and telecommunications data had got some common fields in multidimensional association and sequential pattern analysis: operators with networks can e.g. to find usage pattern for a set of communication services by customer group, by month etc., to promote the sales of specific services or to improve the availability of particular services in a region. Both specific data types need visualisation tools.



98

Visualisation is not only used to represent information, but it also helps to create knowledge. From the data processed with data warehouses and marts, data mining, data archives and KM applications it is easy to create meaningful pictures. For visualisation there are different kinds of categories: data visualisation, mining result visualisation, process visualisation and visual user interfaces. Easy‐to‐use and high‐quality graphical user interfaces and query languages are essential for e.g. user‐guided, highly interactive data mining.

Applications can use different functionality or methodology; e.g. data mining applications may even work with completely different kinds of data sets. System issues are also important: is the system running on only one or on several operating systems? Does it follow client/server architecture or provide web‐based interfaces and allow e.g. xml data as input and/or output?

It is noteworthy, that the same application can be used for different purposes. Relational databases were (and still are) fundamental applications for all data management. SQL Servers with OLAP (On‐line analytical processing) are sources for all knowledge management and delivery: report mining, report databases and ad‐hoc reporting; data mining, document imaging and workflow systems. They have got in‐built OLAP functionality when creating analytical applications. The trend was to go towards relational databases for really large‐scale data warehouses with less maintenance – in the future it may be combination of any data sources.

There are many practical ways to classify the applications. Case I used the following categorisation of applications based on www‐sources and from parallel research projects:

Data management Data Mining, Data Warehousing, Data Marts, Information Retrieval, Knowledge Discovery, Collaboration, Knowledge Mapping, Online Training Systems, Document Management, Project Management Reporting, Organisational Memory and other KM products. All of those categories were analysed, and all systems are listed in APPENDIX 2: Systems

investigated. The technical details are not reprinted here. Combined with a number of interviews from different organisations (specified in Table 12) and recurring observations, a conclusion is that data management is the base for knowledge management. Except the traditional data, also the semi‐ or unstructured content and meta‐data are crucial. The trend is to deliver the data in any form of communication – not only manage the data. Data management is using object‐based technologies like generalisation and specialisation hierarchies, multiple inheritance, user defined data types, triggers and assertions, support for knowledge based systems, recursive query expressions, and additional data administration tools. It does not matter how visible the data is for the end user – there are data in some layer anyhow.

For knowledge management, the multimedia store and search options in Oracle, DB2 and Informix, as well as Object Databases from big vendors like Computer Associates (CA) and smaller ones like Versant, were crucial. Corporate wide knowledge management should be independent from vendors and operating systems. For that reason the chosen products for detailed comparison were Oracle, Informix, DB2 and SOLID.

Already in the beginning of this research, when the universal databases were compared, the vision was to join data from different data sources – independent from the manufacturer and



99

including content like video and image files (See e.g. Varma 23.9.1997). Normally knowledge management systems need to have options to manage complex data. First, Binary Large Object (BLOb) is a part of the standard, but ‘object‐aware’ RDBMSes use storage and indexing technologies customised to each data structure. Second, content‐based retrieval requires special methods. Third, to deliver peak performance for search and retrieval of complex data, query optimisation and the retrieval process itself must be customised to the data being retrieved (Weldon 1997: 110, 114). The complex data types were present in all case studies.

All the concrete KM system requirements were quite trivial, what comes to actual RDBMS platform. All the products met the requirements. The question was how well the requirements were met. RDBMSes must embrace some characteristics of Object‐Oriented Database Management Systems (OODBMS) to become Object‐Relational Database Management Systems (ORDBMS) (Weldon 1997: 109). There are many features in Object‐Oriented Database Management Systems that Object‐Relational Database Management Systems do not include. OODBMS offer full data encapsulation, inheritance, full integration with OO programming languages, and all the other defining characteristics of OO software (Muller 1997/2: 111). The object‐oriented features, interpreted through information species (presented in the Part I section 2.1, Paradigms from philosophy and science) are adaptive to memory management model.

1.3.2 Records and archiving

The importance of archiving popped up in this case study. The record concept is needed in system categorisation as well. The archiving of business records is based on general requirements set by national laws and regulations, by ISO 9000 Standards, by approval authorities, and by internal requirements.

Records are documents captured to provide evidence of the activities carried out in any group or organisation. Essentially, they are an important and valuable asset for enterprise, and their creation, control and preservation should be managed accordingly. The absence of records can open corporate employees to accusations of fraud and impropriety or an inability to defend any organisation in cases of legal action or claims. Records creating organisation (with assistance of Legal, Innovative, Quality etc. functions) defines which subset of all data generated by business processes is captured as records.

There are three roles of records management and archiving process: First, development facilitates archiving and records management activities enterprise wide. Second, records management is a function to be implemented in other organisational units and functions by redefining the tasks of appropriate organisations. Third role of archives is to deliver archiving services regionally.

Appraisal is a process that examines the values of records and determines their retention periods. Appraisal is not just determining the eventual record disposal but a continuous process of determining the records value for one or more of the recognised purposes, related to the costs of preservation.

In an electronic environment, the records lifecycle must be extended backwards from the traditional paper environment — to a stage prior to records creation. Functional requirements for the management of electronic records are addressed in the design and specification of information systems in order to ensure that the records created or retained provide reliable evidence. That is meant by conception.

Activity in the creation stage involves monitoring records creation practices to ensure that the records are created as expected. The business processes and systems of each function must operate to capture records, which provide evidence of all business transactions. It should be ensured that 1) appropriate and complete records are created and captured, 2) meta‐data are captured, 3) links to other records are established and maintained, and 4) records creators create and capture records which are authentic, reliable, and preservable.

During the use some control must be taken. Actions with respect to records during their active phase should be framed to facilitate continued maintenance and accessibility of records after



100

they have satisfied the creator's business needs. Similarly, steps taken for archival preservation must ensure that the records continue to provide reliable and authentic evidence of the creator's activities.

Disposal includes the retention or deletion of records in or from record‐keeping systems or from the archives. It may also include the migration and transmission of records between record‐keeping systems, and the transfer of custody or ownership of records. If the requirements or need to retain the records have been changed, the retention periods, or the requirements concerning different types of records can be altered in reappraisal.

Records must be stored in conditions, which ensure that they are accessible and retrievable in appropriate timeliness for the length of time they are retained. The central goal must be to preserve information integrity; to define and preserve those features of an information object that distinguish it as a whole and singular work.

It should be ensured that physical destruction of records is carried out in accordance with best practice applicable to the type of record and its relative security requirements. The event log about their destruction has to be preserved long term or even permanently to prove that records have existed and have been deleted according to approved process. The deletion of records from the record keeping systems can not be automated but it always has to be approved by an authorised person.

1.3.3 Phase 2 – Users’ perspective

In general, data in the data warehouse, data archiving, or data mining applications are different from data in the operational databases, depending on the length of time since the data were extracted (i.e., the length of time since the snapshot was taken). That is one feature of organisational memory. Data in the data warehouse represent a long time horizon – from five to ten years. The time horizon represented for the operational environment is much shorter – e.g. from current up to ninety days. Also, all data in the DW contain an element of time, such as day, week, month, and year. Data in operational systems may or may not have an element of time.

The data warehouse transforms data into a more useful resource by grouping them more conveniently for end users, putting them into more usable formats, enabling them to be analysed, and dispersing them to appropriate groups to increase availability and accessibility. Preparing data to be loaded into the data warehouse involves extraction, transportation, and transformation. Extraction programs are run periodically to collect relevant data from operational systems. After the data have been transported to intermediate files, transformation programs integrate data from multiple sources, remove unnecessary data, correct poor quality data (e.g., redundant names and addresses), and summarise the data along any dimension.

Data warehouses contain current detailed data, historical detailed data, lightly summarised data, highly summarised data, and meta‐data. Current and historical detailed data are voluminous because they are stored at the highest level of detail (least amount of summarisation). Lightly and highly summarised data are necessary to save processing time when users request them and are readily accessible. Interestingly, older historical data are the least used data in a data warehouse, while highly summarised data are the most used.

Data types can be different: relational, transactional, text, time sequence, spatial etc. The selection of data may require a multiple dimensional view. The data sources can vary from ASCII text files to multiple relational data sources. Support for ODBC (OLE DB, JDBC) connections can vary.

In most big corporations there are many DW applications and therefore many kinds of data in product creation process. For example, resource and project management data is often common. There are critical systems, e.g. Production control systems, where data are normally in better shape than data in the rest of the systems. One common problem seemed to be that the quality of data was difficult to improve if the business representatives were not participating to the work, or they were otherwise difficult to contact. That required special communication besides the data.

The quality of data was not a unanimous evaluation. Product creation process workers could evaluate the quality of data high, while support representatives were considering it low ‐ or vice versa. Even the users of the same systems could have contradictory opinions about the quality of the data. A criteria and a scale to measure the quality of data were needed.

Systems

models fcommon

1.3.4 C

managemapplicatishould bworkflowdata effeapplicatiimportanapplicati

In those is one gudepend factors. Abe based

Fu1.4

mining, dexplorinAlso visuThere wfacilitatesystems with webover the

1.4.1

There wbelongedApplicatre‐defineinterfacedata stoBusiness

requiremhourly).

differentaccuracywill be c

between inf

In corfrom the sysn use for all d

Conclusions

The bment applicaions are morbe selected aw, working mectively and ions before tnt to go throions are not

Data mprocesses, puarantee for on the accurA criterion od on high‐qu

uture ‐ trend

Some data archiving. There willual user interill be new me systematic in industry ab mining. The world.

Future trend

Thereere critical ad to e‐busineions were froe the data we tools for DWrages for pros Objects or C

The sement is to spTherefore O

Third,t source systy, operation reated. Ther

formation an

porations, dtems, but mdata. Data an

from the cu

ig picture of ations and dare and more and archived methods, sysefficiently. Othey are anaough before important. mining and cpre‐processinthe quality oracy, compleof measuring ality data.

ds, threats an

general trenng and know be applicatirfaces will beethods for mdevelopmenand society. e interests o

ds

were some reas like proess area, eveom categorie

warehouse stW maintenaoduct creatioCrystal prodecond trend eed up getti

ODS type syst metrics reptems so that profit, cost pre was a nee

nd knowledg

ata are usedost often nond reports ar

rrent situatio

data warehoata is beginnimportant. T— i.e. normtems, prograOne general lysed and mplanning on creation of nng could takeof data. The eteness, cons or evaluatio

nd solutions

nds for the fuwledge managion‐specific ae more genemining compnt, improve i Privacy protof the users a

general trenoduction, quaen though coes of data inructure and nce. There won process inucts. found was tng informatitems will emorting was gfact based inperformanced for metrics

ge

101

d for many diormal users dre delivered

on of case I a

ousing, data ning to formuThe amount

malised. The bams and mandanger was tade more efthe technica

ew knowlede up to 80 %quality of dasistency addeon of the qua

s

uture were cgement systeand 'invisibleral. Forthcomlex types of dnteroperabitection and iand usability

nds, like to deality and logiould be categntegration orrelated interwill be more n general, ba

he use of Opion all the timerge. growing. It renformation we etc. Insteads also in prod

Part II – In

ifferent purpdo not separato several u

applications

mining, dataulate. Data aof data is grbusiness, supnuals were nthat the procfficient. Theral application

ge were not % of the total ata is the based value, intality was req

clearly foundems will be ie' data mininming scalabledata. Langulity and prominformation y of the know

evelop smaristics. Customgorised to anr data marts.rfaces and restatistical daased on DW p

perational Dame (from mo

equired confowas availabled of creating duct creation

itial case stu

poses. There ate data modser groups.

a archiving aand Knowledowing all thepport and denot always becesses are fixrefore the prn level. That

used much time. Right sis for all, beerpretabilityuired. High‐q

d. The data wintegrating. Ang systems, we methods aages will be mote the edusecurity will wledge will b

rter systems mer Relationnalytical know. In the futueporting inteata warehouplatforms lik

ata Storages onthly to we

orming and pe for decisioreports metn process.

udies

possibly aredels. Report

and knowledgge Manageme time ‐ evenelivery proceest possible txed to the syrocesses weret is not to say

in the curren kind of pre‐ecause all they, accessibilitquality decis

warehousing,Applicationswith built‐in fre based on standardiseducation and be a big issue the main p

using smartenship Managewledge manre there will erface and usses and operke Informatic

(ODS). A genekly or daily

processing dn‐maker, e.gtrics objects o

e data ing is a

ge ment n though it sses, to manage ystems and e more y that the

nt situation. processing e results ty and other sions must

, data are functions. constraints.d ‐ it will use of ue especiallypoint all

er tools. ement agement. be needs tose high‐levelrational ca, Cognos,

neral or even

data from g. planning or entities

y

o



102

The fourth trend concerned business planning. The requirement of seamless view to ‘actual reporting’ and business planning was challenging. Usually DW processes are focused on the history (long or very near ‐ but history anyway) data in source system. The nature of planning data is different compared to history data (actual). It is ‘in the future’, it changes frequently, it is unstable, it must be modified, and the future business structures and product portfolio can be totally different compared to current structures.

The most important business requirement for planning‐reporting process was continuous planning which means the possibility to follow product's lifecycle from development stage to selling last unit and get this knowledge to planning phase as base information. That information is including sold volumes, costs, profitability etc. ‐ from the idea of the product until the last item has been sold. Planning (especially long range planning) required more flexibility for simulating and generating different scenarios.

Knowledge management applications were expanding. The enabling technologies will be based on meta‐data and information access control for KM applications.

1.4.2 Threats and solutions

The conclusions written in this section are all about the future based on the results. One general threat, especially for big corporations, is that enterprise‐level planning is not done or not communicated. The big picture is missing ‐ even though parts of it are drawn in some companies. That is also relevant to the problem of co‐operation between processes: product creation, delivery and support processes should be working together right from the beginning of the projects.

Ownership of data was one general problem in corporations. There were no owners of the data, or there were too many owners. The ownership should be clarified in order to ensure that the business benefits could be achieved. This requires high‐level business sponsorship. Shared dimensions and hierarchies with non‐unique data between business functions was another problem.

Meta‐data management was a second challenge. A common need for identification 'keys' (using of some kind of 'master systems' for identification data) was quite general for many systems. One solution was to manage meta‐data in one location. The meta‐data was in this context structured and machine‐readable, not any wider concept.

Archiving the historical data of data marts should ensure that data are stored always in accessible format and not dependent from the infrastructure. Neither data should be lost, although process, data model or other system component changes. Data lifecycle can be many years ‐ following a product lifecycle. A solution for customer requirements of speed and performance is that data marts will contain only data that is relevant in reporting and analysis.

In a competing corporation, reporting was needed more often: the frequency changes from monthly to hourly data. Self‐service, and reduced data collection time were required. Data were often fed from one data mart to another. There was a risk that data flow will become erratic when rules are changing in some data mart. Solution was that the rules and the whole feeding were not in a 'black box'.

In corporations, there was a constant competition for selection of different products. There could be several tools used for same purpose. The problem has been in the area of data warehousing that not all reporting tools are suitable for product creation organisation where it is not possible buy the tool for everyone. That created small side effects, e.g. linkages to e‐mail system have been missing, the costs per user have been enormous and concurrent licences were not available. There should be a solution for licensing and linkages between systems.

Architecture should be easy to change, following the core business. That meant following of changes from business requirements, specifications and design all the way to the applications. When business requirements are changing, the specifications and design are changing, and the new applications should be generated. ‘Real’ data management tools and platforms should be seen in every day work.

One big problem was that people were stuck on their old methods. Everybody was used to create own reports instead of using comparable, general reports. Another, even deeper problem was



103

that people did not care to improve the quality of data in source system. People modified own figures according to what they should be, although the source data should have been corrected. There should be a culture that does support changes in practices and cares from the quality of data. Quality management could be more involved.

There were some common problems like the costs during application development and use phase. The system must have a user interface for both developers and end users. The definition and management of user rights was one problem. There should be both internal and external resources for support available during the development phase and for ordinary usage as well.

Maybe information management departments should strengthen their own competence in data warehouses and co‐operate with business right from the beginning of DW projects. There might still be lack of common support and knowledge. Version updates, massive growth of data and thousands of new users were big challenges.

One general solution is that data and knowledge management applications will become a similar service as infrastructure is. At least the common platforms should be decided. The development roadmap is hard to draw when DWs are crossing multiple organisations.

Data normalisation should be done on corporation level. That helps to get rid of redundant data. All static data should be separated from dynamical data. Meta‐data is an important issue ‐ the ownership and processes should be clarified to organise the creation, maintenance and deletion of meta‐data. Meta‐data should be an important part of process and concept work in co‐operation with terminology experts. The experts of terminology could be responsible also from storage. There are two separate classes: business meta‐data and technical meta‐data.

Knowledge management areas should concentrate on content specific information management instead of organisation‐based information. Different attributes and other meta‐data were needed for knowledge discovery and information security. More cooperation was needed to avoid reinventing the wheel.

Systems

2 CAS

This caseinterview‐ 210). Himportanresults a

managembecauseand procmanagemknown pconcept

summarorganisagroup to

The issueaccompl

organisaproject cevaluatioavailablechosen twere pro

Ca2.1

administthat couorganisaground oorganisamany pr

specific. the same

between inf

SE II ‐ ORGAN

Alreade study utilisws (Glaser anHowever, thisnce of featurare not based

This sement system the most imcess of documment technoproducts (witwas used w

This fiy about the ation was thao group. It wa

es surroundilished:

Phasefound

Phasefeaturthere Docum

Objecation so that consisted of on framewoe (Documenttools were poceeding as

ase II, phase

The fitrators and uld be derivedation's documof case organation managerocesses cou

Case oEvery projece purposes. T

formation an

NISATIONAL

dy basic groused new sournd Strauss 19s case does nres. Everyond on a samplecond case sm, and the sumportant elemmentation. Tologies and pth some diffeith other proirst section oproject phasat the practicas felt that d

ing documen

e 1: Checked d out the cur

e 2: Further dres of documwas a possibment Manag

tive was to fthe focal cothe previousrk for testingtum and Lotuiloted in the presented in

1 – users’ p

rst phase cousers from vad from the inmentation mnisation was ement thereld utilise docorganisation ct was decidThat was on

nd knowledg

L KNOWLEDG

unded theoryrces for qual972: 167 ‐ 18not include ce involved ine, but compstudy is namubject is documents were This case stupractices. Theerent featureoducts as weof the case stses and theirces of documdocumentatio

ntation at the

the backgrorent practice

discussion anment managebility to test ger).

find out the dmpany will hsly mentioneg purposes (nus Domino Dselected pro

n the append

perspective

nsisted of inarious parts nterviews. In

management.on weak ba seemed to bcument manwas doing aing how to me reason for

ge

104

GE FROM IN

y uses data fitative data ‐83), but also complicated n this case II lete enumered as organiumentation mcontent (infody found oue purpose ofes) suit for mell. tudy concernr results. Thementation maon should be

e case organ

und of knowes and needs

nd quantitatiement systemthe chosen t

documentathave an appred Phase 1 annot includedDocument Mocesses. Thedix.

terviews witof the case onterviewed u The interviesis, because be very rare agement tooalmost merelmanage docu copies from

Part II – In

FORMATION

flexibly (Glas‐ caches of dquantitativestatistics; onhas been resration. sational knomanagemenormation or t the best wf this case stmanagement

ning documee initial situatanagement de managed b

nisation were

wledge and ds by qualitati

ve survey coms in chosentools (e.g. Do

ion and docuropriate systnd 2; but thed in this repoanager) in lig interviews a

th different korganisationusers were noewees felt ththe knowleddocumentatols. ly project woumentation. m each docum

itial case stu

N MANAGEM

ser and Straudocuments ae data (Glasenly numericasponding to

owledge fromnt (not documknowledge) ays to use efudy was to et of knowledg

entation mantion in the andiffered frombetter.

e clarified. Th

documentatioive interview

oncerning then processes. ocumentum

ument manaem to answeere was also rt). It was poght of the peand answer s

kind of speci. There wereot satisfied what the wholedge delivery tion processe

ork. DocumeToo many syment.

udies

MENT SYSTEM

uss 1972: 161nd qualitativer and Strausl data figurinthe question

m informatioment managewritten in dffective docuexamine howge. The same

nagement givnalysed casem project to

he following

on managemws.

e possibilitieDuring phaseand Lotus D

agement neeer these neeadapted a syossible to comerceived neesheet questio

alists: manage lots of genewith any of te documentawas not wores defined, e

ents were proystems were

M

1 ‐ 220). ve ss 1972: 185 ng out the ns, the

n ement), ocuments, ument w the well‐e evaluation

ves a e project and

tasks were

ment and

es and e 1 and 2 omino

eds at case ds. The ystem mpare tools eds, and the onnaires

gers, eral notions he matrix ation rking. In line even though

oject e used for

Systems

(active dtraditionmail andduplicat

not be fothink tha

mail was

the strucpoor intr

system weasily as

dissatisfaabout. Pthat werworkflowundersigpatentin

applicatieven tho

Ca2.2

questioncase orgmanagemfor featu

V

12 As the matter so

between inf

Delivedocuments). nal documend intranet haing conversio

Most ound. Adminat it was com

Mail ws a wrong so

Searchcture of intraranet search

Therewithout conts copying fro

Workiaction with NPC‐docs and Pre sent by e‐w was regulagned scannedng processes

Some managtempltechnthe Ch

How tions were noough that is a

ase II, phase

The sennaire and inganisation. Thment systemures are in de

Very importa

1. Wi2. We3. Sea4. Wo5. Au6. Wo7. Do

‘working papeome systems a

formation an

erable documThere were t format. Difve many disaon from oneof the copienistrators thomplicated to was the tool lution. Direch by keyworanet pages wh engine inde was no genetent. But it wm personal ding Team RoNotes applicPC‐master wmail from abar, but could d pdf files, hcould widelydocuments gement werates were (iically‐orientehapter

the cases weot well knowa (risky) met

2 – system

econd phasenterviews abhere were rem features foescending or

ant features

indows Clieeb servers aarch Resultsorking and pudit Trails orkflow ocument ma er’ metaphor are restricted

nd knowledg

ments were malso plenty offerent kind advantages. file format ts were out oought that it request manto store andctory and fileds (with searwas more imexing. eral easy mewas a poor exdevice to fileom applicatiations, but L

were considerbroad – couldnot be autoowever, couy utilise docucould be sple used for mn the beginned processes

ere reached, wn. Public data‐repository

perspective

e of documenout the impoespondents for their procerder of impo

‐ independe

nts and thin clies Display oppublic versi

anagement

also indicates to manage ‘p

ge

105

managed diffof files that iof documentBetween difto another.of date, or thwas difficultnagement frod search for de names musrch engines) portant than

ethod to accexcuse. Docume server. ions were froLotus Dominored as ‘relictd be seen asmated as lonuld have beenument manait into modu

modular docuning of this cas). Intranet t

Web Site Butabase abouty.

ntation manaortance of dofrom four proesses. In thertance as fol

ent from pro

ents ptions ‐ Docions

system in g

s, the conceptpapers’ in elec

Part II – In

ferent way cncluded ‘knots should befferent syste

e right versiot to manage om administdocuments. Ast still be logicould be mi

n using searc

ess shared dments should

om a narrowo Documentts’. Other app an exampleng as agreemn managed aagement tooules. Softwarumentation pase) rarely nools (e.g. Co

uilder's Toolbt competenc

agement resocument maocesses meae case organillows:

cesses

cument list

general

t of documentctronic form.

itial case stu

compared wiowledge’ con stored withems there sha

on (not necethe documerators. Attached doical. isleading. Soch engines. T

ocuments. Pd be transfer

w sector. The Manager wplications, e. of wrong soments were oas attachmenls. re tools usedpurpose alsoeeded (or noommunities o

box and othecies was requ

search projecanagement sasuring the imsation the st

t has roots in

udies

ith working pntent, which a single tooall not be an

essarily the laent users …an

cument deliv

me users thoThis could be

People did norred to the sy

re were somas not comp.g. Excel‐sheeolutions. Conon paper. Thnts. Inventio

for source co. Strict workot known in of Practice m

er Notes baseuired as a KM

ct consisted oystem featumportance otrongest ave

papers, and fo

papers12 was not in l. Notes, e‐ny

atest) could nd Users

very by e‐

ought that due to the

ot use the ystem as

me plained et invoices ntract e n and

code kflow and less

mentioned in

ed web M tool –

of a res for the of document rage needs

or that

Systems

and LotuIn the saProjects,and quic

Co2.3

in inform

way. Forprocessidocumerevision purposeinformat

docume

trying tothe distrof technclients aorganisaDocumetransactorganisaeffective

control aorganisaproductscharacteintelligenmanagem

storing, retrievalcollaboramain fun

documealways b

Ba2.4

organisa

between inf

8. Sec9. Lif What

us Domino Dame section t, Legal, Admck to glance.

oncepts

The domation system

Documr a record mang package ant may be reof an older o and may betion it contai

Digitant composed

The teo describe whributed repoologies that nd their supations to inpuentation manion‐orientedation. It coveely process th

Documand repositoation. Neverts and more. er recognitiont organisatiment.

For thlocating, andl are the orgaate when crenctions are th

Documnts has growbeen best po

ackground

Documational memo

formation an

curity integfecycle

were the difocument Mathere is analinistrative anPlease refer

ocument mams. Most of ments are coanager a doca document evised, but thone. A docume used for mains may no loal documentsd of various ferm documehat it really msitories of doenable orgapliers. It is thut and retrienagement incd and collaboers as well noheir organisament managory‐oriented theless, effecThe integratn and other ional manag

his case studyd keeping traanisational peating informherefore creaments have bwn. The workossible, as fou

ments form aory. Any succ

nd knowledg

gration

fferent aspecanager answysed every snd Learning r to that sect

anagement fthe requiremontainers of cument has bis a discretehere is generment is, therany other puonger be avas are documfiles and mayent managemmeans. Docuocuments nonisations to he set of teceve these papclude technoorative ways on‐technical ational memgement systeservices for tctive documetion of imagitechnologieement of the

y, documentack of informprocesses. Kemation, to disation, managbeen more akflow, workinund out in th

an ideal examcessful organ

ge

106

cts of differewer to these nspecial requiprocesses). Tion for funda

eatures covements are foinformation.been unders collection orally a clear drefore, a humurposes. Wheailable (see eent in digitay not exist inment is used ument managow disperseddisseminatehnologies, super‐based doologies like wthat docummanagemenory. ems are the the electronent managemng, workflows, together wese documen

tation managmation. The aey characteristribute the gement and and more impng methods, his case study

mple from mnisation has

Part II – In

ent case procneeds, is sperement of stThe statisticsamental bac

er all kind of r KM system. They contatood as the f data kept tdifference beman creationen a documee.g. Bielawskl form. A dign a single parin so many wgement is a td throughoute informationuch as imagiocuments in workflow andents need tont issues that

products anic documentment includew, groupwarewith realisticnts are what

gement wasacquisition, ristics are theinformation,distribution portant for tsystems, proy.

manageable kalways mana

itial case stu

cesses, and hecified in thetudied proces of Phase 2 ckground info

content andms, not only din informatiolowest level together (savetween a nen. It is createent is destroyi & Boyle 19ital documerticular physways that cotechnology tt many organn to their intng and forma conveniend groupwareo be processet organisatio

d services thts located thes the use ofe, document standards‐ct make up eff

s defined as aretention, mae ability to m, and to allowof documenthe focal comograms, man

knowledge coaged its know

udies

how did Docu Phase 2 detss cases (Invdetails are kormation.

d therefore ‘kdocument mon in some sof filing. In aved) by the uw documentd for a particyed or lost th97). nt is often a ical form. onfusion reigthat is used tnisations. It iernal resours processingnt way, too. that managed within anons need to a

hat provide rroughout anf these typest managers, ocompliance afective docu

a systematicaintenance, sanage informw secure accnt‐based infompany, as thenuals etc. use

ontainers in wledge at so

umentum tails section. vention, kept clear

knowledge’ anagement.structured a word‐user. A t and a cular he

compound

ns when to manage is also a set rces, their g that allow

e both the address to

evision n s of optical and ment

c method for search and mation, to cess. The rmation. e amount of ed have not

ome level.

f

Systems

For mostoutside texchangdocumemanagemof the pecoordina

various rpersonnthat knorepresenfuture, o

using clieproductiare elusiquestion

enterprienhanceaspects, they oftecome int

memoryutilisatiotechnolotacit knoalso in thsupportiorganisacompara

Derived structureonly minmuch of indicate with cormanagemDocumecreationinformat

Ca2.5

There w

between inf

t enterprisesthe realm ofes), includinnted. Even wment systemeople who crated attempt

Large‐reasons. Theel results in owledge therntative at evorganisations

Duringent/server teivity and dative in multiven of how it al

Stratese, but technement, and fethey tend toen define theto play when

Documy. Organisatioon of tacit knogy businessowledge playhe utilisationing explicit kations. Intereable with the

Even tfrom experieed or unstrunor part is in f her/his timethat it is posrect IT tools.ment will haent managem and utilisatition system

ase II phases

ere two pha

1) Usein casroles,

formation an

s, vast amouf managed ITg knowledgewhere informm), much of treated the int to tap this ‐scale manage globalisatiothe distributre is no way tery point of s need to be g the past deechnology, ga accessibilitendor applicll fits togetheegic issues innically by neederated ano be much me rules and mn the selectioments are a vonal knowlednowledge wiles born duriys an importan of explicit knowledge mestingly the ce efficiency othough ‘explences (and inctured form structured fe just for seassible to incr. These threeve significan

ment will proion process. managing kn

s

ses in this re

ers’ perspecte organisatiovolumes etc

nd knowledg

nts of corpoT assets (e.g.,e never captumation is capthe value to tnformation ounrealised pgement of oron of organistion of the oto ensure thbest practicegood in re‐uecades, mangraphical usety. Increasingcation enviroer. clude compaw protocols)d virtual bus

more overreamethods by won of organisvery distinctdge is traditill be strengthng the past 5ant role. Howknowledge. Dmanagement case organisaof utilising taicit’ is anothndustry surv(as written dform (as dataarching the inrease the effe conclusionnt impact on vide help esDocumentatnowledge of

esearch proje

tive. Planningon. The purpc. This was a

ge

107

rate knowle, in documenured in digitatured in a mthe organisaor have used potential for rganisationaations is onerganisation'sat business pes for the orgusing the disty enterpriseer interfaces gly, productionments. Dur

any internal ) through intsiness modelching in detewhich an orgsational docu example abonally in twoh for organis5 ‐ 10 years hwever, in gloDue to quickand transferation's efficiecit knowledgher type of knveys as well),documents, abases, exactnformation siciency of kns give evidenthe performpecially to thtion manageorganisation

ect:

g and findingpose was to fqualitative i

Part II – In

dge and organts on persoal form or evanaged formtion comes fit in their wsharing and l memory ane important rs knowledge,practices wilganisation. Ftributed knos have focus(GUIs) and rivity will lie iring the next

communicattranet and exls. Although ermining howganisation is ument manaout knowledo forms: tacisations in somhave originaobal business and global gr have not deency to utilisge. nowledge, th most of expdrawings or t records, etshe/he neednowledge disnce that the mance of worhe decision mement systemnal memory.

g out qualitafind out the enterview, wi

itial case stu

anisational mnal drives orven explicitlymat (e.g., datfrom the knoork. Documereuse. nd knowledgreason. Geog, and withoul be consisteFor being comowledge. sed on applicelational datn cross‐applt decades, en

tion (organisxtranet; costthese issuesw an organisrun. Both ofagement appdge containet and explicime extent. Eted from inns it is very imgrowth of peeveloped wese explicit kn

he distinctionplicit knowledother formlc.). Also, a tys in decisionstribution proimprovemenrk processes making phasem is a fundam

tive samplesexisting needith questions

udies

memory contr in fleeting Ey articulated tabase or docowledge andent managem

ge is importagraphic distrut active manent, effectivempetitive als

cation replacta to improvication synernterprises w

sationally in et control vs. r also have opsation will fuf these typesplications is mers in organist knowledgespecially thenovative teammportant to bersonnel, theell enough in nowledge wa

n goes even dge is in semess informatypical office making. Stuocesses signints in documand businese of the knowmental exam

s from the cuds, practicess like

tents lie E‐mail and cument d expertise ment is a

nt for ribution of nagement ofe, and so in the

cement e rgies, which ill face the

extended revenue perational nction, and s of issues made. sational e. The e new ms where be efficient e systems most s not

further. mi‐tion) and worker usesudies ificantly ment ss processes.wledge

mple about

urrent usage, processes

f

s

e

Systems

productsnumber

Se2.6

person iKnowledbasis forthe case

the variethat atteDocumeneeded. Manage

the caseIt was alsystems)

the like Salready bManageRetDOC,

howeverBasis Plu(http://w

between inf

• •

2) Sysorganwith tcomp

Cross‐

One asthe capsearchsolutio

How impNot impor0

How weNot well…0

Also, ts to discoverof systems.

electing docu

Projecnvolved in thdge Managemr this projectes, the resear

Lots oety of producended to theentum. It wasThe evaluatr / Lotus, Hu

It seeme organisationso possible t), an integrat

OtherSGML/XML pbeen made br, Sherpa/PIM, Omega, OD

Therer, the possibus (http://wwwww.eidon.d

formation an

What are thWhat are th

tem Perspecisation fromthe existing parison. This w

repository se

spect that seppability to seaing has becomons.

portant is crortant…

1

ell Document …

1

there was a sr which requ

ument mana

ct was plannhis project. Tment, Organt. This case strch on the suof preliminarycts (see e.g. e research bys so basic KMtion concept ummingbird /med that then requiremeto compare ated documer systems weproducts webetween DocMS, MetaphDIN, Lotus No are plenty ole systems foww.opentextde/), Sigma L

nd knowledg

he processehe needs fo

ctive. Analys the featurepractice. Thewas a quant

arching

parates producarch across mume a core bac

oss‐repository 2

management

2

simultaneouirements are

agement sys

ed to cover tThe case orgaisational Metudy was plaubject continy system evaBroadhurst &y giving the liM system thawas adaptiv/ PC‐Docs / Fe most practents, and usea separate dnt managemere studied frre studied alcumentum aase II, Relatiotes, ClearCaof XML tools or this reseat.com/produLink (http://w

ge

108

es of documr delivering

ing and derivs of documee needs of caitative quest

cts targeted toultiple serverskbone service

y searching fo 3

t system X is s 3

us possibility e met. It was

stems and a

the two scopanisation’s Demory togethanned and thnued in the baluation was& Hedley 19icenses for eat it covered ve to other pFulcrum or Xical evaluatioe possibly Lotocument ma

ment system rom papers. lso on separaand Panagononal Documase and Interalready avairch were as ucts/), Poet Cwww.step.de

Part II – In

entation in g documenta

ving importaent managemase organisattionnaire, wi

o workgroupss, i.e. multiplee for enterpris

r your work in 4

supporting cro 4

for evaluatios possible to

project scop

pes of this prDocumentatiher with othehe results utibackground as done before97). The proevaluation. Tthe primaryroducts as w

XML content on was to setus Domino Danagement sand on‐line d Documentuate evaluatio, DocsOpen,ent Managerleaf extensioilable. For Dofollows: AstoCMS (http://we/), TC‐Toolb

itial case stu

case organation?

ant requiremment systemstion were theth questions

s from those me repositories.se document m

n case organis 5

oss‐repository 5

on (with full repeat phas

pe

roject. The reon Managemer company'lised during after the three the systemduct supplieThe negotiatefeatures, bu

well, like Lotumanagemenee whether DDocument Msystem (like odocumentatum, Matrix Oon projects. I LiveLink, Bar, Optegra, Pons. ocumentatiooria (http://wwww.poet.cbox (http://w

udies

nisation?

ments of cases and compae basis for ths, such as

meant for the . Cross‐repositmanagement

sation in the f ……very imp6

y searching? ……very w6

support) to e two with u

esearcher wament was an's specific wothe project, ee years casems were selecer was one sted product wut comparisous Domino Dnt products. DocumentumManager as cother Lotus dion managemOne (PDM), AInitial compaasis and also Parlance Doc

on Managemwww.chrystaom/), SGML/www.ovidius

aring them he

enterprise is tory software

uture? portant7

well7

pilot the unlimited

as the only alysed usingork as a but as all e study. cted from takeholder was on was ocument

m suits to omparison. domino ment. Astoria and arison has OpenDATA c.Manager,

ment, al.com/), /XML‐base .com/),

g

Systems

BladeRuis using I

comparemanagemand maradditionbuilt upoarchitect

managemLeveragiDocumemanagemand Notedistributand Domenterprithe Notecollabordocumeimaging,workflow2 discove

Ca2.7

what is hprocess

First the123

4

5 All the inthe numquestion

• osp

• Emea

between inf

nner (DocumInformation

Docued. Documenment and enrkets a platfo to a series oon their prodting and rep

Lotus ment applicaing Lotus Doent Manager ment, to usees users worted documenmino as theirse. Lotus Does/Domino pative documnt and outpu, OnDemandw. The indiviered how im

ase II, phase

Intervhis / her worand 5) Docu

following qu1. What are2. What are3. How long

Manager4. What are

organisat5. What kin

nterviews wember of quotans were thatThere wereorganisatiosome respoprocess. (BaEven in the managemeevaluation,answers.)

formation an

mentation MManager SBmentum andntum has emnterprise conorm for buildof packaged duct and targositioning ofDomino Docation built onmino server provides disers via Web brldwide, Lotunt managemr infrastructumino Documplatform. Lotment manageut managemd for computdual EDMSu

mportant the

1: users’ pe

views capturerk, 3) experiementation e

uestions were your currene your curreng have you b) e your work ption? ds of docum

ere recordedations is incr: barely any on managemondents felt ased on mamanageme

ent, e.g. proj employee’s

nd knowledg

anagement BML/XML datd Lotus Dommerged as onntent managding documeWeb applicaget specific vf core produccument Mann Lotus's Domtechnology stributed docbrowser clienus Domino Dent. It allowure to easily ament Manageus and IBM ament compo

ment (IDOM) er output toite componefeatures felt

erspective –

ed first the bence with doentities on th

re asked: nt job and tant projects aneen doing w

processes an

mentation ent

d, written doeasing from

documentament (Evolulike being r

anagers’ ansent level theject portfolis evaluation

ge

109

System, Destabase by Teino Documee of the top ement softwnt managemations knownvertical markct platform.nager (formemino and Noand its securcument mannts or Lotus ocument Mas organisatioadd documeer is a platfoalso positiononent of IBMsoftware tha laser disk (Cents are tiedt, but the ph

results from

background iocumentationhe case organ

sk functions nd duties in y

work with doc

nd business r

tities there a

wn and analcase I to cas

ation manation insteadregularly ovswers.) ere were proio, contractn, Job applic

Part II – In

sk Top Publisexcel) (http:/ent Managemsuppliers in

ware market.ment applicatn as the Innokets. Docume

rly called Lototes messagirity and replnagement capNotes clientsanager repreons that haveent managemrm‐focused n Lotus DomiM's EDMSuiteat also includCOLD) applica together thhase 1 was in

m the intervie

nformation:n (as a user, nisation leve

in the case oyour departmcumentation

rules that gu

are in the cas

ysed. (To illuse III).The co

gement prod of design)ver employe

ocesses thats, patent prcations, all

itial case stu

shing, Schem//www.interlment were ththe integrat Documentutions and corovation Applientum repre

tus Domino.ing and grouication featupabilities, incs. For the miesents a leade made a heament functionsolution; it hino Documene, a portfoliodes ImagePluations, and Mrough Contendependent o

ews

1) Who we administratol.

organisationment or labon? (As an Adm

ide you in yo

se organisati

ustrate the pnclusions fro

ocesses form). Evolved syed, forced to

at can utiliserocesses, insorts of form

udies

ma and conveeaf.com/). he main prodted documenum companyrporate portication Seriesesents a majo

doc) is a docpware platfoures, Lotus Dcluding lifecyillions of Lotuding solution avy investmenality throughas no relevant Manager o of integrateus for producMQSeries Went Connect. on systems.

are interviewor or manage

? (‘Who’) oratory? (‘Wministrator, U

our work in t

ion?

possibilities oom the first s

mulated in lystems haveo be a 'bottl

e documentvention repms… (Based

erter, which

ducts nt y develops als, in s that are or re‐

cument orm. omino ycle us Domino for ent in Notes ghout the ance beyondas the ed ction orkflow for The phase

wing, 2) er) 4) the

ork’) User,

he case

of reporting, set of

ine e got gaps: leneck' in a

t ports, self‐d on users’



110

• Every project could utilise document management, but every project decided work processes, systems and tools. (Based on Administrators’ answers.)

So this field requires more work. But the processes were still to cover up. The different organisational units are dealing with following documentation entities:

• There were corporation level processes (comparable with Case I) like Product development, Product delivery processes, Administration; Project handbook including R & D process, various tools, patent databases, articles from external sources, project portfolio, contract archive. (Managers)

• The other organisational unit was dealing with instructions, specifications, feature lists, program code and also contracts while the third organisational unit was dealing with projects in two systems. (Users)

• The third organisational unit covered Knowledge bases like Lessons learnt and Best practices, e.g. a very large Knowledge Asset ‐project on System Integration (co‐brain). These are useful project auditing methods also for the whole focal company. (Administrators)

The second set of questions took a closer look to processes, information search and distribution. The intention was to find out the processes of documentation, needs for delivery, how to search for information and finally basic figures on reading vs. writing, modifying and updating documents. The actual questions were:

6. What are the processes of documentation like in your work? 7. What are the needs for delivering documentation? 8. How are you searching for information? 9. How many documents you are reading per day? 10. What amount of documents you are writing or updating per day?

Every set of questions generated detailed findings, where from a summary was processed. The main conclusions concerning processes are as follows:

• Processes were depending on projects (Users), but • There were processes common to each and every project (Managers) • Software source code management done together with other documentation

management formed a narrow process line. Documentation management was needed independently from software source code management. (Users)

The administrators shared a common view based the following findings:

• Also e‐mail was a wrong solution for documentation processes. • Version management was not working. • If file servers are used, group directories must to be specified strictly in the process.

About the delivery we found out that:

• Only the information about updates and links to the documents should be delivered (Managers)

• E‐mail was (and is) a wrong solution also for delivering documents; • What was the use of Intranet? Only delivery. (Managers) • One example from strict delivery and access control was Portfolio management.

(Managers) • The same Portfolio management came up as a result also with users, as an example

from fixed workflow. (Users)



111

• More experienced users were not satisfied with Lotus TeamRoom, or other Lotus applications. (Users)

• Information in organisational unit (not to mention the whole focal company) was not even in the 1st normal form. (Users). In practice this meant redundancy.

• Duplicates and version discrepancy was a real problem. Documents must be unique, no fallacious copies (This is a users’ perspective to normalisation).

• Articles through library and other external sources were also much used, not to mention the collaborating projects of the extended enterprise.(Users)

On information searching the conclusions are:

• Documents in secured systems wasted a lot of effort. (Managers) • Some users considered structure of intranet pages more important than search engines.

(Managers) • Directory and file names must be still logical, even though search tools are used.

(Managers) • Search by keywords (with search engines) can be misleading. (Managers) • Meta‐knowledge about persons expertise was also used (Managers), but • Public competencies' database will be the place to store human resource meta‐

knowledge (Users), and • 'Search by phone' within a well‐defined process was a symptom from unmanaged

documentation workflow. (Users) • Mail was a place to store and search for documents (Users) • Corporation global search engine was ‐ on that time ‐ not working properly.

(Administrators) (The engines have been replaced every now and then.) • There must be a system where 1) keywords can be used and 2) document content can be

indexed. (Administrators) The e‐mail statistics gave implications that:

• Mail is used too much, for wrong purposes • Answers made suspicious about overestimation, because

the amount of mail seems to be a sort of prestige among workers. • Vice versa, the culture should be that load of futile e‐mail is not a merit.

The third set of questions analysed needs and problems, requirements for own work, as well as on organisational level. Questions asked:

11. What kind of documentation you are dealing with? 12. What are your needs for documentation in your current job? 13. What are needs for documentation for the whole case organisation? 14. What are problems concerning documentation in your current job? 15. What are problems concerning documentation for the whole case organisation?

The conclusions based on the question nr. 11 are:

• Documents needed different kinds of access control (confidential documents) • There were a big difference between deliverables and working papers (active

documents) • Most of documentation was free form (Managers) i.e. unstructured. • Templates were rarely used. • Invoices were sent with Excel sheets from abroad (indications of an extended enterprise)

(Users)



112

• Knowledge bases (e.g. TechNet) were independent entities. (Administrators) The respondents’ own work specific needs are summarised in the following list:

• There was a dilemma how to start the system, before there was content (ostensible problem, pretext)

• Different kinds of documents should be stored with a single tool. • Practice for delivery was critical. • Easy access was needed. (Managers) • When a project ended, knowledge (shared in that project) disappeared. • There were strict workflows, like the contracting process. (Users) • Lifecycle and workflow were necessary on some level. (Administrators) • Patent practices and project portfolios were suffering from paper‐based documents,

which was another extended organisation specific problem. The needs on focal company level are as follows:

• The project result knowledge delivery was not working because of the poor applications. (Managers)

• The case organisation needed an easy access to shared documents on focal company level. (Users)

• Extended (e.g. contracts and patenting) processes without paper on the case organisation level was needed. (Users, Administrators)

• Document access between different sites worked at most inside one country. (Administrators)

The summary based on problems of an individual’s work:

• Managers did not actually use the systems — their ‘visions’ were already on the market. (Managers)

• Too many systems were used for the same purposes. (Managers) • Notes, e‐mail and intranet had many disadvantages. (Managers) • Different systems forced to store redundant copies. (Users) • Version control was needed, history could be relevant. (Users) • Mailing (at least between two persons) was a wrong solution. If there were more editors

a document management system was the (only) solution. (Users) However, this depends from the knowledge maturity, or the phase of document lifecycle.

• Strict workflow was not the solution for 'anarchistic model' (Users), but workflow was needed e.g. for contracts. (Users)

• Notes links are a (small) problem. (Administrators) Focal company wide problems:

• Access and delivery in general (e.g. Strategy, Portfolio, project documentation) (Managers)

• Project information vanished after the project ended. (Managers) • Administrators thought that it was difficult to manage the user specific information…

(Administrators) — and users thought that it was complicated to request system management from administrators. (Users)

• Intranet tools were not yet known. (Users) • Uncontrolled versions from documents (caused by e‐mail and multiple stores) rose a big

problem (Administrators)



113

• Paper documents, e.g. invention reports on paper were a problem. (Administrators) The last set of questions was about current tools and future. The questions were to find out what kind of information is included, what tools are used as editors and document management, how to enhance documentation management and any other comments.

16. What kind of information is included in your documents? • Many documents could be formalised: e.g. from human resources management, project

plans, databases, portfolio brochures and contracts, quality handbook. (Managers) • Handbook, and obviously manuals and instructions could be split into modules.

(Managers, Administrators). More generic modularity will be discussed in the results – here modular documents are referred to as compound documents.

• Lifecycles for projects documentation was possible. (Managers) • The implementation of documentation management system could refer to personal,

project and line organisation levels. Project level was the most important. (Users) • There was no version history with Intranet pages.(Users) • Meta‐information, e.g. Access levels needed for inventions. (Administrators)

17. Which products you are using for documentation (e.g. Word, PowerPoint, Excel and Outlook,

but also other than Office products: Lotus Notes, html‐editors, project management tools, etc.)

• Everybody was using Office tools and Outlook. Therefore document management should be compatible.

• There were plenty of other tools. It must be divided, which kind of files are managed with document management system. But the information written on documents should be independent from the tools.

• Some parallel administrative systems e.g. Projects were used. (Managers)

18. Are you using any systems or tools for documentation management (e.g. Documentum, Lotus Domino Document Manager, or PC‐Docs…)?

• Working TeamRoom, Lotus Notes (Domino) applications were from a very narrow sector. • Even Notes developers criticised the tool ‐ not to mention some users. However, this is

mostly a question of habit. • Some other (PC docs/master) systems were used as ‘relicts’ (Administrators)

19. How the documentation could be managed better in the focal company?

• Easy access was important for all (version management is more important for other staff than managers) (Managers)

• Project documentation needed instructions and templates. (Managers) • Web and e‐mail co‐operation were needed from the document management system.

(Users) • Research topics could be published on personal and project level. (Users) • Lessons Learnt from projects must be published (Administrators)

Based on the last comments it can be summarised that one system for the whole focal

company is needed. (Later it was realised that it is a question of storage platform, and not the client application.) The managers were not so interested in piloting documentation management systems, but some experienced users wanted to pilot Documentum. However, they ended up using version control and own configuration management system – they store software code and documents in the same

Systems

system, used in p

Ca2.8qu

importantook alsoselected

Administintegratebased onframewosubject w

importanselected

between inf

and they arepractice all th

ase II, phaseuestionnaire

This sence of featuro a closer loo to present d

The acproce• Le• Ad

co• Pr• Le

ex• In There

trators and Med documenn system evaork was usedwere investig

This sence of system and targete

1 Pe

2 Sys2.1all2.2all2.3all2.4all 3 Ar3.1a 3.1m 3.1u 4 Se4.1u 4.1m 4.1a 4.2u 4.2m

formation an

e using softwhe time duri

2: system pes

ection preseres – how thok to the orgdifferent kin

ctual organisss: egal process dministrativeompany rojects – varearning procxternal trainenventive proc

was the alreManagers oft managemealuation framd later on casgated (e.g. frecond phasem features foed for differe

ersonal bac

stem Backgl Docl Docl Coml Obj

rchitecture UN

WW All

arch and ReSeaCro

ComSea

Filt

nd knowledg

ware source cng the resea

perspective –

ents results fre systems caganisational pds of conten

sational unit

– Contracts:e process – H

ious collaboress – HR Devees cess – Invent

eady definedf all processeent softwaremework (Bolose III as well rom Verkasae of case II coor the case oent groups ba

ckground inf

ground Inforcumentatiocument mampound dojects as attr

NIX and WinWW serverstypes of un

etrieval arch from aoss‐repositompound seaarch results tering

ge

114

code managearch.

– results from

rom structuran fulfil the uprocess (casent — from ad

s were in the

: strictly confHuman resou

ration projecvelopment: m

tion and Pate

d dimension es. Questionse (Shegda 199oix & Robilla— and lots oalo 1997 to ponsisted of aorganisationased on phas

formation ‐

rmation on managemanagement scument maributes

dows clients and thin clstructured

ttributes anory searchinarch abilityrepresenta

Part II – In

ement anyw

m structured

red numeric users’ require organisatiodministrative

e same orde

fidential prources with co

cts in internamore about t

ent Reportin

available: ans were derive99 & 2000), rd 1995) – foof articles anapers and fie questionnaprocesses. Tse 1 findings

‐ name and

ment in gensystem is inanagement s

ts lients information

nd contents g

ation

itial case stu

way. The syste

d numeric ev

evaluation qements. Thison) dimensioe to more ‘cre

r, based on t

cess in an exonfidential le

al project mateaching and

ng in an exten

nswers were ed from the the comparior the continnd studies exeld observatire and interThe followings.

selected pr

eral ntegrated system

n

simultaneo

udies

ems evaluate

valuation

questionnairs second phaon. The proceeative’.

the formality

xtended enteegal matters

anagement d collaborati

nded enterp

gathered froperspectivesson of systenuity this samxplicitly on thtions until 20rviews aboutg questions w

rocess or pro

ously

ed were

es on ase of case IIesses were

y of the

erprise inside focal

on with

rise

om Users, s of ms was me he same 008). t the were

oject



115

5 Library Services 5.1u View checked‐out documents 5.1a Manual version numbers 5.1m Working and public versions 6 Control 6.1u Notifications 6.1a Audit Trails 6.1m Security levels 6.2a OS security features 7all Workflow and Lifecycle 7.1all Workflow 7.2all Lifecycle 8u Customized applications 8a Standards 8m Business Approach 8.1m Price of the software 8.2m Vendors of document management systems 9all Other requirements Most of the questionnaires were filled during an interview. There are statistical

representations from the answers to clarify the importance of the features of document management systems. The arithmetic mean is presented as an average value of observations and it tells how big the observations should be if every observation is equal and the feature described by these values would be still the same as it was originally. The geometric and harmonic mean are not needed in this presentation.

Using this classification, the arithmetic mean for a discrete quantitative variable X was

counted with a basic formula xn

f xi ii

A

==∑1

1

In the tables there are following kind of classification used:

0…25 %

Not important feature

> 25…50 % Slightly important feature

> 50…75 %

Important feature

75…100 % Very important feature

The formula for variance (s2) used here is sx x

n

ii

n

2

2

1=−

=∑ ( )

and the standard deviation

(s) used is simply the square root of s2. The divider is n (instead of n‐1) in the following presentations, because we are measuring the whole population and not a sample. And of course, the lower s is, the better the mean is as a summary measure.

Graphics describe the averages and standard deviation in one glance. However, it should be emphasised that this is not a statistical analysis mainly because of the number of the respondents (total 36) – number is not included in the graphics. The questions were about the future, but the current



116

opinion was measured. Therefore the questions were formulated as: how important is – for your process in the future?

2.8.1 Common questions

There were five questions for every respondent group. Questions are either general or divide the products already in the first steps. They belong either to series 2 System background Information or to series 7 Workflow and lifecycle. Here are presented brief summary from the most interesting findings.

Document management systems in general (2.1all) ‐ System background

Documentation management system, also sometimes referred to as document repository management, is a type of middleware service for organising electronic documents, managing content, enabling secure access to documents and unstructured data, routing documents and automating related tasks, and facilitating document distribution. Core document management services functionality includes library services (check‐in/‐out, version control, and document‐level security), cross‐repository searching, common system administration, and unified system management. Recent product introductions point to an increased leverage of Web technologies and standards (HTML, XML, WebDAV13), the addition of Web publishing and enhanced search and retrieval capabilities, and support for compound documents and document component management. Content and Knowledge Management applications are the new drivers for system vendors such as Documentum, FileNET, Open Text, and PC DOCS (now the Enterprise Knowledge Portals unit of Hummingbird Communications).

Document management system in general was very important (importance over 75 %) for

four processes out of five. Inventive process was the only one in the 'important' category (under 75 %) though their deviation was the greatest (21 %).

Inventive process managers told that documents can be managed well enough with file servers and directories while the administrators and executive managers (as found out in the first phase) said that the present Lotus Notes application (IMS) meets most of the needs.

The best practices are not clear‐cut. The Inventive process users had got very great deviation (29 %), and the lowest ranking user commented that document management system is inherent in the document database but should be available if selected to be used for other work.

Projects administrators considered the overall importance lowest, but one of them commented that ‘a need for document systems will increase in future’. Contracts user commented: ‘It's good to know contracts are signed by business units, but I hardly ever get queries from project managers.’ That is a detail from the process. The contracts managers, however, were also interviewed in the Phase 1, and they represent very deep experience from the focal company and also practice from outside. Learning process was top ranking with an extremely low deviation (7 %). Administrative Process and Inventive Process managers gave lower points than about the importance in general.

13 The two previously mentioned standards, WebDAV and XML, could not be omitted, although they were not especially investigated in this research. XML was in good practice already in the focal company. WebDAV was crucial to include in the beginning of the research, as the protocol consists of a set of new methods and headers for use in HTTP, and the two protocols were related ‐ WebDAV as the first protocol ever to use XML. It was also supported by industry leaders: Microsoft, Netscape, Novell, Xerox, FileNET, Documentum, PC Docs and others, and in shipping products: especially from Microsoft ‐ IIS, Office and Windows.



117

Figure 19. Document management systems in general

Integrated vs. non‐integrated document management system. (2.2all) ‐ system background

Integrated documentation management software typically provides a document repository and the means to search for and find documents stored in it. Documents enter the repository in one of two ways. When a user creates a word processing document or a spreadsheet and selects Save, a profile screen supplied by the document management application appears. The user enters attribute information about the document (called meta‐data), such as title, subject, document type, etc., while other information such as author, date created, version number, and the document's physical location is automatically captured by the document management system. The attribute information is typically stored in a relational database. Documents may also enter the system via import utilities or through optical character recognition (OCR) performed on scanned documents. Text documents are typically submitted to a full‐text indexing engine, with the resulting file sent to a full‐text database. The combination of attribute indexing and full‐text indexing allows greater searching capabilities. Compare two cases a) Integrated document management. You are writing a Word document and you are able to store it to document management system straight from Word.

Document Management system in general (2.1 All)

64%

90%

78%

86%82%

0%

25%

50%

75%

100%

InventionProcess

LearningProcess

Projects AdministrativeProcess

Legal Process

Process

Impo

rtan

ce

AVG by ProcessManagersAdministratorsUsersTOT. AVGSTD DEV P. (%)



118

b) Not integrated document management. You are writing a Word document and you should start a document management application separately to store the document into the document management system.

2.2all How important it is that document management system is integrated for your process in the future?

Integration was important, but not very important (under 75 %) on an average. Invention process was the only one in the 'slightly important' category though their deviation is the greatest (27 %), while for projects integration was important and for all others very important – legal process was on the limit (75 %) but with smallest deviation (6 %).

Invention process was an exception (‐29 %). Their manager said that ‘the documents to be stored are generated mostly by outside associates’ (in an extended enterprise). Their administrator and executive manager (as found out in the first phase) said that all correspondence goes to patent attorneys, but a centralized document management system would serve transparency and process objectives. The Invention process users had got very great deviation (36 %), and the lowest ranking user commented that ‘an integrated system will be painful’.

For projects managers the deviation was also high (over 21 %) and the lowest ranking manager commented that ‘the only important thing is that the system works properly’. Legal process representatives were on the same tracks, and an administrator commented that ‘document management program should be indistinguishable, not disrupting the use of other applications or tools’, while the user said that ‘integration will reduce workload to the person storing documents with separate systems now’.

Integration was very important for both Learning and Administrative processes.



119

Figure 20. System integration

System evaluation scored that Documentum integration was two times better than Lotus Domino Document Manager. DominoDoc integrates only with the Office Applications from Lotus and Microsoft.

Compound vs. non‐compound document management system (2.3all) ‐ system background

Compound document management systems allow objects to be assembled into pages and pages into documents. The challenge is to track the links between each element and the documents in which it is used. In compound document managers, the items entering the repository, the attribute database, and the full‐text database may be objects, i.e. elements such as a single paragraph, drawing, or chart. Compound documents are also known as virtual documents.

2.3.all How important it is that documents can be assembled from independent elements for your process in the future?

Compound document management system deviated most from the answers from everyone ‐ deviation is over 28 % from the whole population. The average importance goes therefore down to 61 %.

DM system is integrated (2.2All)

43%

90%

71%

86%

75%

0%

25%

50%

75%

100%

Inventi

on Proc

ess

Learn

ing Proce

ss

Projec

ts

Admini

strati

ve Proce

ss

Lega

l Proc

ess

Process

Impo

rtan

ce

AVG by ProcessManagersAdministratorsUsersTOT. AVGSTD DEV P. (%)



120

Compound documents were important, but not very important (under 75 %) on an average. Again, Invention process was the only one in the 'slightly important' category and their deviation was very high (over 40 %), while the only 'very important' ranking group was Administrative process.

Invention process manager said that it was possible to cut and paste documents even if they were not compound. First time the administrator thinks the same. The ultimate disagreement (deviation maximum 50 %) came from one user, who stated that Compound System could be valuable because of the always present unique keys, like patent numbers and application numbers and invention report numbers. At that moment Invention process was using a handmade system for 15 000 documents and required attaching a prefix to the title before filing.

Invention process and administrative process system administrators were in the opposite ends. Legal process administrator pointed out that ‘compound documents can be especially contract or process guides, not necessarily the actual contracts’. It was one challenge how to make dynamic content legal – for example, some agreements were still requiring autographs. Learning process administrator said that independent elements are a question of education: ‘It demands from users the ability to understand this kind of process’.

For administrative process this was a 30 % exception from the average.

Figure 21. Compound system

System evaluation scored six out of seven to Documentum and zero for Lotus Domino Document Manager when comparing how documents can be assembled from independent elements. This is simply because Lotus Domino Document Manager did not support Compound documents.

This was an essential limitation in Lotus Domino Document Manager, and made a clear difference between the products. It made also a big difference between processes (Administrative 90 % and Invention process 46 %) and even within Invention process (users 86 % and administrator 0 %).

Compound DM system (2.3 All)

46%

57%

90%

54%

67%

0%

25%

50%

75%

100%

Invention Process Learning Process Projects AdministrativeProcess

Legal Process

Process

Impo

rtan

ce

AVG by Process

Managers

Administrators

Users

TOT. AVG

STD DEV P. (%)



121

Workflow (7.1all)

Most integrated document management software includes some workflow functionality. At a minimum, document versions can be sent along predefined routes for review or approval. At the higher end is workflow that includes task assembly, task assignment, and task tracking to completion. Library managers generally contain ad hoc workflow features such as routing and e‐mail notification, but may permit integration of third‐party workflow packages when greater processing and control features are required. Compound document managers also vary in the degree of workflow functionality they contain, although those with roots in the publishing industry tend to have workflow specific to the editorial process (e.g., Xyvision's Parlance Document Manager).

7.1.all How important is workflow functionality for your process in the future?

Workflow was very important on average (approx. 80 %). Projects were the only group that scored it to the 'important' category though their deviation is the greatest (28 %). The best possible case was legal process ‐ their managers thought workflow is definitely important, and the user probably was thinking just her role in the current practice, and not the future process as an entity.

Invention process manager said that ‘workflow is not important from patent application point of view, but from invention report point of view it is’. The administrator said that ‘ad hoc workflow is important’. A user said that ‘workflow is presently only implemented to a small degree in Lotus Notes but not at all in the main process workflow’. He commented also that ‘workflow is most welcome’.

All inventive process users thought that workflow is very important, while the deviation among projects managers was very high (nearly 36 %). The lowest ranking manager stated that ‘personal contacts are more important’. Workflow did not directly affect to personal contacts one way or another.

More reasonable comment came from one legal process manager, who put a strict condition to the high importance of workflow: ‘Contract acceptance requires electronic signature, otherwise workflow is not needed’. Another important condition was whether the document is accessible corporation wide or not (access rights for signatories).

Administrative process manager commented that global employment management system was already including a workflow, while there was the current (R/3) system used as a portal. According to the administrator it was very important, that the ‘documents flow from one person to another quickly and reliably — especially the documents that need to be accepted’ (e.g. job requirement)’. Their administrative user gave an example from Salary and Employment Notifications (PTI in Finnish).



122

Figure 22. Workflow

The workflow functionality of the systems evaluated scored 1 : 2 for Documentum’s benefit.

Lifecycle

A typical document goes through a number of states during its existence within a business. To provide a user with structure to move a document from one state to another state integrated document management system provides a document lifecycle and uses the promote and demote commands. Any appropriate user can apply a document lifecycle to a document, or promote the document from one state to another. As the document lifecycle progresses, user may have to perform other actions on the document. Typically, a document lifecycle will be incorporated into a workflow, and user will be alerted role in a document lifecycle when a workflow task appears. Simplified document lifecycle could be like:

Figure 23. Simplified document lifecycle14

14 Workflow, lifecycle and other features can be utilised in revision.

Workflow (7.1 All)

86% 86%

63%

95%

86%

0%

25%

50%

75%

100%

InventionProcess

LearningProcess


Legal Process

Process

Impo

rtan

ceAVG by Process

Managers

Administrators

Users

TOT. AVG

STD DEV P. (%)



123

Enhanced lifecycle services manage the relationships between content, business processes, and user roles. (Enhanced lifecycle could be like the one presented later in Figure 28). These services automate how content and its associated processes relate to one another and automatically promote a document or object through the stages of its lifecycle. Each object has a default lifecycle. Using drag‐and‐drop editors, developers define the relationships and business rules for each stage of the content's lifecycle. Lifecycle definitions can be applied to different types of content and are reusable. Lifecycle services also enable the application of different levels of security for each step of the process.

Lifecycle belonged still to the very important category (77 %). ‘Projects’ was the only process under the 'very important' limit and this time their deviation was not the greatest. The legal process continued the workflow answers — their managers thought lifecycle was definitely important, and the user probably was thinking just her role in the current practice and not the future process as an entity.

One interesting thing was that even though the correlation with workflow and lifecycle is very high among administrators and users, the Invention process manager thought workflow is important, while lifecycle is not. For lifecycle, the contradiction between Invention process manager and other members of the group was huge (deviation over 27 %). Invention process administrator said that ‘lifecycle is an in‐built functionality of current (Notes) application (IMS), and not of great importance outside IMS’. A user said that ‘lifecycle is for some Invention process documents very convenient.’ Users' average was 93, managers’ 29.

One legal process manager commented that ‘lifecycle is the basis for Workflow, and the contracts are going through different phases like drafting, commenting, freezing and signing.’

Lifecycle and workflow were the most important for administrative process. The administrator emphasised especially the case of accepted job requirements and related official (PTI) forms: The manager sees if the job advertisement is already in the web.



124

Figure 24. Lifecycle

The workflow and lifecycle correlations could be calculated, while it should be noted that this case study is mixing qualitative analysis and just presenting the evaluations on graphical format.

System evaluation scored how well the requirement was fulfilled by the products, ranking Documentum twice as good as Lotus Domino Document Manager (6:3).

2.8.2 Very important features

In this subsection the most important features ‐ independent from the processes – are analysed. Nine very important features were chosen based on the over 75 % average on the scale. After each feature, there is a table about the system evaluation, where the cumulative difference between Documentum and Lotus Domino Document Manager is counted. Figure 25 refers to the number of the feature in question, e.g. Lifecycle is 7.2 All. The rest of the features were analysed in a similar manner, but this final report focuses on new findings.

Lifecycle (7.2 All)

75% 76%

65%

90% 89%

0%

25%

50%

75%

100%

InventionProcess

LearningProcess


Legal Process

Process

Impo

rtanc

e

AVG by Process

Managers

Administrators

Users

TOT. AVG

STD DEV P. (%)

Systems

3.1.1a

3.1.2a

clients. Alowest, u

of Unix cimportanand excea negativ

the findithe oper

on the edifferent

between inf

Windows C

TradiclientinterfdatabMicroplatfomanasuppo How iyour p

How iproces

The mAll except onunder 7 %.

Howeclients was rant. This was eptionally higve exception

One fiing is more trating system

Even tvaluated syst roles (admi

formation an

Fi

Clients (3.1.1

tionally, intt/server moface residesbase or suposoft SQL Seorm for dagement prort Window

important it process in the

important it ss in the futu

most importane administra

ever, Unix clieanked as 43 exceptionallgh for Projecn for the Legainding is thato direction, wm. though Windstems, nor oninistrator, us

nd knowledg

igure 25. Very

1a) – Archite

tegrated dodel. Core fus on netwoport populaerver. Microdocument roducts supws and occas

is that docume future?

is that documure?

nt feature wator ranked W

ents’ deviatio% average. Oly low for Invcts (+43 %) aal process ‐2t projects vawhere the d

dows clients n the answeser, manager

ge

125

y important D

ecture

ocument munctionality orked clientar SQL dataosoft Windomanageme

pport a varisionally UN

ment manag

ment manag

was that docuWindows clie

on was on thOnly the datvention procnd Administ29 %. ry much comocument ma

was ranked rs from the dr). This is not

Part II – In

DM system fea

managementresides on ts. Productabases suchows NT hasent solutioiety of UNIXIX clients.

gement syste

gement syste

ument manaents as 100 %

he opposite eabase objectess and Learrative proce

mpared with anagement s

very importdifferent prot considered

itial case stu

atures

t software a server or ts may comh as Oracle, s become thons, althouX servers a

em will suppo

em will suppo

gement syst% must. The

end (nearly 3ts stored as arning processss (+29 %). O

more fixed psystem is tot

ant, it makesocesses (case critical in th

udies

has been bservers, whme with a pInformix, S

he predominugh some as well. Mos

ort Windows

ort UNIX clien

tem supportsdeviation w

39 %). The imattributes was (‐43 % fromOn the contr

processes. Inally indepen

s no distinctie organisatiohe results.

based on ahile the userproprietarySybase, andnant serverdocument

st products

s clients for

nts for your

s Windows was also the

mportance as less m average) ary this was

n general, dent from

ion, neither n) or on the

a r y d r t s



126

Web servers and thin clients (3.1m) – architecture

The Internet has had a significant impact on document management technology. Web capabilities have become essential for traditional document management software products, and new products have emerged that are designed from the ground up to leverage the Internet infrastructure. Such products are built around three‐tier architectures supporting Web servers and thin clients that provide access through Web browsers. The document management architecture will continue to move toward a multi‐tier (or n‐tier) client/server architecture. Going forward, browser‐based interfaces and Internet protocols will become the primary means of providing document management functionality.

Q3.1m How important it is that document management system supports web servers and thin clients for your process in the future?

The second very important feature was that document management system supports Web servers and thin clients. On an average it was ranked with nearly 88 % importance with only 12 % deviation. Invention process manager commented that the system was easier to deploy, since the users are spread world‐wide in different organisations. One Legal process manager said that ‘it is good if you don't have to install new system’. (That means the client side). Another Legal process manager commented that this is very important rather from the generic (extended organisational) viewpoint than narrowly in his case organisation process.

For the Learning process this is a ‐16 % exception from the average. It is the only very important feature, which the case organisation didn't respect over 75 %. The system evaluation did not find any crucial differences on web servers or thin clients between the applications researched.

Search results display options (4.2u) – document list: search and retrieval

Search results can be displayed as a document list, as a list with documents displayed, or as a pictorial representation showing how each document is related to others. Most document management products also deliver the search results list in order of the documents' relevance to the search term(s), known as relevancy ranking.

Q4.2u Please rank the importance of display options of search results for your process in the future. (7=most important, 0=no importance, many options can be marked with same number.)

Document list

List with document content

List with relevancy ranking

Representation about document relations

The third very important feature was that document management system shows search results (at least) as a document list (86 % average). The deviation was nearly 19 %, and a – 29 % negative exception between Legal process and Administrative process.



127

About list with document content, which is nearly on the limit of a very important feature (75 %), people were more unanimous. This was a positive exception for Projects and a negative exception (‐29 %) for Learning process and Legal process (‐32 %).

The relevancy ranking (71 %) was important according to Learning process, especially with web pages. (Such display was e.g. in the web pages of Helsinki University). It was exceptionally important also for Administrative process. For Legal process, this was a negative exception ‐29 %.

Representation of document relations (63 %) was another big exception for Learning process, +38 % difference above average. There was huge deviation in general, over 33 %, even within groups (see e.g. projects). Even individual respondents were unsure, as one Invention process user commented: ‘might be a useful feature or it might be of no use?’ It depends from the contents ‐ the information that is managed.

Figure 26. Search results representation

System evaluation told that this requirement was fulfilled with the products similarly. The other display options vary: for example representation about document relations is ranked 100 for Lotus Domino Document Manager vs. 0 % for Documentum. Also list with relevancy ranking was a combination of attributes and content in Lotus Domino Document Manager. We will return to search result display options in case III.

Working and public versions (5.1m) – library Services

Sophisticated products recognise versions and sub‐versions and can make a distinction between working versions for limited distribution versus public versions for wider delivery.

Q5.1m How important it is to make a distinction between working and public versions of the documents for your process in the future?

Distinction between working and public versions was very important (86 %) with small deviation (15 %). Even the project managers agreed from this.

Search results representation (4.2u)

0%

25%

50%

75%

100%

Invention process Learning process Projects Administrativeprocess

Legal process

Users, Document list

Users, List with content

Users, Relevancy ranking

Users, Doc. Relations

Usr, Doc. list STD DEV (G) %

Usr, List with cont. STD DEV G (%)

Usr, Rel. ranking STD DEV G (%)

Usr, Doc. rel. STD DEV G (%)

Usr, Doc. list AVG

Usr, List with cont. AVG (%)

Usr, Rel. ranking AVG (%)

Usr, Doc. rel. AVG



128

Invention process manager commented that e.g. patent applications older than 18 months will become public anyway, so those should become public inside the company as well. One legal process manager commented that after the contract is published, it should 'freeze'. The status of the contract is important to display: is it valid or not. It can affect to the visibility, but certainly to the locking of the contracts. Another legal process manager commented that there was a certain lifecycle (which relates to public and private versions): internal draft, draft for customer negotiations, and the version for signature, which results a frozen contract.

For administrative process this was a negative exception from average (‐29 %). Public versions are not needed. The flat result is that systems evaluation does not make any difference between the applications.

Audit trails (6.1a) – control

Audit trails are kept of who accessed which documents and what functions were performed.

Q6.1a How important are audit trails for your process in the future?

Audit trails were very important (83 %) with small deviation (15 %) — and not only for technical reasons. For business unit contracts any project plan and its progress needed version control and traceability.

Audit trails were a ‐26 % exception for Invention process. Another administrator (who understands the content) from Administrative process said that ‘It will be good to see which manager is searching for people through internal recruiting, which applications she/he has looked through, so Human Resources can more proactively take a part to teaching of the system and sees which kind of applicants every manager is looking for etc.’ This is the first example from one sort of ‘who‐knows‐what’ meta‐information in a system itself.

According to system evaluation, so far there was no difference between Documentum and Lotus Domino Document Manager. This audit trail, however, was a requirement for the system in case III – the meta‐knowledge popped up also during case III.

Workflow (7.2all) as a very important feature

Workflow functionality was next very important requirement for all. Invention process managers and legal process user didn't find it very important — and neither projects administrators, nor managers, nor users considered workflow very important. It is interesting that there was a difference between workflow (80 %) and lifecycle (77 %) even though the deviation is quite similar (24 and 22 %). We can look also the relations of workflow and lifecycle by the role:



129

Table 13. Workflow and lifecycle ‐ importance by role

Role Workflow Average

Workflow Deviation

Lifecycle Average

Lifecycle Deviation

Administrators 86 16 88 11

Users 77 27 73 23

Managers 78 30 71 29

ALL Average 80 24 77 22

Among administrators the lifecycle was even a little bit more important, and the importance went perfectly analogously from process to process between lifecycle and workflow. Systems evaluated fulfilled the workflow requirement: Documentum very well (6/7), Lotus Domino Document Manager not so well (3/7). Also the lifecycle management was better in Documentum (7/7), versus Lotus Domino Document Manager (5/7). This is the first feature ranked highly important for the focal company, where the products differ.

The Pearson's correlation coefficient between Workflow and Lifecycle was not ultimately analogous. These same indications as from the cross‐tabulations could be seen from a scattergram:

Figure 27. Workflow and lifecycle scattergram

Normally in a scattergram the horizontal axis represents the values of the independent variable and the vertical axis represents the values of dependent variable, but lifecycle and workflow are here as in the same role.

Workflow & Lifecycle

0%

25%

50%

75%

100%

0% 25% 50% 75% 100%

Workflow

Life

cycl

e



130

We should note that the more concentrated the points are, the stronger is the relationship, and in a positive relationship slope of the points is upwards moving from left to right direction. Also when the points roughly form a straight line the relationship is linear nature, where they form a curved line it is curvilinear and where there is no pattern there is no relationship.

Therefore we can observe, that there is no regression between workflow and lifecycle, it is just so that people who think workflow is very important, think also that lifecycle is very important and vice versa. The one exception affects to the Pearson's correlation coefficient and we can make an assumption that the workflow was considered more important for some other reason, e.g. because of the order of the questions.

Document management system in general (2.1all)

The generic importance of document management system was not meaningful to compare with the features or requirements, but the answers could be compared by each role. All groups ‐ administrators, users and managers ‐ considered document management systems in general very important (over 75 %).

Even though every group was certain that document management was very important, the managers were the leading group with this approach. On average, managers ranked documents management systems’’ importance to 83 % with 13 % deviation. Systems evaluated gave a biased score for organisational use, and those results will be discussed in the findings.

Security integration (6.2a) – control

Many integrated document management software products leverage the built‐in security features of the operating system, i.e., Microsoft Windows NT (user profiles) or Lotus (access rights, compared here with an OS). PC DOCS' DOCS Open can define security by group membership or restrict access to particular functionality by user session. For example, a user may access a particular set of document on Monday at 3 p.m. to perform certain functions, but not thereafter.

Q6.2a How important it is that document management system can utilize built‐in security features of OS for your process in the future?

In general, integration was very important (77 %) with 22 % deviation. Administrative and learning process considered this very important, and others important. Legal process administrator ranked this the lowest compared with others. Administrative process wanted to comment that ‘there are many separate programs and databases that need own user ids and it will be better if no more log‐ins are required.’ However, the access rights are defined separately, and it is more a question of organisational single sign on and not the system feature.

Lifecycle (7.2all) as a very important feature

Lifecycle was the last very important feature. Activities during the lifecycle occur in most knowledge management processes, so we’ll return to the lifecycle process as described in the following picture:



131

Figure 28. Document lifecycle activities (IEC 82045‐1:2001)

The archiving mentioned in the figure was discussed about in case I.

2.8.3 Exceptions

In this subsection there are explained all exceptions that are over +‐ 25 % distance from the average. This will give us new findings, how the processes differed from each other. All exceptions were analysed, but not reported here.

Development tools (8u) – an invention process and projects exception: customised applications

Increasingly, integrated document management software is really an application services middleware layer that enables customized applications to be built. In most integrated document management offerings, the development environment is graphical instead of script based and provides native support for industry‐standard development tools such as Visual Basic and Visual J++, PowerBuilder, C++, etc. Document management vendors typically use either Sun's Java or Microsoft's ActiveX as tools for the rapid development of reusable software components that can be linked to produce applications quickly. The two products are built on different object models, and neither is subject to an independent standards body. Development tools make it possible to tailor applications to users' needs, although the integrated document management system itself is adjustable depending on the corporate process. For a developer, tools are critical ‐ for the end user they are critical only indirectly. However, more and more end‐users are doing some developing.



132

Q8u How important are the case organisation specific, customised document management applications for your process in the future?

This was a remarkable exception for Invention process, because the total importance of development tools was only 63 % and deviation over 23 %. The difference was therefore over 30%. For projects, this was a big negative exception (‐34 %).

The great importance for Invention process might be true, even though one user said that development tools were very important ‘because outside developers never get it right’.

Systems evaluated tell that the requirement is fulfilled by the products Documentum 6/7 and Domino.doc 5/7. Lotus Domino Document Manager is not open but accessible only via Domino API. However, this depends what kind of development is required.

Types of data (3.1u) – an invention and learning process exception

Current document management software architectures will possibly evolve from a repository designed to handle specific types of information to an integrated repository designed to handle multimedia i.e. all types of unstructured information (text, images, Web pages, video, audio, report data, etc.). These integrated repositories will add functionality required to not only manage the information, but also to understand the information content and structure. Being able to understand content and structure is necessary for interacting with content and component management systems that rely upon the information contained within a document in order to execute an application.

Q3.1u How important it is that document management system handles all types of unstructured information for your process in the future?

In general this was nearly a very important feature (importance 73 %) but deviation was very high (33 %). For Invention process this was a big (+27 %) exception.

An Invention process user commented that invention reports and patent applications were compilations to one document from different systems. According to another Invention process user there were file types as presented in the following list.

Invention document statistics:

60 % Word text 10 % Word Pictures 10 % Power Point Slides 5 % Designer Documents 5 % Scanned Images 5 % Links (markup language documents) 5 % Other

Most of the above files in Invention process were stored in Lotus Notes. One possible need for Learning process was data import from SAP R/3 or Docent. For

Learning process this was a +27 % exception. Both systems evaluated are fulfilling this requirement seven out of seven with no

difference.



133

Check‐out / check‐in (5.1u) – a legal process exception: View checked‐out documents

All document managers have facilities for document check‐out and check‐in. In order to guard against conflicting revisions, checked‐out documents are generally locked so others can not use them. Some products, however, will allow users to view checked‐out documents.

Q5.1u How important is that it is possible to view checked‐out documents for your process in the future?

Invention process user wished that simultaneous read out must be allowed, while for Legal process this was strong negative ‐46 % exception.

This question did not make a distinction between Lotus Domino Document Manager and Documentum. With both Document Management systems the users can see the latest (approved) version since that is made visible in the system. Checked‐out documents are usually stored on the local drive and can not be seen by others (or they must have access to the drive).

Vendors of document management systems (8.2m) – a learning process exception

The integrated document management software market generally consists of two categories of vendors ‐ platform focused and universal. Platform‐focused vendors develop integrated document management software for a specific hardware/software platform or communications environment such as Lotus Notes/Domino or Microsoft Exchange/Outlook. Leading platform‐focused vendors include Lotus Development Corp., Keyfile, and Eastman Software, as well as Novell and Microsoft. Leading universally focused vendors, including Documentum, Hummingbird (ex PC DOCS), FileNET, and Open Text, are repositioning themselves as purveyors of content management and knowledge management solutions for enterprise applications. All of them have released major new versions of their software within the last year to reflect this strategy shift.

Q8.2m How important it is that document management system is 'universal', and not 'platform‐focused' for your process in the future?

For Learning process this was a ‐ 34 % exception. One Legal process manager mentioned that ‘Lotus is not usable, it limits functionality’. Also relational database behind the documentation management and the portfolio functions in front of it were important. Another Legal process manager said that ‘documentation management in general should be universal, but not for contracts necessarily’, and his opinion was that ‘commitment to Lotus is not bad.’

This question made a distinction between Lotus Domino Document Manager and Documentum, because Lotus Domino Document Manager needs Domino and is in that sense platform focused, but we must remember that both platforms integrate.

Tools for filtering (4.2m) – an administrative process exception

Despite improved search technologies, users often are faced with ‘information overload’ as a result of the explosive amount of information being posted on the Internet and corporate intranets. In response to user demand for better tools to assist them in retrieving only the relevant and most up‐to‐date information,



134

vendors have begun adding more intelligence, or filtering capability, to their document management products. Document summarization and content categorization features, along with intelligent agents and Web crawlers, are being added to improve document retrieval from internal repositories as well as from external Web sites.

Q4.2m How important is filtering for your process in the future?

As the Invention process manager mentioned about filtering: ‘If it makes finding relevant information easier, it is definitely important’. Legal process manager said that filtering was used all the time with PC‐Docs. Administrative process manager evaluated this exceptionally low, ‐ 42 % from the average.

System evaluation scored Lotus Domino Document Manager filtering only by one out of seven.

Search from compound documents (4.1a) – an invention process exception

Compound document managers have facilities that allow searching by object type and searching for all documents that contain a specific object. In compound document managers, a document is really a collection of pointers to the locations where component objects are stored. This approach, sometimes called the virtual document, results in a document that configures dynamically only when it is retrieved, and what displays depends on several factors, including the viewing privileges of the requester.

Q4.1a How important is compound documents search ability for your process in the future?

Due to Invention process administrator's zero line, their difference from general average is 63 % and deviation 32 %. The whole Invention process personnel is not supposed to be on the same line, if we compare with the subsection Compound vs. non‐compound document management system (2.3all) ‐ system background.

Maybe compound search ability is definitely not important for Invention process, but e.g. Legal process administrator considered it important for checking of existing contracts and non‐disclosure agreements.

The difference between products is as drastic: Documentum 7 / 7 and Lotus Domino Document Manager 0 / 7. The explanation is simple: Lotus Domino Document Manager did not support compound documents.

The Pearson's correlation coefficient between the questions 2.3all (compound documents assembled from independent elements) and 4.1a (compound search ability) is very high (95 %) among administrators. The small difference appears because administrators thought that it is even more important to search from compound documents (64 %) than to build compound documents (60 %). The administrators deviation is 10 % higher compared with total deviation concerning the compound documents in general.

Cross‐repository searching (4.1m) – legal, learning and administrative process exception

One aspect that separates products targeted to workgroups from those meant for the enterprise is the capability to search across multiple servers, i.e., multiple repositories. Cross‐repository searching has become a core backbone service for enterprise document management software solutions.



135

Q4.1m How important is cross‐repository searching for your process in the future?

The Invention process manager did not consider this very important, because ‘process is mainly interested in our own documents and would like to control who is allowed to search those even within the company.’ This is understandable with patent applications.

Cross‐repository search was an exceptionally low ranking for Administrative process (‐46 % from average). Administrative documents obviously are neither so public. But of course we must remember that cross‐repository searching does not force to make documents public. For Administrative and Legal processes this was just a significant exception (+26 %).

One Legal process manager did not answer, due to corporation contracts were different documents and stored in one place. Another Legal process manager said that ‘there will be retrieval from different sites.’

Document management systems are of course capable in searching from their own repository. For overall searches — like combination of Web, Documentum and Lotus bases — a separate tool is needed. There were various search engines that have been used in corporate wide search. The search possibilities are talked about more in the following case study.

Security Levels (6.1m) – a Legal process exception

Document security may be applied at the network operating system level, the repository level, the document level, or the document element (object) level, depending on the document management product's capabilities.

Q6.1m How important are different security levels of document management for your process in the future?

One Legal process manager emphasised that ‘security must be controlled on Document level.’ Another manager described the value chain from customers to (sub) projects and also towards a laboratory, line organisation group, programs, projects, etc. The different security levels were hard to update in a fast living organisation.

An Administrative process manager said that their system (R/3) included the user profiles and access rights. One Invention process manager, who was answering to the questionnaire, wrote that he did not understand the question. (All others were interviewed.)

The system evaluation gave scores Domino.doc 7 / 7 and Lotus Domino Document Manager 5 / 7 when comparing the security levels.

Standards (8a) – an invention, legal and learning process exception.

Document management software vendors have adopted interface and interoperability standards, including those of the Document Management Alliance (DMA) and the Open Document Management API (ODMA) task force, which are complementary standards sponsored by the Association for Information and Image Management (AIIM). The move towards Web‐centric platforms and technologies has led to the adoption of Internet standards such as HTML, XML, and soon WebDAV.

Q8a How important it is that document management system is following standards for your process in the future?



136

This was a big negative exception (‐33 %) for Invention process as well as Legal process and even bigger positive exception for Learning process (+38 %).

The System evaluation scored how well the requirement is fulfilled with the products as follows: Documentum 7 / 7 and Lotus Domino Document Manager 5 / 7.

Notifications (6.1u) – a learning process exception

Some document management systems will automatically notify document owners that an element in one of their documents has changed and will report on all documents affected by the change. This capability is particularly crucial in enterprise wide repositories containing critical documents.

Q6.1u How important it is that you'll get a notification when your document is modified for your process in the future?

This was an exceptionally important feature for the Learning process (+32 %) and a slight exception for Administrative process.

An Invention process user commented that ‘normally certain documents are strongly protected against any kind of changes...’ However, the protection is different from notifications: it is possible to protect documents cleverly and get just the right notifications ‐ to mobile devices as easily as any clients.

System evaluation scored Documentum six out of seven, while Lotus Domino Document Manager got two points.

Objects as attributes (2.4a) – an Invention, legal and administrative process exception

A key difference in document management software is whether the database contains merely the attributes describing the object or the object itself, stored as a binary large object (BLOb).

Q2.4a How important it is that objects are stored merely as attributes for your process in the future?

Even though Invention process ranked this to zero, there is only a ‐39 % difference to average. There is the same exception also with Legal process, ‐39 %.

Legal process administrator said: ‘important is, that document management is working, the model of implementation is not so important. Security is noteworthy’.

Administrative process representative commented quite similar: ‘Most important is that information (recruiting instructions etc.) can be accessed as easily as possible. It's a service.’ Anyhow, this was an exceptionally important feature, +46 %, for Administrative process.

This feature made only an indirect difference between Documentum and Domino.doc and Lotus Domino Document Manager, because there was no RDBMS with the latter. Documentum was storing only attributes. Domino.doc did not store anything.

System evaluation gave also a result that Documentum was better in storing the meta‐data since the attributes were stored SEPARATE from the documents in a RDBMS.

This was a feature with the lowest importance (39 %) with a great deviation (33 %). This and UNIX clients were the only requirements under 50 % of importance.



137

Search methods (4.1u) – no exception

Being able to quickly find the right document is an essential function of document management. All document management software products provide built‐in search tools (e.g. Verity or Fulcrum search engines) that enable users to search for documents, typically by index terms or by content. Users can search for documents by attributes (or profile) using Boolean (AND, OR, NOT, etc.) operators or by full text using such methods as text strings, proximity searches, and wildcards. Some document management products allow the use of both profile and full‐text search methodologies simultaneously as a way to improve the relevancy of search results.

Q4.1u How important it is that you can search from document attributes (e.g. document author and date created) and contents (e.g. including acronym 'UMTS') simultaneously for your process in the future?

This was a unique feature, because it was not very important, but neither an exception for anyone. We should take a look to it because search methods were different with Lotus Domino Document Manager and Documentum.

One Invention process user stated that they used just standard codes in every document, while another user mentioned that the basic tool in Invention process is search… There was a huge deviation within Projects users as well. Even though the search ability must be important (compare with Phase 1 interviews) one user ranked it null. What are the generic requirements will be discussed in other sections about search methods.

The system evaluation scored search from document attributes as full points to Documentum and to Lotus Document manager three out of seven. Lotus Document management had poor search possibilities although they were a little bit better in the new release. It was not possible for example to run queries over the system that list all documents created by a certain author.

Talking about existing systems, another Projects user stated in general that he is using Notes Reports for internal group reporting. On the other hand Intranet Web pages were created by those tools. However, Notes was not working properly ‐ there were some employees not using it at all. Also bulletin boards were not working due to the same reason, even though they were needed. Web access has made situation better, but there are still problems with access rights. They must be defined separately for each and every document.

All the exceptions were drawn in graphs and presented to the respondents.

2.8.4 Phase 2 summary

In this subsection there are summarised all relevant differences between processes and system evaluation results biased by the needs of each process. However, the system evaluation results were so obviously pro Documentum, that there is no need to compare the products case by case.

Invention process

The following figure shows graphically the suitability of Documentum and Lotus document management (Domino.doc) features to the requirements of Invention process (IPR) in order by the importance.



138

Figure 29. Invention process requirements

The same numbers are in the following table, and also those numbers weighted with each

other, e.g. IPR * Doc = Documentum feature importance for Invention process (IPR), compared with Lotus Document Management (LDM).

Table 14. Invention process requirements

Ref. Importance Requirement IPR Doc. LDM IPR* Doc.

IPR* LDM

4.1u Very important Search from document attributes 93 % 100 % 43 % 93 % 40 %

8u Very important Customized document management applications

93 % 86 % 71 % 80 % 66 %

3.1m Very important Web servers and thin clients 86 % 86 % 100 % 73 % 86 %

7.1all Very important Workflow 86 % 86 % 43 % 73 % 37 %

4.2.4u Very important Representation about document relations

79 % 100 % 0 % 79 % 0 %

7.2all Very important Lifecycle 75 % 100 % 71 % 75 % 54 %6.1u Very important Notifications 71 % 86 % 29 % 61 % 20 %2.1all Important DocM. system in general 64 % 100 % 57 % 64 % 37 %5.1a Important Manual version numbers 57 % 86 % 43 % 49 % 24 %2.3all Slightly important Compound documents 46 % 86 % 0 % 40 % 0 %2.2all Slightly important Integrated DocM. system 43 % 86 % 43 % 37 % 18 %8a Slightly important Standards 29 % 100 % 71 % 29 % 20 %

AVERAGE 58 % 31 %

The table above is telling that according to system evaluation and Invention process analysis, Documentum met the process needs 58:31, i.e. 1.87 times better than Lotus Document Management.

Invention process requirements and system suitability

0%

25%

50%

75%

100%

3.1m

7.1all

2.1all

7.2all 4.1

u6.1

m2.2

all 6.1u

4.1a

4.2.4u 8u 8a

IPR AVGDocumentumLotus DM



139

Some kind of favouritism could be measured from the all over ranking of the importance of the features. The average of all Invention process answers is 69 %: Managers 63 %, administrators 46 % and users 85 %.

Invention process administrator talked about knowledge management as the synonym of content in the documents, and required it to be included in this case study. Content is of course the purpose of the documents and it will be talked about in the next case study, as well as in the results. Invention process user wished for an ‘easy migration of critical process documents as better future document management‐systems’.

Learning process

The following figure shows graphically the suitability of Documentum and Lotus document management features to the requirements of the Learning process in order by the importance.

Figure 30. Learning process requirements

The same numbers are in the following table, and also those numbers weighted with each

other, e.g. LP * Doc = Documentum feature importance for Learning process (LP) compared with Lotus Document Management (LDM).

Learning process requirements and system suitability

0%

25%

50%

75%

100%

3.1m

7.1all

2.1all

7.2all 4.1

u6.1

m2.2

all 6.1u

4.1a4.2

.4u 8u 8a2.3

all5.1

a

Learning Process averageDocumentumLotus DM



140

Table 15. Learning process requirements

Ref. Importance Requirement LP Doc. LDM LP * Doc.

LP * LDM

4.1u Very important Search from document attributes

100 % 100 % 43 % 100 % 43 %

6.1u Very important Notifications 100 % 86 % 29 % 86 % 29 %4.2.4u Very important Representation about

document relations 100 % 100 % 0 % 100 % 0 %

8a Very important Standards 100 % 100 % 71 % 100 % 71 %2.1all Very important DocMan. system in general 90 % 100 % 57 % 90 % 52 %2.2all Very important Integrated DocMan. system 90 % 86 % 43 % 78 % 39 %7.1all Very important Workflow 86 % 86 % 43 % 73 % 37 %8u Very important Customized DocMan.

Applications 86 % 86 % 71 % 73 % 61 %

7.2all Very important Lifecycle 76 % 100 % 71 % 76 % 54 %6.1m Important Different security levels 71 % 100 % 71 % 71 % 51 %3.1m Important Web servers and thin clients 71 % 86 % 100 % 61 % 71 %4.1a Important Compound documents

search ability 71 % 100 % 0 % 71 % 0 %

5.1a Important Manual version numbers 71 % 86 % 43 % 61 % 31 %2.3all Important Compound documents 67 % 86 % 71 % 57 % 0 %

AVERAGE 73 % 36 % The table above is saying that according to System evaluation and Learning process

evaluation, Documentum is meeting the Learning process needs 73:36, i.e. 2.03 times better than Domino.doc.

Some kind of favourableness could be measured from the all over ranking of the importance of the features. The average of all Learning process answers is 80 %: Managers 88 %, administrators 77 % and users 76 %.

Learning process manager required speed, clearness (ease of use), and security, while administrators required ‘the protection of files as well as different safety classes for files and users.’ Learning process user found important ‘defining the rights for groups of documents’.

Projects

The following figure shows graphically the suitability of Documentum and Lotus Document management features to the requirements of Projects in order by the importance.



141

Figure 31. Process requirements

The same numbers are in the following table, and also those numbers weighted with each other, e.g. Pr * Doc = Documentum feature importance for Project (Pr) compared with Lotus Document Management (LDM).

Projects requirements and system suitability

0%

25%

50%

75%

100%

3.1m

7.1all

2.1all

7.2all 4.1

u6.1

m2.2

all 6.1u

4.1a4.2

.4u 8u 8a2.3

all5.1

a

Projects averageDocumentumLotus DM



142

Table 16. Projects requirements

Ref. Importance Requirement Proj. Doc. LDM Pr.* Doc.

Pr.* LDM

3.1m Very important Web servers and thin clients 86 % 86 % 100%

73 % 86 %

4.1a Very important Compound documents search ability

79 % 100 % 0 % 79 % 0 %

2.1all Very important DocMan. system in general 78 % 100 % 57 % 78 % 44 %2.2all Important Integrated DocMan. system 71 % 86 % 43 % 61 % 31 %8a Important Standards 71 % 100 % 71 % 71 % 51 %7.2all Important Lifecycle 65 % 100 % 71 % 65 % 47 %

7.1all Important Workflow 63 % 86 % 43 % 54 % 27 %

4.1u Important Search from document attributes

62 % 100 % 43 % 62 % 27 %

5.1a Important Manual version numbers 57 % 86 % 43 % 49 % 24 %2.3all Important Compound documents 57 % 86 % 0 % 49 % 0 %6.1m Important Different security levels 57 % 100 % 71 % 57 % 41 %6.1u Important Notifications 52 % 86 % 29 % 45 % 15 %4.2.4u Slightly

important Representation about document relations

38 % 100 % 0 % 38 % 0 %

8u Slightly important

Customized DocMan. applications

29 % 86 % 71 % 24 % 20 %

AVERAGE 54 % 28% The table above is saying, that according to System and Projects evaluation, Documentum

met the Projects needs 54:28, i.e. 1.93 times better than Lotus Document Management. If the one and only one feature the case organisation Projects needs have been web servers and thin clients, Lotus Document Management would have been 1.18 times better. However, already the second requirement, compound document search ability, changes the situation 76:43, i.e. 1.77 for Documentum.

Some kind of favourableness could be measured from the all over ranking of the importance of the features. The average of all Projects answers is 68 %: Managers 68 %, administrators 71 % and users 64 %.

Projects administrator said that in an organisation like the focal company, ‘the scalability of the document management system is a must’. Technical information about scalability is available about all systems, but not discussed here. One very often occurred requirement was ‘ease of use’. Another Projects user required that ‘it should be possible to easily share the documents over organisational and geographical borders’. This is a special challenge for an extended organisation.

Administrative process

The following figure shows graphically the suitability of Documentum and Lotus Document Management features to the requirements of Administrative process in order by the importance.



143

Figure 32. Administrative process requirements

The same numbers are in the following table, and also those numbers weighted with each other, e.g. AP * Doc = Documentum feature importance for Administrative Process (AP) compared with Lotus Document Management (LDM).

Table 17. Administrative process requirements

Administrative process requirements and system suitability

0%

25%

50%

75%

100%

3.1m

7.1all

2.1all

7.2all 4.1

u6.1

m2.2

all 6.1u

4.1a4.2

.4u 8u 8a2.3

all5.1

a

Administrative processaverageDocumentum

Lotus DM

Ref. Importance Requirement AP Doc. LDM AP * Doc.

AP * LDM

3.1m Very important Web servers and thin clients 100 % 86 % 100 % 86 % 100 %7.1all Very important Workflow 95 % 86 % 43 % 82 % 41 %7.2all Very important Lifecycle 90 % 100 % 71 % 90 % 65 %2.3all Very important Compound documents 90 % 86 % 0 % 78 % 0 %6.1u Very important Notifications 86 % 86 % 29 % 73 % 24 %4.2.4u Very important Representation about


2.1all Very important Document mgmt system in general

86 % 100 % 57 % 86 % 49 %

2.2all Very important Integrated document mgmt system

86 % 86 % 43 % 73 % 37 %

8u Very important Customised doc. M. applications

86 % 86 % 71 % 73 % 61 %

4.1a Very important Compound docs search ability 86 % 100 % 0 % 86 % 0 %4.1u Important Search from document

attributes 71 % 100 % 43 % 71 % 31 %

8a Important Standards 71 % 100 % 71 % 71 % 51 %6.1m Important Different security levels 57 % 100 % 71 % 57 % 41 %5.1a Slightly

important Manual version numbers 43% 86 % 43 % 37 % 18 %

AVERAGE 70 % 35 %



144

The table above tells that according to System and Administrative process evaluations,

Documentum met the Administrative process needs 70:35, i.e. 2 times better than Domino.doc. If the one and only one feature the case organisation Administrative process needs is

web servers and thin clients, Domino.doc will be 1.18 times better. However, already the second requirement, Workflow, changes the situation 84:71, i.e. 1.18 for Documentum.

Some kind of favourableness could be measured from the all over ranking of the importance of the features. The average of all Administrative process answers is 75 %: Managers 80 %, administrators 65 % and users 89 %.

Administrative process manager required ‘fast access to information. ‐ For the administration it will be good, if the project folders can be used for shared documents.’ The process administrator mentioned the following points: ‐ ‘When e.g. search engine is used in the administrative system, the manager can get a notification about the filling of other forms concerning recruitment ‐ Notification about workflow, responsibilities, work stages in the process to less the amount of work: ‐ all possible information will flow from one filled document to other documents, e.g. salary to payroll information’ – ‘The control of administration situation, e.g. queuing times: how overloaded administration is in the time present: e.g. when a form is filled, there will be a notification, where it says, how many other similar documents are waiting, and possible waiting time estimation’. Administrative process user was sure about real utility of the system: ‘Even if some of the features could be used today, it would make certain processes faster and lessen futile paperwork.’

Legal process

The following figure shows graphically the suitability of Documentum and Lotus Document management features to the requirements of Legal process order by the importance.

Figure 33. Legal process requirements

The same numbers are in the following table, and also those numbers weighted with each other, e.g. LgP * Doc = Documentum feature importance for Legal Process (LgP) compared with Lotus Document Management (LDM).

Legal process requirements and system suitability

0%

25%

50%

75%

100%

3.1m

7.1a

ll

2.1a

ll

7.2a

ll

4.1u

6.1m

2.2a

ll

6.1u

4.1a

4.2.

4u 8u 8a

2.3a

ll

5.1a

Legal process averageDocumentumLotus DM



145

Table 18. Legal process requirements

Ref. Importance Requirement LgP Doc. LDM LgP * Doc.

LgP* LDM

6.1m Very important Different security levels 100 % 100 % 71 % 100 % 71 %3.1m Very important Web servers and thin

clients 93 % 86 % 100 % 80 % 93 %

7.2all Very important Lifecycle 89 % 100 % 71 % 89 % 64 %7.1all Very important Workflow 86 % 86 % 43 % 73 % 37 %2.1all Very important DocMan. system in

general 82 % 100 % 57 % 82 % 47 %

2.2all Very important Integrated DocMan.System

75 % 86 % 43 % 64 % 32 %

4.1a Important Compound documents search ability

71 % 100 % 0 % 71 % 0 %

5.1a Important Manual version numbers 71 % 86 % 43 % 61 % 31 %4.1u Important Search from document

attributes 57 % 100 % 43 % 57 % 24 %

6.1u Important Notifications 57 % 86 % 29 % 49 % 16 %8u Important Customized DocMan.

applications 57 % 86 % 71 % 49 % 41 %

2.3all Important Compound documents 54 % 86 % 0 % 46 % 0 %4.2.4u Slightly important Representation about


8a Slightly important Standards 29 % 86 % 71 % 29 % 20 % AVERAGE 60 % 32 %

The table above tells that according to System and Legal process evaluation,

Documentum met the Legal process needs 60:32, i.e. 1.88 times better than Lotus Document management. The Legal process user however was the only defensive link, and her answers affects to the figures quite much. The answers of the experienced managers are not weighted.

Some kind of favourableness could be measured from the all over ranking of the importance of the features. The average of all Legal process answers is 65 %: Managers 86 %, administrators 68 % and users 52 %.

A Legal process manager repeated ‘easiness of use as well as easiness to build actual applications’ as most important requirements. Reliability was also important. Legal process administrator emphasised archiving: undersigned contracts are scanned to the database ‐ search of from scanned documents is needed. The input of scanned documents; and electronic signature was mentioned by another Legal process manager. Also corporation internal contracts are in portfolio, so also contract meta‐data should be in portfolio. Now the archive is in another system. Concerning extended organisation, the project contracts with associates are stored separately for law specialists.

Case II, phase 2 exceptions

Finally, all exceptions that were not listed as very important features were analysed (not listed here). Percentage (e.g. ‐26 %) told how much those process answers differentiate from average, and all significant (>+‐25) exceptions were marked from every process. Even though this was not a statistical analysis, those exceptions showed the differences pending from the process level of formality. In the results part 3 we’ll return to the case organisation on average level.



146

The case organisation on average

The following figure shows graphically the suitability of Documentum and Lotus document management features to the requirements of the case organisation in average order by the importance.

Figure 34. Case organisation requirements

The same numbers are in the following table, and also weighted with each other, e.g. CO

* Doc = Documentum feature importance for Case Organisation (CO) compared with Lotus Document Management (LDM).

Case organisation average requirements and system suitability

0%

25%

50%

75%

100%

3.1m

7.1all

2.1all

7.2all

4.1u

6.1m

2.2all

6.1u

4.1a4.2

.4u 8u 8a2.3

all5.1

a

Case averageDocumentumLotus DM



147

Table 19. Case organization (CO) requirements

Ref. Importance Requirement CO Doc. LDM CO * Doc.

CO * LDM

3.1m Very important Web servers and thin clients

87 % 86 % 100 % 75 % 88 %

7.1all Very important Workflow 80 % 86 % 43 % 68 % 34 %2.1all Very important DocMan. system in

general 79 % 100 % 57 % 79 % 45 %

7.2all Very important Lifecycle 77 % 100 % 71 % 77 % 55 %4.1u Very important Search from document

attributes 75 % 100 % 43 % 75 % 32 %

6.1m Important Different security levels 74 % 100 % 71 % 74 % 53 %2.2all Important Integrated DocMan.

system 71 % 86 % 43 % 61 % 31 %

6.1u Important Notifications 68 % 86 % 29 % 58 % 19 %4.1a Important Compound documents

search ability 64 % 100 % 0 % 64 % 0 %

4.2.4u Important Representation about document relations

63 % 100 % 0 % 63 % 0 %

8u Important Customized DocMan.applications

63 % 86 % 71 % 54 % 45 %

8a Important Standards 62 % 100 % 71 % 62 % 44 %2.3all Important Compound documents 61 % 86 % 0 % 52 % 0 % 5.1a Important Manual version numbers 60 % 86 % 43 % 51 % 26 %

AVERAGE 61 % 31 % The above table says that according to System evaluation and the case organisation

average evaluation, Documentum is meeting the case organisation needs 61:31, i.e. 1.97 times better than Domino.doc.

Some kind of favourableness could be measured from the all over ranking of the importance of the features. The average of all the case organisation average answers is 71 %: Managers 73 %, administrators 70 % and users 73 %

The final results about Documentum and Lotus document management based on the data available here in the Phase 2 there is no reason whatsoever to select Lotus document management instead of Documentum to fulfil the need for document management system detected in the Phase 1 of Case II studies. But the question is, whether document management is enough for knowledge management.

Moving from case II to III we can tell that applications are not enough. A group of engineers tried to leverage their knowledge by storing their documentation in a common document management system. They needed to understand the logic that other engineers used. The documents themselves were neither enough. The constructing of the knowledge within its context is required.

Systems

3 CAS

services managemof the ovmethodsneeds ofinformatengineerWISE sysobjective

O3.1

complexsafety anproject taccess ewith an e

during eknowledwas less approac The case 1. To p

organisapracticalprocessescenariodevelopmdecisioninterfaceand by wmanaginmanagemtaken fo

2. To dev

engineerof‐the‐aand recoprocesseannotati

between inf

SE III ‐ EXTEN

This cfor engineerment in partverall pictures but rather tf industrial ution sourcesring knowledstem, but gees were desc

bjectives

Enginex and safety‐nd quality. Htherefore dessential knowengineering

Case Iach stage ofdge managemon developihes oriented

e III objective

propose work

Basedations and thl and effectives. This incluos of how thement; (2) kns and rationaes to existingwhom will knng and stimument or teamr each detec

velop a proto

Systemring processert KM instruord mission‐ces. The featuing and trans

formation an

NDED ORGAN

hapter presering (WISE). Tticipatory dee. The main oto integrate sers. This re. In this contdge. The collneric requirecribed also o

eers need a critical prodHowever, accsigned practwledge and knowledge pII developmf the project.ment methoding new specd towards th

es were as fo

kable knowle

d on an analyhe world‐widve KM approded approace existing knoowledge acqale be captug systems arenowledge in tlating knowlms detect thcted KM nee

otype softwa

m architectues. This webments into tcritical knowures of this psforming kno

nd knowledg

NISATIONAL

ents the resuThis Europeasign processobjective of and exploit search projetext, this caseaboration poements for ton project fac

large amounucts. Reusingcessing this ktical methodinformation portal solutioents followe. Supporting ds and ways cific KM toole needs of in

ollows:

edge manage

ysis of KM nee state‐of‐thoaches that cches for (1) eowledge canquisition, i.e.red during ae required; (the knowledledge exchane needs for d, what are r

are to suppo

ure was deve‐based archihe workspac

wledge in a uslatform incluowledge obje

ge

148

L KNOWLEDG

ults from thean Union resses of complethe project wexisting statect was protoe III is an exaossibilities behem were octs sheet (W

nt of reliable g existing knknowledge iss for engineefor their tason. ed a human‐cstudies addrfor managins and methondustrial use

ement appro

eeds and worhe‐art and stcan be seamlefficient usagn be provided, how, when typical prod3) maintenadge base be vnge and knowknowledge mrelated succe

ort engineeri

loped to supitecture enabce of engineeser‐friendly wuded: a persoects needed

Part II – In

GE AND A KM

e third case, nsearch projecex products, was not on de‐of‐the art otyping a meample to shoetween orgabserved fromISE 2003).

knowledge owledge sav often difficuers and theirks. Case III in

centred pararessed the acng their indusods, but ratheers.

oaches for en

rk processesate‐of‐practlessly integrage of existingd to and utilin and by whoduct developnce of corpoverified, validwledge devemanagementess factors, e

ng knowledg

pport the probled the seaers. This enaway – integronalised knoin specific ta

itial case stu

M SYSTEM

named Web‐ct was conceputting the

developing napproaches eta‐system foow what findanisations wem a theoretic

and informaves resourcesult and time‐r managers tntended to s

adigm: it wascceptance ofstrial introduer to integra

ngineering p

s in the particice in the fieated with exg knowledgeised by enginom will evolvpment cycle, orate knowledated, updatelopment, i.et, what are tetc.

ge managem

oposed KM amless integrbled engineeated in theirowledge worasks; CAD ac

udies

‐enabled infoerned with knengineer in tew specific toriented towor different kings are obsere not inhercal framewo

tion when ds and improv‐consuming. to easily sharupport these

s involving stf the proposuction. Case ate and explo

rocesses

cipating engield, case III suisting enginee resources, ineers in prodving knowledwhat tools aedge, i.e., hoted and delee., how can he approach

ment

approaches iation of exisers to accessr everyday wrkspace for ccess, e.g., fo

ormation nowledge the centre tools and wards the kinds of erved from rently in a rk. The

esigning ves product Case III re and e methods

takeholders ed III focus oit existing

ineering uggested eering i.e., provide duct dge, and w, when ted; (4)

hes to be

n sting state‐s, re‐use work ollecting, or searching



149

existing parts and assemblies; access to knowledge communication services such as finding and contacting reference persons, FAQ’s, etc.

3. To take into account human factors issues in engineering knowledge management

There were user needs and requirement investigation to qualify the difficulties engineers encounter with knowledge management in industrial organisations. The design of the KM architecture was user‐centred. Several issues were taken into account, including the following:

1. Work practise and behaviour. KM solutions must fit into typical engineering

work practise. Otherwise, a realistic scenario must be proposed to adapt the work practise to new ways of working.

2. User‐friendliness of developed processes and tools. The solutions should be straightforward and reliable to use and facilitate the reduction of workload.

3. Usefulness. The solution must have a clear added value. 4. Professionalism. Engineers must be able to develop their products faster, with

higher quality and increased safety. 5. Socio‐cultural obstacles. Knowledge management initiatives deeply impact

existing corporate structures, processes, and balance of power. Therefore they often meet resistance from anyone who might be affected by these changes.

6. Responsibility. In the development of safety‐critical systems, reliability and ability to trace provided knowledge are key issues. Therefore clear responsibilities and rules must be defined for acquisition, maintenance and usage of knowledge items.

7. Motivation. An essential success factor for convincing engineers in working with KM is to make their work more interesting and more rewarding. Routine work should be reduced and empowerment supplemented by induced learning processes.

8. Communication. As product development processes become more complex and involve more players, communication is becoming a bottleneck. KM solutions should provide means to supplement and improve existing communication structures.

All in all case III was a research project that was putting a high premium on human co‐

operation. Therefore the main concern of this chapter is all about extended organisational knowledge. The work packages of the project are shown in the following picture.

Systems

Before t

informatpractical

Ex3.2

project, report. Tmain res

stuck on23.9.199has just

in techn

between inf

he project, t

StateEvalu

deKnowAppro

deKnowKnowKnowProto Case I

tion sourcesl knowledge

xtended orga

Some but the reseThe division osearch done

Whenn the assump98). The situareplaced the

Extenical context)

formation an

he expected

‐of‐the‐art uation methevelopment wledge proceopriate metesign wledge systewledge‐basewledge appleotypes to illu

II research p. In In this comanagemen

anisation

participantsearcher condof case III wafor the focaln starting thisption that moation remaine word ‘teamded organisa. Organisatio

nd knowledg

Figure 35

d results wer

report od for poteand designess guidelinthods for kn

em architected organisatets class librustrate the

project madeontext, this snt system.

s from the buucted from tas not equall company ws research, inost knowledgns the same, m’ with otherational knowons are nowa

ge

150

. Work packa

e like:

ential knowl nes nowledge ac

ture and inftion guidelinrary results

e a prototypeystem is an e

usiness unitsthe focal com– the protot

were about thnteractive spge is generatdespite the r sales concewledge is socadays turnin

Part II – In

ages of case II

ledge‐driven

cquisition in

frastructurenes

e of a meta‐sexample to s

s of the focalmpany all thetype system he human knpecialist areated in teams marketing jaepts. iotechnical (g into netwo

itial case stu

I

n processes

n product de

e for effectiv

system for dshow what w

company we analysis wowas just thenowledge issas like discus(see e.g. Kitargon of man

in generic morks, strategi

udies

s in product

evelopment

ve exchange

ifferent kindwas observed

were involvedork included tip of the icues. sion databasson & Baltupnagement co

meaning, sociic alliances, a

t and

e

s of d from

d in this in this eberg. The

ses were p onsultants

al networks and virtual



151

organisations. In this case III, the product creation and engineering work is done in global, distributed organisations. They are actors in the social networks, while highly technological in nature.

This third case ‘Extended organisational knowledge and a KM system’ is taking a look into co‐operation between companies, which has increased significantly, especially in the research and development of high technology. The ability to produce innovations efficiently in the rapidly changing competitive environment has become dependent on the inter‐organisational networks. There are various reasons for connections between organisations, such as striving for increased revenue or cost reduction, and accessing the complementary resource or crucial competences of the partner company (Ebers 1997). The first reason is straight business‐beneficial, while the second could have other consequences as well.

The new organisational forms have been developing in an environment that is characterised by global competition and globalisation, fragmentation of the markets, international decentralisation of production, and the demand of customised products. The organisations have responded to these challenges by outsourcing, flexible manufacturing, and innovating fast. Work has become complex, and it requires cross‐functional teamwork, continuous learning and knowledge‐intensive skills, abilities to collaborate with expert colleagues, and capability to adapt to constant change (Hatch 1997).

Extended organisation has got its own requirements. The emergence of the virtual work demands improved KM activities, while at the same time the virtual way of working makes KM more complex than before. The internet‐based technologies may offer one solution supporting the communication of knowledge between co‐operating organisations or colleagues (Schwartz et al. 2000).

Knowledge management began to emerge strongly on the systems after the proliferation of the Web‐browser (Maurer 1998). The browser is used as a common interface, even though the standards are divergent. Case III KM system is based purely on web technology.

3.2.1 Social constructivism and learning

These subsections are mainly based on a literature survey, but observations from case studies made for the WISE project are used as references. The method is overall qualitative. The aim of this subsection is to verify how clearly the theories from sociological context suit to information services and ‐ in this case – engineering in an extended organisation. The traditional interpretations of organisational knowledge are compared with modern requirements for collaboration between organisations.

As described in the Part I, Chapter 1.6, Sociotechnical view and the actor‐network theory, information system as a case can be observed through an actor‐network theory. Sociological context presented here is based on constructivist theory. It is a framework for interpretation of knowledge. In that theory actors are considered to enact and construct realities based on their mental models, which are shaped through interpretation and discourse between different members. The socially‐oriented constructivist theory (compared with cognitive) is more appropriate when talking about organisational knowledge.

Lev Vygotsky, a Russian psychologist and philosopher in the 1930s, is often associated with the social constructivist theory. He emphasises the influences of cultural and social contexts in learning and supports a discovery model of learning. This type of model places the teacher in an active role while the students' mental abilities develop naturally through various paths of discovery. Here are his assumptions adapted to engineering as interpreted by 21st century schools.

Vygotsky’s Principal Assumptions

1. Making Meaning The community places a central role. The people around the person (student, engineer) greatly affect the way he or she sees

the world. 2. Tools for Cognitive Development



152

The type and quality of these tools determine the pattern and rate of development. The tools may include: important experts to the engineer, culture, language. 3. The Zone of Proximal Development According to Vygotsky's theory, problem‐solving skills of tasks can be placed into three

categories. These are as follows: a) those performed independently by the person; b) those that can not be performed even with help; and c) those that fall between the two extremes, the tasks that can be performed with help from others.

Vygotskian Principles – from classroom to engineering

• Learning and development is a social and collaborative activity that can not be ‘taught’ to anyone. It is up to the learner to construct his or her own understanding in his or her own mind. It is during this process that the expert acts as a facilitator.

• The zone of proximal development can be used to design appropriate situations during which the engineer can be provided the appropriate support for optimal learning.

• When providing appropriate situations, one must take into consideration that learning should take place in meaningful contexts, preferably the context in which the knowledge is to be applied.

• Out of working (learning) environment experiences should be related to working environment experiences. Pictures, news clips, and personal stories incorporated into working environment activities provide the workers with a sense of oneness between their community and learning.

Those principles should be considered in an enterprise as well as in a classroom. There are still more detailed branches from socially‐oriented constructivist theory. When exploring the world including information systems, the actor‐network theory could be used, as information systems are often a kind of meta‐model of the world. It studies stabilisation processes of technical and scientific objects as they result from the building of human actors and natural and technical phenomena. Strong social constructivism is criticised for giving special preference to social elements, whereas natural or technical elements are allowed to be actants or actors in networks through which technical or scientific objects are stabilised in actor‐network theory.

Still more precisely for this essay current interpretations from the social locus of technological practice by Edward W. Constant II are used as the basic categorisation model. Other literature findings are adapted to these categories. Theories used in this work are mainly based on social sciences and philosophy. However, the long roots are leading us also to current knowledge management literature (see e.g. Barnes 2002). WISE‐specific observations are mostly about engineering, but also from international co‐operation with different organisations from different countries.

3.2.2 Culture and cooperation

Culture is the unseen hand of all functions in an organisation. For knowledge management, the infrastructure, technology and measurements are all necessary but they will not work without the right kind of organisational culture for managing corporate knowledge. Culture is the combination of shared history, expectations, unwritten rules, and social mores that affects the behaviour of organisation and its members. Important aim for KM is to enhance knowledge sharing and collaboration, but nothing will happen if company’s nature is to hoard knowledge with structural barriers (see e.g. O’Dell & Grayson 1998, p. 71). The culture influences systems, communities and organisations, but also on a micro level: the strongest barrier to transfer e.g. best practices or lessons learnt is the lack of credibility between source and recipient (see e.g. Szulanski 1998). The value and relevancy of knowledge always depends on the recipient. KM should be part of the corporation culture



153

instead of part of HR department, not to mention a single Chief Knowledge Officer in a position to understand hundreds of different specific contexts. The technological practice in corporations is the key issue in most engineering, but it easily becomes an aim itself also for cross‐organisational culture.

Culture is very likely a permanent feature of the evolution of all human life. Yet there is equally likely no one universal human culture, characterising all human beings in the same generic way at all times and places. The concept of culture is mainly used to refer to specific historical forms, such as the state (national culture), the economy (market culture) of the corporation (organisational culture). Culture and characteristics of people are different depending on the nationalities. Experts in the USA want their knowledge to be used and acknowledged differently compared to their European colleagues. The epistemological cultures of knowledge creating companies are also different. Knowledge creation varies depending on the field of research, development, or engineering (see e.g. Knorr 1999; Nonaka & Takeuchi 1995). In any case, knowledge sharing must be encouraged, and positive rewards are giving positive result on individual recognition, as well as corporate level in any culture. Because of this similarity, and the specific scope of this research, this section is concentrating only on organisational knowledge context.

In the corporation culture, knowledge shall not be seen as a source of power. Neither the executives nor the experts shall create ‘knowledge fiefdoms’. Knowledge engineers should be naturally motivated to search and share knowledge. The ‘pull’ philosophy put the responsibility to knowledge workers to seek out knowledge to improve the performance and processes. When the efficiency and effectiveness of the corporation grows, there are not displacements but retraining. Technology must support knowledge creation, sharing and searching (O’Dell & Grayson 1998, pp. 76, 80, 82). However, this depends from the type of knowledge. Technology is easy to share, but we should still remember that pending from the philosophical paradigm, or information species presented in the Part I Chapter 2, Information and knowledge species, knowledge can not be ‘shared’ – it is always reconstructed both on individual and organisational level, and especially on cross‐organisational collaboration. Storing (or presenting) and communicating pragmatic information are different qualitative species.

3.2.3 Findings within the social loci of technological practice

Technology in an organisation has got its own cultural context. Originally there were three different social loci for technological practice; 1) technology as knowledge (community), 2) technology as system (not IS), and 3) technology as function (complex organisation – in this case III, corporate). Conceptions of community, system and organisation in technological practice are not contradictory. The culture is including different communities of practitioners, and the organisation and the wider system are separate unions within and around them (Constant 1999, pp. 223, 238). Culture in all of these groups is the key to knowledge transfer. Learning and sharing knowledge are social activities of people belonging to these three categories (O’Dell & Grayson 1998, p. 72). To these categories can be added a 4) meta‐level for cross‐organisational technology (extended enterprise infrastructure).

Across all cultures, mutual obligation is one of the most powerful social forces there is. It is an advantage that individuals engage with organisations for deeper reasons besides solely financial benefits. But inside any organisation, official project teams as well as communities of practice are the vehicles by which the rich and tacit knowledge gets shared among people who feel an obligation to help and support each other (O’Dell & Grayson 1998, pp. 73). There is e.g. a continuum of the most effective approaches for encouraging a best‐practice or lessons learnt transfer. Transfers that are entirely spontaneous encounter the least difficulty in gaining acceptance by a department. Because spontaneous transfers are rare, organisations could consider the second‐best approach: ‘strongly suggesting’ the adoption of the new best practice. Less effective were transfers that were ‘mandated’ – implying that employees are more likely to accept a new best practice if they have some choice in the matter (and suggesting that a ‘mandate’ is not always enforceable). ‘Favoured’ transfers were even less effective, and least likely of all to be adopted are ‘optional’ best practices (Szulanski 1997). It has been argued, whether best practices or lessons learnt are the right definitions, and are they working in computerised systems or not. However, from the example of best practice it can be found that community of practitioners is the first channel for knowledge transfer.



154

Collaborative relationships enable tacit knowledge and high‐value practices to transfer. They can not be put on the information systems. Communities of practice, the basic units of innovation and collaboration, bring together like‐minded people or people with a shared goal (O’Dell & Grayson 1998, p. 79). The core purpose of community of practitioners model is to explain technological development and revolution, conceptual or cognitive novelty, invention and innovation. A typical model for community practice of technology, knowledge is expressed in well‐winnowed traditions that are the possession of well‐defined communities of technological practitioners. Individual membership in a community is different from disciplinary commitment – what distinguishes a practitioner is adherence to the tradition, not disciplinary training. Communities may be composed of either individual adherents to the tradition or, equally, valid organisations. The normal practice of such communities, however they are defined, is the extension and articulation, or incremental development, of the received tradition. In case III, the KM system was aware of developmental problems for a tradition of technological practice: the incapacity to function under new or more stringent conditions (Constant 1999, pp. 224 – 225).

Corporations involved in case III are making technological products that require exceptionally high quality. Traditions of technological practice and their relevant communities usually embody higher‐level traditions of technological testability, which in turn is composed of both some set of testing technologies and techniques and some set of normative values. The technological systems, however, are intrinsically hierarchical, and decomposable. This hierarchical decomposability suggests the absolute relativity of all change and implies multiple traditions of practice of multiple communities of practitioners (Constant 1999, pp. 226 – 227).

Communities of practice are good arenas for communicative collaboration, close to ‘knowledge sharing’. But talking about problem solving, the situation is different. Why an expert, who has solved a problem, will participate to a community without any other reason beside the problem? Their solutions should be captured otherwise. Sometimes people do not know or understand or can not write down how they become successful in a particular area of expertise. That kind of knowledge is hard to transform into a KM system, as it is between information and knowledge, mostly tacit in depth of memories. All meta‐level descriptions of knowledge are thin and pale, like a map. Practices embedded in people, culture and context are complex and rich, like a journey itself. Dialogue and demonstration are needed to enrich the knowledge transfer.

To transfer deep, rich and tacit knowledge and practices effectively, connecting people is needed. Tacit knowledge, as well as explicit, has got some subcategories. Social information (along with physical and other information) can be split to individual and organisational (group) information. Individual tacit knowledge consists of language, identity, membership practice and work practice. Tacit organisational or group knowledge have got three practices: work, identity and membership (Linde 2001, pp. 160 – 162). Communities of practice share on a micro basis to help meet common needs and goals. Corporations must be inspired by a common goal as well – that is a sense of social and organisational purpose that defines the raison d'être of the entity. Knowledge and practices are meaningless if not put in context of purpose. That is the thing even in wider systems. The system including context is sociotechnical (O’Dell & Grayson 1998, pp. 73, 81, 82).

The systems perspective seeks to explain the emergence and development of large‐scale sociotechnical systems – especially invention and entrepreneurship. The hierarchical complex technologies of case III corporations constitute complex systems. At some lower levels, systems can be considered ‘hardware’, for example, an aeroplane, or a GSM base station. Larger systems, however, are invariably sociotechnical and organisational – for example, an air transportation system, or a UMTS network. In case III we were aware also of the sociotechnical aspect of that engineering knowledge. A systems portrayal of technology is hierarchically higher and more general than the community model. It may therefore be thought to include the previously introduced community model as well as the organisational memory section (Cf. Constant 1999, p. 228).

Case III covers an information service system for engineering. That should cater for wide sociotechnical system. Case III as a research project was not too strictly focused on information, but also focused on people. A manager’s choice of actions and investments decisions will be very different depending on her/his viewpoint, but focus on information, as well as focus on people, are both included



155

in knowledge management. Maybe competitiveness in a long term rests in the sociotechnical system of people, not in the explicit information in the computer system (Based on Sveiby 1.9.2000).

The problem solving, knowledge sharing and transfer can be rewarded. But if the process of problem solving, knowledge sharing and transform is not inherently rewarding by the system or even by the culture, then artificial rewards will not have much effect. Different experts may even become cynical (O’Dell & Grayson 1998, p. 82).

The knowledge workers are different in the sociotechnical system perspective. In this system viewpoint there must be separated at least three different kinds of problem solvers: inventor, manager and engineer entrepreneurs. Also the different developmental phases produce a specific culture of technology, composed of distinctive values, ideas, and institutions. Some values are quite general, e.g. technical efficiency, and are shared with much of engineering. For each phase such general values have system‐specific referents. Cultures of technology also comprise specific organised knowledge, where technology itself, its systematised knowledge, and its culture are embodied in a variety of economic organisations and social institutions. Technology is not autonomous; it requires integration with community and organisation (Constant 1999, pp. 229 – 230).

Organisational portrayals seek to explain technological change, innovation, entrepreneurship, and economic and organisational growth. Technological knowledge and systems are not the products of organisations. The products are artefacts (a GSM base station) or output of complex systems (not only an airbus, but flights to the passengers), both of which embody and require an immense amount and variety of technological knowledge. Technological knowledge is never pure: this circumstance implies the co‐operation of adherents of multiple communities of practitioners in the creation of any complex system. Integration of this multiple expertise in turn implies complex organisation. Case III was providing interaction and processes ready for intra‐organisational interfaces among functionally differentiated departments and between hierarchically higher and lower levels, each of which have slightly different goals. Technology as function is mediated by organisation (Constant 1999, p. 231). The organisational success is a function of the people that make it work – no matter if it is intellectual capital, business intelligence, organisational know‐how or structural knowledge.

The organisational divisions or departments, as well as unofficial communities have different missions and visions. So they produce different cultures that inhibit the transfer of knowledge and lessons learnt. On a micro‐level there are personal relationship either pro or ante the knowledge sharing culture (O’Dell & Grayson 1998, p. 72). The structure and functioning of business enterprise have been well described already in Alfred Chandler’s Visible Hand (1977). Large business organisations rise as the result of a process of successive inclusion, not as the result of a process of internal differentiation and specialisation. There are often functional modules of an enterprise, e.g. R & D, sales, or engineering. For case III, it was noted that each module of organisation, e.g. an engineering, assembly or design community is a locus for potential change. Knowledge can simply change autonomously. People learn by doing. However, the computer systems seldom follow the changes.

The modules of an organisation represent also an environmental interface through which the organisation both acts and receives information, most often with specialised communities of practitioners to which organisation members belong (Constant 1999, pp. 233, 236). Even though the content was information for engineering, case III took a notice of established networks of people extended outside the corporations. Also within the corporation, knowledge transfer has got many formats: serial, near, far, strategic and expert transfer (Dixon 2000). The extended organisational knowledge sharing is very crucial to understand by present day management.

Systems

corporatorganisaknowledorganisabroader much of informatinformat

Ca3.3

ethnomeintention

group sptalking aMany qusubsecti

3.3.1 W

what’. Evart in thecurrent K

between inf

Case Itions. Case IIational culturdge. Seeing cations as the dynamic andf the history tion infrastrution technolo

ase III, initia

Thereethodologicanal time brea

Whenpecified in deabout totally ualitative surons.

Who knows

As perven though te focal compKnowledge M

formation an

Figure

II was a reseI should notre must provommunitiessocial locus d holistic strof technologucture with eogy will be b

l survey – sta

were duringal research inak between n this case III eep theoreticdifferent thrveys were d

what – preli

rceived also this is not a ppany. The obManagement

nd knowledg

e 36. Extended

earch project be looked thvide purpose of practitionof technologuctural contgy, business aespecially thbased on hist

ate of the ar

g the years mn the workplparticipationbegan, the scal study of kings with theone during t

minary inter

in the later iplain questiojective of tht activities g

ge

156

d Organisatio

t for the futuhrough a nare and motivatners as the sgical functionext of both mand econome human asptory of peopl

rt

many differenace. To reacn to work ansituation of tknowledge me same namethis project.

rviews

interviews, aon to ask, thee first brief soing on in th

Part II – In

on include com

ure of Europerrow view oftion, rewardocial locus on and sociotemay offer a wmics. Case III spect that hasle.

nt interview ch the objectd to write ththe focal commanagemente, as we can The results a

a good startine case III begsurvey was tohe focal com

itial case stu

mmunity mem

ean Union anf KM tools fod and incentivof technologiechnical macway of reconsucceeded as got long roo

sessions, antive distance he results. mpany was twt. However, tsee from theare combine

ng point is togan by investo assess exispany. A set o

udies

mbers groups

nd the involvor storing knove structure cal knowledgcrosystems anciling the inss a state‐of‐tots. Also futu

nd continuouthere was a

wofold: therthe businesse following rd in the follo

o find out ‘wtigating the ssting materiaof questions

ved owledge: for sharing ge, as the sights of the‐art ure

us n

re was a s units were results. owing

ho knows state of the al on the was sent to



157

KM experts throughout the company per e‐mail. The intention was to create an overview of what kind of formal and informal support the organisation provides for KM in terms of KM tools, processes, and strategies.

For the focal company, as well as for the collaborators, it is crucial to know the responsible persons for every task. Inside the focal company the personal relationships played an important role (that was also found out in the later interviews, e.g. # 13), and with the collaborators some formal lists were required. A simple meta‐information on ‘who knows what’ (see the Figure 37. Topics related to reference persons from case III application tested) is either too vague or quite fraught to get on the appropriate level. The relation between topic and a person could be more than a list.

Figure 37. Topics related to reference persons

When the collaboration begins, the collaborators’ level of knowledge is assumed. The

expertise on information system usage is also an open issue. It is necessary to find out what the user knows and what she or he does not know. Despite the state of the art of the focal company, this preliminary interview also told the level of research collaborators.

For collaborators, it is especially important to define in the planning phase the right contact persons within a program to help to find persons by their roles. When the person changes, it should be a traced, so that the role can be fulfilled and the right people will be found continuously. There can be a kind of catalogue about contact information with job titles at least. This could be available for collaborators (as revealed also in the phase 2 interview # 9),

The first set of questions was sent by e‐mail to 27 respondents asking for a response within 9 days. Additional four people received the e‐mail two days later, after they had announced their interest in participating. It was a purpose that the respondents will themselves rank the confidentiality of their answers, and therefore an external e‐mail address was used as a main channel, together with the focal company address for confidential answers.



158

The list of experts, to whom the questions were sent, was created by assumptions based KM related group on intranet. The researcher sent an introductory e‐mail to potentially interested KM experts using focal company’s internal KM‐forums as the background information. If interested in participating to the survey, the KM experts replied to the researcher, who then forwarded the addresses to a researcher at a collaborating organisation. The collaborating organisation researcher then sent the actual questionnaire to the respondents. The intention was that the respondents would note (from the explanation text, but also from the sender’s address) that the e‐mail was sent from outside the focal company, in order to consider the confidentiality of the material requested. Also, it was an intention to test the ‘extended organisation’ recursively from the beginning.

The correspondence with interviewees was stored, as well as other details on their expertise area. Concerning extended enterprise, it is worthwhile to store meta‐information about contact persons. This can be stored with a common document base, where there is more generic information available (as mentioned also in the interview # 3). This should strictly follow laws of data privacy.

One big issue with this ‘who knows what’ concept is that it ignores the division of knowledge and the position of a person. We should think separately what an employee knows, and what her or his duty is to know. In a fast changing organisation the updating of the ‘who knows what’ system falls possibly behind. Employees will get questions based on their former office. In general, it is possibly better to let the employee publish his or her formal education and work related courses than let the colleagues (i.e. competitors) to evaluate their knowledge.

About the answers to the questionnaire, the respondents were asked to mark the material public, project confidential, or focal company confidential, since the conclusions are available to externals. So we ensured that the researcher would not write confidential information in the report. This preliminary survey was intended to be fairly ‘light‐weight’, since it requested only already existing material like documents or previous KM reports.

As presumed, it is easier to get the right information within the focal company than from the collaborators. This was proofed also in the later interviews with comments, such as: ‘It is more difficult to get the real information … e.g. because of the personal reasons — no will to tell….’ and ‘if you are not present in discussions, you can not be sure about the validity of information’ (interviews # 3 and # 9).

3.3.2 Conclusions from the state of the art of the focal company

This initial survey gave first conclusions that a formal policy with managerial commitment was needed to achieve a KM strategy. On an enterprise level there was no strictly defined KM material – neither public, nor specific for the focal company. Even though the respondent’s didn’t have much time to answer the survey, the formally defined KM policies would have been mentioned, as focal company’s objectives and other well‐known subjects. Concerning the KM strategy, the responses varied from ‘yes’ to ‘no’ and in between: ‘We generally do not refer to this as Knowledge Management partly because the term has only been known widely in recent years’. Based on divergent stati, we can observe that organisation (not to mention extended organisation) is a diffuse concept – the level of organisation could be group, business unit, and more interestingly unofficial community of practice.

Since the respondents focused on different aspects of KM and used different terms, the answers were also different. Some of them were positive, since there were organisations that manage information and knowledge in different product development phases. Another respondent distinguished between KM goals and objectives versus formal policy and strategy for KM. The respondent found that even though everybody agreed on the goals and objectives, there was no management commitment, which resulted in no KM policy nor a strategy, and lack of resources, targets and plans. Yet another respondent answered that there was a (KM) policy, since it was encapsulated in one of the four corporate core values.

As a generic result from the first question, the overlapping data should be avoided (in extended enterprise level) by normalising the systems, and continuous learning concerned human knowledge. The ‘re‐creation of data’ could, however, be considered as a kind of knowledge construction



159

in communication through systems. The respondents interpreted the subject Knowledge management somewhat differently. For example, one respondent emphasised the management aspect of KM: ‘if we manage information and knowledge well, we will not re‐invent the wheel unnecessarily’. The other respondent stated that ‘KM should change the normal way of working’, whereas the third respondent focused on ‘continuous learning with the objective of customer satisfaction’. Also the terms ‘formal policy’ and ‘strategy’ were understood differently. One respondent needed to introduce ‘goals’ and ‘objectives’ to express what is working on an abstract level, while another used the term ‘policy’ in a fairly similar sense, without going into daily practices.

About the KM Tools the answers were like ‘KM in product creation field tried to link data with the data requirements through intelligent data repository information management solution. –‐ This is more a co‐ordination/data collection type of thinking method rather than a simple one machine solution for a certain environment.’

From the case II inherited Documentum was playing an important role. All the customer project documentation should have been in a global Documentum database, available for the organisation's personnel in the other projects. But it didn't work globally, some sites wanted to keep the information secret, not available to other sites. The respondent claimed that ‘from my point of view this is a cultural issue.’

With Documentum project the customer documentation representative had a lot of expectations. But unfortunately the ‘performance of the tool is poor and our internal supplier couldn't solve the problems in a proper time frame.’ So they lost the 'time of KM' with the slow tool. Even if role of the tool was 20% and the processes 80% ‘you can't achieve anything on process level if the tool is poor.’ People start to use local network drives and team rooms which are available for a small group locally.

The R & D, however, used different IM tools. They were generally based on text‐searchable Lotus Notes databases which contain documents (usually Microsoft Word). R & D also tracked ‘open items’ and ‘errors’ in Lotus Notes databases. In particular, all of their R & D programmes used a common text‐searchable Lotus Notes database for ‘recording, publicising and tracking errors.’

R & D had formal approaches to practises, processes and responsibilities but they didn't generally refer to it as knowledge management. They mentioned globally the following points:

• For many years now each of R & D product programmes has produced a ‘final

report’ which focuses on lessons learnt during the programme. It was the responsibility of the Product Programme Manager to produce it, generally as a joint effort by a team. In certain areas R & D had recognised that it caused delay to formally collect the information and therefore they created interim reports. Following programmes were required to review this information.

• R & D product programme managers met regularly to report on the status of their programmes. There were both formal and informal information sharing as a result of this regular meeting ‘inside and outside the main meeting’.

• R & D also had a (long‐standing) strategy of re‐using successful designs from one product to another and had set formal targets for re‐use.

• R & D product development ‘Milestone review’ process (a conventional ‘stage‐gate’ process) also required to use one or two independent experts as the key members of the technical review teams. The experts were formally required to come from outside the Programme team. This results in a formal requirement to leverage the ‘tacit knowledge of the expert reviewers’ – as a sort of organisation extension.

The formal policy, hence, as well as the KM practice, process and responsibilities were quite rare, but on the informal policy the conclusion are more revealing. There were designed office concepts to support 'knowledge accidents' (in informal discussions). Also there were established

Systems

'Interestdependiknowledhave theadvice/c

the bestacknowlgenerally

perceivinorganisa

Accordinprovide tany prof

‘...passincommonis sometthey belunderest

move at engineerorganisaarroganmanagem

managerdone thisthem.’ (managerpath som

got concwas an ithe case

Ca3.4

do the pprojects

various ecreationinvestiga

between inf

t Groups' eveng on contridge. As in aneir own persocomment, as

In add of these locedged expery operating a

Thereng new proceational unit t

Everybng to one resthat after yofit’.

In mong on my knn result of thtimes regardieve that it istimate how

Therethe leading rs who wereation will stagce and I losement was to

Perhars usually felis before, becThe ‘build‐ors needed tometimes ‘pai

Initialcrete experienvestigatione III questions

ase III, phase

The reproduct deve(in a strateg

The foeffects of ne engineers’ pated. The act

Th

be

formation an

en with half‐fbution.) Finany organisatioonal networmanagers th

dition to thescal practices wrts whose prat an inform were some esses and meo find 'the bbody saw anspondent, ’if ou have finali

st R & D orgaowledge takhis perceptioed as a low ps either not rmuch their o was a very fedge in any convinced tgnate). Quoe the benefit o achieve a bps it was thelt far more recause I have n’ bit is optioo accept a muid off’. ly, the respoences, plans on state of ts are listed in

e 1: intervie

esearch queselopment enggic partners’ ocus was on w organisatiperspective. tual question

he design of etween the n

nd knowledg

formal objecally, there weon with morks which thehey will knowse there werwas based oimary role wal ‘one‐to‐onobstacles ofechanisms’ west tool' fromnd understoof you have toised the proj

anisations thkes time andn is that conpriority activrequired or nown experienfine balance field: In ordethat they canoting the respbecause I acalance in thiere that knowelaxed whenstudied all oonal.) If the uch higher ri

onse rate wasor other matthe art in then the Appen

ews including

stion in this sgineers worknetwork) shthe inter‐orgional forms, Work practins were:

the inter‐orgnetworked c

ge

160

ctives and plaere many dife than few yey utilise. ‘Asw who to brire local practon the ‘gardewas to circulane’ level. f KM perceivwas not seenm the markeod the imporo provide an Aject and you

he primary od does not adtributing infovity. Sometimnot welcomence could aswhich an R &er to be trulyn do the job pondent: ‘If tccumulate ers area. wledge manan an engineeof the previou‘studied theisk profile, bu

s 10 %, but itterial, they we focal compdix 3, and th

g external co

second set oking in collabare informatganisational and on the dces, work cu

ganisational ompanies?

Part II – In

ans around. fferent informyears of histos engineers ting‐in to helptices which weners’ approaate their kno

ed. For examn as threat, aet. Content mrtance of KMA4 of Lessonare starting

bstacle was dd direct valormation to mes people ae. It was verysist their les& D organisay ‘cutting edbetter than athis self‐belirrors too fast

agement becr told them ‘us solutions e previous solut their expe

t was assumwould have spany, and thehe observatio

ollaborators

of Phase 1 intborative and tion and knonetworks fodistributed wulture, and th

networks: H

itial case stu

(E.g. some pmal approacory, the morethey will knop their currenwere relativeach used by wledge with

mple, ‘culturaas it was natumade the diffM, but the prans Learnt froma new one, u

a common plue to my cura knowledge

are unwilling y commonplas experienceation neededge’ they neeanyone has dief goes too ft’. So the cha

come an ena‘I can do thisand I can bulutions’ elemeriences told

ed that if moshown more e real surveyons continue

terviews coudistributed

owledge?’ rmed by parways of workhe ways of co

How is the co

udies

profit for the ches based oe experiencew who to asnt team.’ ely formalisedHP and Agilein their orga

al obstacles ural for that ference. actices were m a project, unless you w

perception thrrent task’. e managemeto offer advace for engined colleaguesd to achieve eded to empldone it beforfar it becomallenge for R

abler becauss better thanild on and im

ment was misd that permit

ore people reinterest. Hoys gave moreed after surve

uld be summproduct deve

rtner compank from the prollaboration

ollaboration o

members, n tacit ed people allk for

d. One of ent – anisation

in

different. you never

won't get

hat The ent ‘system’ vice because neers to s. in order to loy re (or their

mes R & D

e the anyone has mprove uponssing, tting this

eally have wever, this e results. All eys.

arised ‘How elopment

nies, on the roduct were

organised

n



161

The tools of knowledge work: Is the networked nature of work reflected in the information systems used in the daily work?

Describing the work practices in the new inter‐organisational networks: How do the individual product creation employees experience the networked mode of product creation work?

So people from both the focal and partner companies were interviewed. Fourteen product development engineers and other professionals such as technical writers were interviewed in their working environments. Five of them were working for the partner companies and nine for the focal company. The interviewees were chosen on the basis of being involved in inter‐organisational collaboration, but they did not represent one particular product creation project. Instead they were involved in the focal company’s product creation in different activities, such as testing or documentation. The interviews were carried out in Finland, in the cities of Helsinki, Espoo and Tampere in the premises of the focal company and its partners initially during spring 2003, and follow‐up for about five years. The original quotations are printed on italic and the number of respondent are marked with ordinal character and sequence number in brackets, like (# 1), to enable the reader to identify the same respondents.

The interviews were complemented by observations in the workplace. For these interviews, a ‘rapid ethnographic’ approach was chosen (Millen 2000). Data were gathered mainly by semi‐structured thematic interviews, and these were recorded and transcribed and finally analysed with qualitative content analysis techniques. For example the following KM‐related issues were investigated:

– Creating knowledge – Storing and retrieving information – Re‐using knowledge from other people and projects – Sharing knowledge between people and organisations The themes of the interviews supported the analysis, and the interview framework was

used as a backbone according to which the interview data was processed. The data was gathered asking the questions defined in the interview framework themes, and the answers found in the data were grouped under the rubrics formed on the basis of the themes. The basic unit of analysis was a set of sentences – including the instance of one sentence – forming a meaningful expression of the subject matter.

3.4.1 From organisational to social structures through the extended enterprise

Results from the interviews showed first that the experience of the interviewees varied a lot. Also their styles of discussion were different. Most of them were dealing with documentation editing and publishing. The information and knowledge were about products – the technical details and features that must be managed in real time with external partners. The published content is mostly XML, SGML, XHTML, HTML and PDF (# 10.) XML and SGML are quite close to a knowledge management, where the information and meta‐data are separated and the information will become an independent entity.

Knowledge was ‘shared’ in face‐to‐face situations and through information systems. The role of the information systems was to store the created documents and to enable the reuse and exchange of these knowledge assets. According to the interviewees, knowledge sharing was encouraged by setting common goals and deadlines for project members, but no direct incentives – that would be destined only for knowledge sharing – were used. The interviewees were not aware of established ways of sharing knowledge, e.g. the lessons learnt, which were informal knowledge in most cases, seemed to be poorly documented. That was not a problem in the Cases I and II, but when the collaboration is the main method to disperse organisational learning, the enterprise borders will become an obstacle.



162

The interviewees represented different educational backgrounds. There were different professions, and also the employers were varying. Mainly they were working with documentation, but from different viewpoints. The collaborators had also other partners besides the focal company (referred mainly to interviewee # 1). One common aspect was that all various professions saw the information flow as their duty, no matter what was the subject of their everyday work (# 4). Their documentation was published. The publishing process – in rough level – consisted of planning, editing and publishing. In the half way of the process a draft trial set was accepted (# 10). The information flow and publishing is in the results seen as communication from the information species point of view.

Even though in case III communication there was no real ‘knowledge‐border’ between the different companies (as one could have expected), and information and knowledge seemed to flow quite fluently across the company borders, one clear bottleneck was identified: Project managers and contact persons controlled the flow of information and knowledge. More direct contacts and communication would have been needed between people working in the collaborated projects, since the project managers and contact persons were continuously ‘too busy’ to answer all the questions directed to them. ‘…we have worked together for so many years that we know each other’s ways almost as well as we were working for the same company. We have kind of grown together.’; ‘Sometimes it’s not nice to stay at work late in the evening just to have a net meeting with the colleagues in the States. The time difference is some kind of an obstacle.’; ‘…I personally don’t see a knowledge border. There is a gap but it’s not that bad. We usually get all knowledge we ask for, if we ask for it’ .The partner companies did not always ‘dare to disturb the busy project manager’ and thus suffered from insufficient knowledge. Information sharing was considered difficult due to cultural or geographical distance; meeting people face‐to‐face in the beginning of the project helped to create an atmosphere of trust that facilitated communication and knowledge sharing via e‐mail and phone. The control of information could be seen as one form of power – although shown as business.

In addition to start‐up meeting, the face‐to‐face meetings are considered important only when there are problems to solve or real decisions to be made. Monthly follow‐ups could be virtual meetings via communication technologies (# 4). There were no significant differences between the focal company and the partners in the ways and culture of working. Due to a long partnership there were common or compatible ways of working together. In some cases there were minor problems with foreign companies due to different surrounding culture, languages, or time zones. The programs and projects were quite enormous, and especially working with other organisations and external partners made the knowledge and information management challenging (# 12).

.

3.4.2 Processes and practices in new inter‐organisational networks

The extended enterprise concept is needed here as we talk about new way of collaboration, but some observations could be found when comparing to traditional collaboration. First, the division of work between the focal company and the partners was very accurate. The documentation – in product creation process and contracts – covered everything. There was a specified change management for projects. Project meetings were once a month and all kinds of reports were delivered to various workers. The processes of the focal company were comprehensive. Product creation process was not a reason of problems. The possible reasons were overload of work and lack of time. (# 1) That came out from the comments, such as ‘When I began to work we were in a hurry twice a year… now twice a week’ (# 13).

Most teams investigated were professionally homogenous, i.e. at the focal company they consisted of mostly engineers. In contrast, the technical writers at the partner companies were usually professionals of linguistics or humanistic fields. Therefore consulting technical support was needed in documentation teams. However, there were no significant difficulties between different professions’ representatives, as the corporate culture was the connecting factor creating a relative uniform way of working, even if the different sites also had unique characteristics depending on location and function. ‘Most of us are engineers. There may be some other educational backgrounds, too… But the engineers may represent many different fields of engineering.’; ‘We need many kinds of people, many kinds of personalities, but I’m not sure if we need people with different professions. …and this (work) requires a



163

certain mindset’.’Corporate culture was perceived similar in the partner companies as a result of close collaboration with the focal company, or in some cases the partner being a former part of the focal company. Moreover, there were some persons with hybrid‐skills that bridged the gap between the engineers and humanists.’ In contrast, some of the technical writers reported that the culture was not as relaxed now as the focal company had become the customer, compared to when everyone were working for the same company. This new organisational independence that had occurred two years before the interviews was still very present in the discussions with all technical writers. On the one hand, it had brought about a centralisation of the technical writers away from the designers into their own offices, and facilitated closer collaboration between technical writers. This change had mixed reactions within the technical writers, of which some felt an increased hesitation to call the engineers being afraid of disturbing them, while others were very pleased with the strengthened technical writing culture. In addition to these dilemmas of project continuity and whether to work with engineers or technical writers, the previously mention outer threat emerged in almost all interviews: scarcity of time. Some said that they have to manage with the available resources; others blamed time to be the reason why the communication sometimes did not work very well.

3.4.3 Extended organisational networks

There were many collaborators in an extended organisational network. They, however, could not cooperate without the focal company in an extended enterprise. That was especially valid for communication. More formal channels were needed for information exchange – personal relationships did not count from the collaborators’ point of view (# 9).

Project management meetings helped to share information between collaborating organisations. What comes to an extended organisation, meetings were needed regularly for follow‐up. The material for the meetings was better to deliver beforehand, although they could be shared in virtual meetings (# 10). Usually the focal company did the innovative work related to the product creation. Partners carried out very limited and pre‐defined parts of the product according to the requirements and specifications of the focal company. This kind of collaboration could be seen as an extended enterprise that consists of a focal company and several collaborators that participate substantially in the operating of the focal company. ‘… We usually implement the knowledge the other company (the focal company) has created. It’s a clear division of work. They make the innovation and we do what we are asked to support them.’

This division of work functioned well in most cases, but was challenging when the product changed during development. It was considered normal to put the technical development on hold, while business representatives renegotiated the deals concerning schedules, resources and money. The project schedules were most often too optimistic. Some of the interviewees considered this as overestimation of own capabilities (# 3). However, it could also be underestimation of hidden delays.

These ‘project freezes’ required the partner companies to be very flexible and re‐allocate resources within very short timeframes, e.g. in a week’s notice. Some focal company interviewees wished that they could get more feedback from the partners, who they thought did not report their problems early enough. Insufficient communication or trouble in understanding the requirements might have increased the uncertainty of the estimated work efforts, which then increased the risk of project delays. Little if any collocated work was done between the focal company and its partners. Instead, there were short working visits to the partner site, but the work was done at each partners’ own premises. ‘Quite little (collocated work) as far as I’ve understood. …There are these kinds of information transfer situations. Then they come to Espoo (town in Finland) for a visit and then they go back to India.’

Training was offered to the partners regularly, as well as familiarisation to the product and processes used during the product creation, and a contact person was defined with a responsibility of the partnership collaboration. Usually the partners followed the focal company’s process models, which framed the partner’s own internal processes. The information systems were usually shared to some extent: not all documents were shown to the partners, but the partners were given access to all documents they needed in their work. However, these practices varied considerably between different projects and partners. ‘As far as I understand, all information systems are available to the partners. ‘We



164

use the focal company’s product creation process model. In general it is very clear and well defined. It makes it easier for us to work.’

The processes, practices, and interfaces of collaboration were in the process of development. Two of the interviewees had been in a position where they had to create new project organisations. It was supported by the existing processes, but they reported that there had been some ‘fighting’ to get their process innovations included in the official processes. There was a continuous dependency between documents and products. The discrepancy between realisation specification and the actual (software) product was due to lack of time. But the real reason could have been in the process – normally it should include the document update before the product update.

3.4.4 Organisational memory content in practice

First, the documentation for a project required a plan for documentation itself. This meta‐document was normally a list about the documents, the responsible persons, the deadlines and the purpose of each document (# 7). Most projects reused information from parallel and previous projects in some form or other (e.g. processes and templates). Many projects as a whole were based on the previous projects (e.g. new releases of an existing product), and some projects needed information from other projects. Some processes really could utilise all of their documents in further projects (# 4). Also respondents claimed that the knowledge in the documents was not personally attached. In the case of respondent number four, the case (testing) might be so.

The traditional documents were managed with different tools. Version management included inspection phase and the authors had got personal portfolios for various documents. Also most document bases were replicated to collaborators (# 4). According to the results, information from other projects was difficult to find: the interviewees did not know where it was and who could access it. According to the interviewees, information should be stored in one place or at least it should be accessed through one portal. Even more important would be information system and tool integration. An informant claimed that information reuse from systems required personal contacts: ‘…I’d say that I find the information when I look for it. But it requires a lot of detective work. Because you never know where the information is stored and who can access it. And you have to ask from several persons before you find the person that can tell where the information is and who can use it…’ More rational methods for finding are the search tools, discussed in the system specific sections.

Also a meaningful classification of documents will correlate with ability to find information fast. Whether document should be classified by content, substance or production phase, depended from whom you are asking (e.g. # 4). This was a problem only when using different document bases. The logical classification should be as multiple as needed, and if there is only one physical classification, the only common factor seems to be the time when the document is created.

As mentioned, information was (tried to be) reused from previous projects — if this information was available. In too many cases it was known that a specific document with specific information existed, but it was not known where the document could be found. The old documents could be used as models and they could be modified according to the current project’s needs. For example, project plans, templates, and experiences were reused. More information could have been reused if it was available. The re‐use of information or knowledge still dealt mainly with explicit knowledge and information that could be expressed in writing. However, reusing tacit knowledge was difficult, because it had proved to be almost impossible to capture and to store this kind of highly personal and contextual knowledge. Some parts of tacit knowledge (of e.g. practical skills) could be captured with e.g. video devices. Of course this would be just one method to help to share and learn those skills, not capturing and storing the knowledge itself.

There was also organisational distance between employees. The geological distance was not a big problem, except the different time zones. For example, calling by phone was possible only on certain time of day. The distance was also slowing down the e‐mail exchange. Replies were sent once a day.

With the partners, most of the communication was based on e‐mail. The colleagues working in the same building were closer. Even though the external partners were met every now and



165

then in all kinds of audits and meetings, they were still distant. That organisational distance reflected in all communication (# 7).

Communication between different sites was always more difficult compared with communication inside one site (# 4). The ‘distance’ was also dependent on the purpose of the cooperation. For example, standardisation gathered all kinds of cooperating organisations together. However, the motivation to participate was not as clear‐cut as it is in business collaboration. All the participants were voting and making decisions, even though their competence areas were not necessarily known (# 6).

One phenomenon in collaboration was that there seemed to be one person who became as a contact point – was it meaningful or not. When making first contacts with the partner, the persons, who gave answers, were contacted regularly. Of course this was a risk, when the person gave answers based on vague knowledge. The comparison with the right expert could not be done while he or she was away (# 7).

When a new collaborator came to a new project, the organisational collaborating methods were just beginning to formulate. In that kind of cooperation the lessons learnt could have been reused in quite a narrow sense (# 8).

Completely virtual work with no face‐to‐face meetings was not considered feasible, because the current information systems were incapable of supporting the sharing of tacit knowledge. Tacit knowledge is local, contextual and closely tied to place; reconstructing tacit knowledge in a new context is difficult. The sharing of partial tacit knowledge could have been facilitated by supporting communities of practice with the help of information platforms. When this implicit knowledge is distinguished from tacit knowledge, it is possible to derive explicit knowledge from implicit, in discourse and collaboration. The explicit, implicit and tacit view is not meaningful in constructivist view to objective knowledge in information systems. This also guides as towards a new concept of knowformation presented later in the results.

Explicit knowledge from history as well as current documents from other people helped in new challenges. ‘We usually work so that we move from release to another. So we can use the test cases etc. again in the next release. We can use it as a model and make a new version.’; ‘…I use the process documents and the slides from other projects. It simply saves so much time when I don’t have to make it always from the scratch…’When storing and sharing this kind of knowledge, e‐mail and CSCW tools were an advantage. Although implicit knowledge is subjective, and has no firm factual basis, it is cognitive and therefore able to be made explicit. That is the case also on organisational level. Information and communication technology helps to move the borderline from implicit towards new explicit knowledge. It should be noted that the tacit, implicit or explicit knowledge were only analysed by the researcher – the respondents considered information and knowledge mainly the same, even when talking about information from the systems or communication with colleagues.

3.4.5 Knowledge capture and reuse

The knowledge capture is related to organisational memory, as the verb ‘capture’ in this context normally refers to memorising or learning. For selection what to memorise by predefined process there could and should be rules, as the natural selection of human memory could lead to secondary factors influencing, instead of pure information itself.

The examples from interviews lead to the direction that all new ideas were not captured. Normally a new idea is presented by an individual, but the audience is revising it (# 5). On the other end, one practical kind of reuse on data‐to‐information convergence mentioned was project work amount planning. Based on the existing statistics of working hours, a new project work amount was measured (# 13). This required experience‐based knowledge on how to compare tasks between different projects.

Knowledge reuse could be interpreted as a target for organisational learning. Knowledge reuse was not as trivial as reusing data. Learning means cumulating knowledge – any lessons learnt can not be copied from one project to another (# 6). The generic substances – e.g. technologies and tools – could be taught and studied, but more project generic findings were harder to share (# 10).



166

Information that was poorly documented in the systems was living with experiences. Those are somewhat emotional, and could lead to prejudices, as commented by interviewee # 12: ‘this has seen before, try something else’. In generic language, the distinction between human and artificial memory (retrieval from the systems) could be seen in analogy with knowledge and information.

Project‐specific knowledge reusability is restricted into processes and documents that are moved (not copied) from one project to another. That is happening when the substance and schedule of the project is modified (# 12). The project itself is changing to new, but the knowledge is not transferred from a finished project to a new one.

What comes to knowledge presentation, the result is that (like in philosophy) there is a separate communication discipline, and the ‘presentation’ should refer to storing of non‐interpreted information (including knowledge).

3.4.6 Knowledge creation and feedback

Knowledge creation was mentioned using learning new skills, and new documents about inventions as examples. The most practical knowledge creation examples were models and concepts about technical solutions (# 12). So creating knowledge was understood in various ways depending on the interviewee. For example, to create knowledge referred to inventing some new product or method, to getting a new idea, to producing specifications and other documents, and to coding innovative software. The knowledge creation was seen as an interactive process between individuals and a team; the ideas formed in someone’s mind were further refined in one’s team, and the results of teamwork and discussions could be further refined as individual work. ‘…I’d say that we create knowledge when we create a new product. Then we make something new in a concrete sense. But of course we may invent new ways of doing something. I think that’s creating knowledge too.’; ‘Creating knowledge…I think it happens in your head, but you need the rest of the team for it. It’s partly inventing something on your own but you have to work it out with the colleagues.’ And from a collaborator’s point of view: ’…we usually implement the knowledge the other company has created. It’s a clear division of work. They make the innovation and we do what we are asked to support them. Then of course we may create new ways of making our work in a better way, but that’s a different story.’ In most interviews the knowledge creation was located at the focal company. The work between the partner companies and the focal company was divided so that the latter would develop, create and make innovations (e.g. product architectures, specifications — in some cases innovative, but not always) related to the product. The partners would implement some well‐defined part of the product. On the other hand, competence on the environment of the focal company, products, and tools accumulated.

The most fruitful phase for knowledge creation was in the beginning of a project. New ideas and solutions were generated in discussions, although not necessarily presented immediately. Also existing individual ideas could be ‘published’ in discussions. For example, during the specifications requirements and implementation specialists could teach each other, but also generate new solutions (# 5). Meetings were generally mentioned as fertile situations for knowledge creation (# 12).

Actually, this indicates a result that knowledge creation and sharing live in parallel. The most fruitful knowledge creation is done in communication, same time as this newborn knowledge is ‘shared’. However, the rare knowledge‐related gems could easily drown to all other communication – and justification or truthfulness of this shared belief is rarely required. Strange was that most of the interviewees interpreted knowledge as being tied up to physical products and other artefacts, even as ‘implementation’ of external knowledge. Creation was not seen as an initially iterative generation in discourse as in collaborative engineering. Also the engineering aspect emphasised practical ‘doing’ that was going hand in hand with knowledge creation. All of these observations supported a new division of tacit vs. explicit knowledge, with an implicit layer including short term holistic knowformation in organisational consciousness or cognition. This will be explained more in detail on later chapters.

Some of the interviewees thought that they are not receiving enough feedback from their work (# 5). Feedback does not come especially from collaborators to the focal company – the other way round feedback is somewhat given. One good kind of practice with the feedback is first to give positive feedback and then giving proposals for improvement. It is not fruitful to tell what went wrong, without



167

telling how thing should be done (# 4). The feedback was a problem also in focal company (# 12). Feedback from customers was delayed. One reason might have been that the customer did not know what she or he wanted. The technical details have ‘triggered’ part of the feedback.

3.4.7 Knowledge sharing

Knowledge sharing with or without systems was discussed in many instances and here are some of the interviewee’s opinions. First, it was considered crucial to separate unofficial knowledge sharing from information flows. Information flow is a designed kind of knowledge sharing process. The shorter the chain is, the better the right information is in the right place, in the right time. ‘If the information is coming from the customer, it is not coming fast – many people have buffered and processed it, so it necessary is not the same compared with the original information’ (# 12). ‘The problem in knowledge sharing with partners is not that the information isn’t shared, but there are delays‘(# 13).

Another formal example from knowledge sharing was associated in software interfaces: these kinds of specifications should be given to collaborators, including the knowledge on how to understand them in practice (# 12). There were signs all over that openness and knowledge sharing was important and also fulfilled in practice. There were official organisational channels for knowledge sharing, the importance of which is varying depending on the company (# 12; compare with # 5 & # 1). The meetings were categorised to the official channels .However, one interviewee noticed that knowledge sharing depended on the economical situation, and some level of ‘hedging’ could be noticed. He also mentioned teaching as an example from knowledge sharing that is systematically rewarded (# 12).

Many interviewees associated knowledge sharing with information systems. ‘The tools (of the focal company) are in good shape’ (# 5). However, there were also personal relationships on the first place. ‘It helps to share information if you know the people better than by e‐mails. Information systems will ease the sharing if they are working’ (# 3). ‘Working’ here could be interpreted as ‘used’.

Also e‐mails, meetings and announcements were considered high in association to knowledge sharing (# 4). It was often acknowledged that the announcements should be made better, but the problem was, as in most knowledge sharing cases, that people need different information. One obvious fact was that most of the information about announcements was dispersed unofficially before the official news. So the unofficial dispersion was quicker than official.

Knowledge sharing in a traditional way was mentioned as interviewing more experienced colleagues. This was normally done during a project start‐up. ‘During a project, the situation before the milestones is asked for, as well as many kinds of trivial details, mostly by e‐mail’ (# 13).

Extended organisational knowledge sharing required rules. First, the targets of the cooperation must be defined. ‘The roles and rewards for each collaborator should be written down, to avoid misleading expectations and assumptions’ (# 6).

All the respondents argued that the focal company supported knowledge sharing. However, the concept was understood differently. While others talked about sincerity of people, others associated it with rules and tools (# 6 vs. # 7). Concerning external partners the knowledge sharing was even associated with patenting (# 7).

The sincerity of people reflected in knowledge sharing. It seemed that openness was tough to be taught. Employees participated to technical courses, but very rarely on any courses on human subjects. Attitudes were reason for the lack of transparency as well. The values of the company did not necessarily support openness: ‘Everyone has his or her own objectives’ (# 3). ‘Concealing’ of information was also possible, but it was highly dependent on the sort of information in question (# 4). Knowledge sharing should have been rewarded with incentives. Experiences from knowledge sharing bases have given good results (# 6).

When talking about knowledge sharing it is crucial to make a classification what kinds of knowledge we are talking about. In a competing organisation, there are certainly ‘sacred’ content that is not shared. They are talked about in inner circle, and personal levels they are not meaningful to make public. Based on the interviews, the most common excuse was ‘lack of time’. It could be assumed that ‘business’ – as a real reason or as an excuse – is causing business secrets.



168

In general, the knowledge sharing motivation is fuzzy. The employees’ are on some level aware that the final aim for knowledge sharing is to reduce the number of workers. On the other hand this is keeping the private sector enterprises in shape, able to compete. Another negative result is that neither lessons learnt nor bulletin kind of one‐way communication are not working as for knowledge sharing because the information requirements are varying so much depending on the receiver.

Geographical friction in knowledge sharing was a practical obstacle mentioned in many interviews. The collaborators used to work at the premises of the focal company before the interviews. Nearly all the respondents said that working in the same premises would have helped to solve problems. On the other hand, all the employees were so busy that the usefulness of the same premises could be small (# 1). Face‐to‐face meetings were needed for common understanding. ‘Depending on the distance of the collaborators, regular meetings should be organised – every other week or every other month’ (# 6.) All in all, the same premises make the sharing easier.

3.4.8 Trust, culture, education and language

One interviewee condensed opinion to one sentence: ‘Trust is the most important basis for knowledge sharing: To share your knowledge, you should be confident that the audience is not using it as a weapon against you’ (# 6). The focal company encouraged to work in groups (# 13), which required trust. ‘There should be a (preferably more experienced) colleague supporting your opinions… and there are plenty of meetings where you have to validate your opinions’ (# 13).

Cultural differences were observed between countries, but between organisations as well. Some of the partners have modified their working procedures in common projects according to the focal company’s requirements (# 13). One interviewee claimed that education is quite similar all over Finland, which makes the situation different in small collaborating companies that employ mostly Finnish students, compared with the focal company, where there are people from different cultures (# 12). The difference is even bigger with foreign collaborating companies.

Language is not the only cultural difference. Collaborators from various countries are different. The culture of countries is reflected in working cultures. The approach to duties varies. Most of the interviewees saw cultural differences as a disadvantage (See e.g. # 3, # 7).

One difference between people is how binding their promises are. This can be due to another disadvantage that interviewees mentioned: To accept own mistakes and their consequences is not equally easy for representatives from certain cultures. In one interview opposite countries, like China and USA were considered similar in that sense (# 4). This cultural problem is serious especially in software business: the earlier a mistake is found, the less it takes time to clear a fault. The (negative) feedback and commitments should be given as early as possible (# 12).

In parallel with culture, some countries had practical infrastructure problems with networks. For example, e‐mails must be encrypted due to security reasons in certain countries (# 12).

Educational background gave one basis for knowledge construction. Language barriers did not constrain communication only between people from different countries, but also between different professions. When people from different educational backgrounds talked about the same issue, it could sound very unclear. One interviewee recommended to impugn own understanding and discuss the same issues within smaller circles (# 6). To confirm the understanding it was recommended that the e‐mail correspondence is not used and that the discussants are from the same professions and from the same (professional or educational) level. For example, a marketing person and a technical expert could discuss fluently, but the outcome is often facile (# 6).

In Finland the technical writers were language students (linguistics and humanists), while in other European countries engineers. Even though the technical writers had somewhat studied technical sciences, they were making a lot of questions to engineers. Engineers saw the documentation for customers quite as non‐productive work (# 1).

Humanists and engineers felt that they did not understand each other. The linguistics felt there was no problem to communicate with each other (# 10) but from the engineers the same results could not be found. The engineers found this more like an individual matter (# 12).

Systems

other edin the inand marconstantmust be

engineerscope: ifwhile mo

harmful foreigneconductdiscussio

of Finns.

3.4.9 S

growing were chathat whaKnowledwere noall the timotivati

organisamotivatimore anable to cwas possgeograp

companincreasininterviewwas criti

Ca3.5

requireme.g. grouand twopinpointwas stud

knowled

between inf

Most ducational baternal networketing profet discussion wunderstood

It wasrs, even thouf you know yore wide edu

On lanfor knowled

er (and the reed in Englishons or to ask

The ca. Most of the

Summary of

The tuand changinanging in all atever is dondge managemt always utilime the conteion.

One cational chartsion (# 4). In pnd more usedconstruct comsible to solvephically.

The reies involved ng requiremewees underscal to receiv

ase III phase

The thments gatherup elicitation sets of intert the positivedied in many

As predge, there we

formation an

often it was ackground inork. Other gressionals. To was required. The interprs a surprise tough the humyour narrow sucation was nguages, thedge‐intensiveest of the woh. Some issuek questions. ase corporate respondent

most impor

urbulent chang heavily nodepartmentne can be losment can be ising the systext is lost’ (#

onstant phes. The social practice, virtd as an extenmmon undere problems m

esults of thisin this studyents for intestood it) was ve the meta‐i

e 2: one dedi

hird case inclring to the acn sessions to rviews aboute and negativy ways, includesented in Fiere initially t

nd knowledg

assumed thn their team roups, in addachieve undd (# 6). Also tretation of eo find out th

manists felt thsector well erespected asere was an obe projects. Oorkers Finnises were not u

tion uses Engts admitted t

rtant notes

nge of the foothing is consts, as well as st. The same formalised wtems as supp# 14). The ch

nomenon inmeaning of ual collaboranded enterprrstanding as more easily, c

case study hy. The lessonr‐organisatiodynamic, aninformation

cated KM sy

luded lots ofctual systemfind out andt KM in geneve features. Tding e.g. Silkgure 38. Inithe data, info

ge

169

at technical (# 4). Technidition to prevderstanding bthe importanxtended enthat in some chat they werenough, noths a framewobservation thne typical sith) but becauunderstood

glish as the othat it is mor

ocal companstant. The orthe contactsthings must with the systposed. ‘In anhange for cha

turbulent chmeetings waation of experise. Howeveeasily as peocompared wi

have been uts learnt mayonal collabornd also lots oabout chang

ystem – WIS

f interviews om evaluation. d compare theral. Also theThe usabilitykTest. tial plan of caormation and

Part II – In

specialists feical professioviously mentbetween sepnt but more erprise conccases a genere not respeching more is rk for detailehat the variotuation was ause of that tinand it was n

official languare fluent to d

ny was a big crganisationas in external be done agatems only to n organisatioange’s sake c

hange was thas affecting terts and onlier, people sitople meetingith problem s

tilised for pry also be usefration and neof informatioges just in tim

E portal

on engineeriAfter the stahe requiremee system evaly and techno

ase III to inted knowledge

itial case stu

elt they do nonals were thtioned humaparate professilent interecept used heric educationcted (# 7, # 4required in ted knowledgous levels of a project whny minority aot so easy to

age, while thdiscuss in Fin

challenge. Inl structure wcollaboratorain within ona certain po

on where peocaused futile

he unofficialto knowledgne communtting in differg regularly. Insolving of pe

ocess develoful for other etworking. Kon changed inme.

ing knowledgate of the arents for an eluation framological perfo

egrate data, ie layers in th

udies

ot need peohe only ‘consanists, were essional groupest group – cure includes cn was respec4). It is just a the everydayge. language skihere there waall the meetio participate

he headquartnnish (# 4).

n a company was changingrs. There is ane or two yeaoint. Changinople are chane work, as we

relations bee sharing anities of practrent premisen ad‐hoc meeople dispers

opment withcompanies f

Knowledge (an time. In en

ge managemrt survey, theengineering Kework was uormance of t

information e plan. In the

ple from stant’ group economists ps, a ustomers – customers. cted by the question of y work,

ills could be as just one ings were to

ters are full

which is g; the peoplea big danger ars (# 14.) ng workers nging places ell as loss of

ehind the d tice was s were not etings it sed

hin the facing as the ngineering it

ment, from ere were KM system, used to the system

and e focal

e



170

company, however, the system aspect remained on the prototyping, while the human factors were considered more valuable.

Figure 38. Initial plan of case III to integrate data, information and knowledge

The human factors and organisational results were processed more intensively –

especially the interviews gave results how to enhance knowledge work processes. On the long run the results of the interviews were beneficial.

Project software application was used as a pilot. The expected benefits were to give a concrete example from a prototype system for engineering. The evaluated version was not used in production because it was not enough specific for the focal company. But it was beneficial as it gave results for this research and showed generic requirements.

As case III was from the world of engineering, it referred to concepts from that discipline. For example, Capability Maturity Model Integration (CMMI) definition for engineering says: Systems engineering covers the development of total systems, which may or may not include software. Systems engineers focus on transforming customer needs, expectations, and constraints into product solutions and supporting these product solutions throughout the lifecycle of the product (CMMI: FM108.HDA101.HDB105.T101). As mentioned, the project produced also a KM system for engineers. CMMI defines that Software engineering covers the development of software systems. Software engineers focus on applying systematic, disciplined, and quantifiable approaches to the development, operation, and maintenance of software (CMMI: FM108.HDA101.HDB106.T101). Maturity seemed to be a critical question already in organisational memory.

One target defined in the scenarios was to adapt a new way of working, defined as follows. Each task of the conventional design process in engineering was supported by a preparation step and completed by a review step:



171

• Classify task and context • Search for suitable best

practices (solution and process)

• Search for lessons learnt • Search for other

information • Select and adapt best

practice(s)

• Apply knowledge to task (manually or automatically)

• Action breakdown into subsets, with own preparation and review activities

• Generalise and classify the task and the context • Identify, discuss and describe the lessons learnt • Identify and describe other useful information • If action is best practice:

describe action in a general adaptable manner • Best practices with expected high repetition rates:

formalise knowledge for computer execution

• Some lessons learnt (what went wrong) should be formalised as denials

So far we can sum up that the lessons learnt and best practices as concepts were visible in all parts of the focal company, as well as in its partners’ definitions. These were common methods for extended organisational knowledge sharing in the beginning of case III project. One key theme of case III interviews and system was knowledge sharing.

3.5.1 Requirements for case III KM system

Case III experimental system was integrating and exploiting existing state‐of‐the art approaches oriented towards the needs of industrial users, so the project prototyped a meta‐system for different kinds of information sources. The three industrial user organisations came up with the following system architecture figure and list of requirements.

Figure 39. Case III system architecture

User requirements were analysed in many phases, but could be summarised as: ‐ Portal‐like access to essential knowledge and information sources in your organisation –

‘single entry point’ for different applications in order to avoid time and system‐resource intensive changing between tools and services when performing a task.

Action Preparation Review TASK

Case 3 System Architecture

Case 3 KM system supports engineers with context-sensitiveknowledge management functions based on web technologies

DB1 DB2 DBn

knowledge base distribution mechanismknowledge base distribution mechanism

storagemanagement modelling indexing

Eng. toolplug in 1

WISEKnowledgebase

WISEKnowledgeserver

Rulesdatabase

capture re-use monitoring

Process flowcontrol

Queryengine

display managementdisplay management

Mobileclient

Web Browser

WISEclients

Adm

inis

trat

ive

fun

ctio

ns

retrieval

Eng. toolplug in 2



172

‐ Finding experts and ask questions: provide an easy access to the name of contact persons in a special field to allow a user to easily ask him a question; automatic storage of the answer of the expert.

‐ ‘Who can help me?’ This function takes into account the work context (roles, project phases etc.)

Search for information in a tree structure, which reflects the organisational logic – either by keywords or by selected meta‐data values.

‐ Similarity search: the user can find also objects with meta‐data values similar to the ones specified.

‐ Context‐sensitive presentation: display first the information that corresponds best to the work context of the user. Context (or view) can be changed by the user.

‐ Notification (pull): subscribe (and unsubscribe) to notification about changes on specific ‘knowledge objects’ (or on types of knowledge objects specified by specific meta‐data values).

‐ Notification (push): subscribe others to receive information about specific knowledge objects.

‐ Annotate existing knowledge and information: a user should be able to annotate a knowledge object. There is also a flag indicating the privacy level of this annotation. (Private, personal, limited to a workgroup…).

‐ Basic document workflow functions for each knowledge object to undergo – in this specification creation or import, validation, publication and dismissal.

‐ Tracing: trace all the knowledge used for a specific task specified by the user. For this the system needs to record who uses which existing information objects for the creation of new objects (Design memory). This is important for assuring consistency of product information and thus product safety. Office functions: provide: e‐mailing, printing screens or documents, copying and pasting between applications.

‐ Integration to engineering applications, e.g. opening a CAD part after searching it with case III system search functionality, creating meta‐data for CAD parts being modelled

‐ Version management: system must be able to distinguish between different versions of a knowledge object.

Administrative functions: ‐ Build a tree structure (ontology) to categorise documents. ‐ Build keyword list for different knowledge domains. ‐ Define sets of meta‐data for different types of knowledge objects (which information

needs to be stored with a knowledge object in order to re‐use it efficiently). ‐ Modify and complete meta‐data of knowledge objects. ‐ Access and modification rights to different knowledge objects and domains for specified

persons or groups. From case II to case III the subject changed from documentation to content management

and especially engineering knowledge management. So the previous list of requirements was defined when moving from document to content management.

The appearance of tight relations between knowledge objects and authors (references, notifications and other human/system requirements) were later in this research discovered as another indication of new knowformation concept15. The communication with context on information presentation was used in case III knowledge management portal as summarised in the following picture.

15 As the IT research have traditionally dealt with artificial phenomena rather than natural phenomena, the new concept of knowformation (between human knowledge based communication and content in the systems) can be located even in the information species based on probability (see Figure 6 and March & Smith 1995).



173

Figure 40. Starting page from case III prototype

The remarkable features in the screen are context, collaboration, knowledge base, knowledge objects, searching possibilities, some of which will be discussed elsewhere in this book. As the requirements indicated, many new features of the prototype were based in knowledge sharing.

3.5.2 Knowledge sharing – with or without systems

Knowledge sharing was a popular topic, although it did not seem to be as idealistic as it sounds. Sometimes people used the term knowledge sharing even for selling and buying processes and for ordering things from the suppliers.

From one viewpoint, people can not share knowledge, as specified in Chapter 2 Information and knowledge species. People construct their own model of the situation and convert it to communication after which the person, with whom they share knowledge, will construct his or her own thoughts about the original message in accordance with his or her own interpretation of reality. However, the conversationalists guide each other through their insights to help to see situation better. To do this the partners must know something about each other — about the problems they are trying to solve, the level of detail they need, and even something about the style of thinking — to interpret those insights.

The knowledge useful to novices is different from the knowledge useful to experienced practitioners. Sharing knowledge requires construction and context — who will use the knowledge and for what purpose (Mc Dermott 1999: 108). Language and linguistics are crucial factors in communication. The communities’ terms should be used when organising information in a system. The organising is then more natural. Since knowledge can be interpreted as the sense people are making from information, it will help if information system supports knowledge construction on more abstract level.



174

People evidently observe that there is rarely one single truth when they hear presentations and find more information themselves by searching. The key factor, when making information easy to find, is to organise it according to a scheme that tells a story in the language of the discipline (see e.g. Mc Dermott 1999: 114). Finding and memorising were also investigated with case III prototype document structure.

Collaboration by systems

Case III system was planned as a portal to existing systems – with value added. The collaboration side of knowledge management includes many systems. Some definitions emphasise that knowledge‐based strategies should not focus on collecting and disseminating information, but rather on creating a mechanism for people to reach out to other people. According to this view, knowledge belongs to communities. Organisationally thinking, individual knowledge is not as important as shared knowledge. But individuals are constructing their worldview based on the information retrieved from the surrounding world. Knowledge is based on subjective and objective sides of a worldview.

Non‐formal knowledge communities were natural way to share knowledge. In general, organisations are full of communities where people discuss, share their information, help each other, and form opinion and judgements. Organisation’s ability to leverage knowledge involves finding, nurturing and supporting the communities that already share knowledge about key topics. If they are too formalised, learning communities can become bureaucratic structures, keepers of the ‘official story’. Communities supported with systems give opinions based on presented information.

Focusing on the topics that are naturally important for the participants give remarkable motivation for follow up. On the other hand the topics should of course be strategically important also to the organisation. Since the development of communities takes effort, the benefit both to the company and to its employees’ gives motivation.

Organisation should support the infrastructure for a community, but in naturally motivated communities, members take care about the community. In practice this means holding the community together, keeping people informed what others are doing, and creating opportunities for members to get together to generate shared ideas and opinions. Intentional communities, however, need to define coordinator’s role. If the role is not accepted, the survival of the community is not so sure.

Systems behind the portal

First, the results from the interviews concerning existing systems are presented. Some (not only one nor too many) systems are normally used, but preferably from behind one portal interface. ‘The main benefit from the systems is that they exist. They make the participants do what they have agreed on’ (# 13). This was happening on the group level and on the project level. Separate systems for separate documentation were required: groupware, word processing, e‐mail, content management (e.g. Documentation WorkBench) (# 9). The usability of the systems should be tested with enormous loads of objects (# 12).

Collaborators and the focal company used similar systems, shared systems and different systems. Organisations learned about the content and the systems by doing their own work (# 8). The same information could be stored in different systems during the document lifecycle: Initial drafts in portable machines or personal computers, shared drafts in file servers, shared internal versions in groupware, and published versions in intranet (# 10).

One general assumption was that a generic portal is an advantage (# 13). However, it should be criticised, if there is no added value. Maybe one portal makes the information sources harder to memorise. The context is part of content in human memory.

Normally collaborating organisations have nominated key users (who can help other users) as well, even though the systems are selected by the focal company (# 9). In some projects the



175

file servers, content management, groupware and other systems can be accessed by externals, but administration is managed by the focal company.

There were plenty of systems related to knowledge management explicitly and implicitly: requirements, project management, testing, error and more generic information management systems. There were lots of them, and the systems were also constantly changing.

Collaborators used e.g. different project management systems compared with the focal company. Some of the newly evolved systems were not stabile, e.g. XML document management systems. However, organisations learned experienced working methods, tools and processes (# 10). Extended organisational learning was happening based on projects with collaborators, not only within one organisation.

Project planning was one application area using groupware. It took an advantage from the use of (groupware) information systems: The work could somehow be done, although there were not enough resources. Specialists did their work independently from assistants etc.

Document management includes version management also in content publishing, e.g. with Mozilla document management or Documentation WorkBench. Groupware was used also for content publishing (# 14). Most simply the documents were stored in a shared fileserver. This could be a solution in a small group with no strictly defined process (# 6). Program code included some level of documentation per se, but that was available mainly for other programmers (# 7).

Some Lotus Notes based groupware applications were available for external partners. However, the documented information must be verified by representatives from the focal company (# 1). The groupware document bases were normally replicated for the collaborators, and also other kinds of databases could be ‘prefabricated’ to partners.

Some of the respondents claimed that the use of groupware was decreasing. Seemed more like the solutions were finding their place, and groupware was not used only as a documentation tool. Groupware was used also for example error management (# 9). Groupware applications were available for externals mainly through web technologies.

There are various documentation management systems. It is not worth to change anything just because of the change. The tools should have something in common to reach one set of information. The intranet portal is just the beginning. Document templates varied depending on whether Extended Mark‐up Language (XML) or Standard Generalised Mark‐up Language (SGML) was used (# 2). For publishing purposes, the focal company had Document Types Definitions (DTD) to be used by the collaborators. The documentation systems of the focal company consisted of so many variants, to make the published documents look similar for the readers. More mature SGML tool support was transferring towards XML (# 9).

The interviewees made presentations (slides) based on the information from various systems. In that sense the organisational normalisation was not working. But the interviewees notified that human knowledge was derived from the compilation of the presentation, based on the data and information on the systems (# 1).

Some of the systems were shared with collaborators, so that a set of documents was available online outside the focal company. However, in rare cases e‐mail was used for delivering of documents, especially with a list of recipients from the focal company and collaborators. This was seen as a problem (# 5). Too active use of e‐mails can easily become a substitute for real document management. And the lists of recipients were not always up to date.

One common finding with the first case is that knowledge system should include some sort of intelligent reporting, such as ad‐hoc time schedules for product accomplishment by feature, linked with error management – not only realised numbers or alike. This kind of reports can be linked with testing and project management and reported formally. Also the minutes could be linked with reports (# 8).

There are systems that the workers are ‘forced’ to use – e.g. tracking of working hours. These systems could give some added value if they are used for more intelligent reporting. The systems need to be developed further on – there were some real challenges, reflected in comments, such as ‘I can not do without systems, but I would be happier if I could’ (# 13).



176

Personal workspace was accepted by all interviewees. One reverse side of personal workspace might be that the meta‐data about applications is different, and does not help organisational learning and memorising. Interviewees claimed also that unified storage of documents is another advantage. ‘All information is available from the same source for every project.’ This is somewhat the look‐and‐feel for the end‐user, not the technical solution. Nearly all respondents considered one portal as a benefit (cf. # 2). One portal was already impugned: The solution based on one portal might lose its value because of information overload. What is the advantage to get information from the same source, if it should anyway be differentiated into separate entities based on optimal (not maximal) capacity of human memory?

Notifications and traceability

Notifications were mentioned two times in case III – first on interviews concerning requirements, and also when testing the case III specific application. Notifications from every change were not considered useful. More important was to define to whom the notifications are sent (# 1). But automatic notification is an opportunity, when e.g. schedules are updated.

Also during the inspection a notification from opening a document was considered important. The notification, however, does not mean that the person is actually reading the document (# 4). Some innovative ways to use existing technology were possible, like shared mail folder, which could be used for notifications about new or updated documents. The shared mail folder could easily be used as a newsgroup or bulletin board (# 6.) Now also blogs are used as personal kinds of news services.

One difference between voice and text based communication was the traceability. It was not so sure if the tracing feature was really needed. Some of the respondents considered it as a ‘nice to know’ feature. It was another kind of feedback, that someone has read (or opened) your document. According to many comments, this feature could be used – but the interviewees didn’t specify for what purpose it will be used (# 12).

Case III portal included many ways for tracing (like annotations, add to favourites, subscriptions, views and e‐mailing) as presented in the following picture.

Figure 41. Tracing as context logging in WISE

One idea was that the informal discussions should be commonly memorised (# 12). This would partially solve problems in everyday communication, but also with ad‐hoc meetings, where no



177

minutes are written (# 12). Notifications on document creation were considered more important than notifications on reading (# 4). This was mainly of project management point of view – it was considered easy to check whether a certain person has published a document. Traceability was also relevant to for organisational memory and searching.

Detective work – searching and finding

Information systems provide various search methods for the end user. Classification is the first step to make the content retrieval easier. Groupware documents were classified directly, e.g. by author, project, or document status or class (# 8). Then there were various search functions in different systems. However, the information search differs from search of knowledge.

One threat for a portal – where various information systems are accessed through a web browser – is that the actual source of the data is lost. One interviewee committed that from the web (intranet) it is even more difficult to find the right information, compared with groupware (# 12).

Web search engines lack possibilities.

• The context of words are missing • Boolean operands should be available • SQL‐type of queries should be utilised, but without a cryptic syntax • Wildcard characters, like ? and * are missing from some search engines • The conjugation and inflection of words is important in certain languages • Irregular verbs, plural forms, compounds, and other irregularities are common in

many languages • Synonyms and homonyms are not processed • The distance of the words is ignored • Meta‐data keywords are missing • The classification of the results (not only by the document format) will give a hint

about the context. As one interviewee stated, ‘web search is giving results, where there are ‘twenty

thousand links of which one is relating to your search and all the others are something else’ (# 12). It is a different question, could that kind of random flow of associations be sometimes useful. Formal search and informal stream of machine‐given connotations could be separated and used for different purposes.

The web search in the WISE portal has been presented already in Part I Figure 2. Search based on meta‐data in external documents. The results require quite much knowledge on the context, like shown in the following picture:



178

Figure 42. Search results require context

So the portal based advanced search could be complicated. But there were different problems with groupware. It was not enough that the documents were difficult to find. Even the document bases were hidden in groupware environment (# 12). Information systems could provide answers to most of the questions when the user dedicates his or her time to this ‘detective work’. However, more meaningful is to dedicate time to updating the documents in a real time when the changes happen. It should be noted that this ‘detective work’ metaphor was adduced by both a focal company and a collaborator employee (# 10).

In large collaboration projects it was considered crucial to dedicate time to human networking. Partners must know each other – ‘to whom to call and ask’. The expert should answer immediately, find the answer from the systems he or she is familiar with, or delegate the question forward. ‘If you must dig something from groupware, you can put the task aside immediately’ (# 12). On the other hand, the verbal result is not confirmed by any colleague, manager, verified document or alike, so the correctness of the answer is an open question.

Another result is that there might be some level of insecurity –The interviewees talked about information search with a highly personal binding. ‘I’ve been in this business so long I know the people’ (# 14). ‘It is necessary to find the one who knows’ (# 5). Too open communication caused insecurity especially in extended enterprise. On the other hand, finding the right information from documents without asking anyone required experience. ‘Even though projects are defining information sources, there are always replicate documents with slightly different viewpoints’ (# 2).

Documents are quite easy to find when the substance is narrow enough. For example, documents on testing could be centralised (# 4.) In that context, it is also easier to know what is not published: when a document was found missing, the supposed author was contacted immediately. However, this requires two pieces of meta‐data 1) awareness that the document is missing and 2) who the author is.



179

The subject, phase and the author were considered quite important for searching (# 4). It is interesting, that the author was ranked as top priority, even though information on author is solely meta‐data, and often changing. The reason could be communication. Finding the relevant person is considered crucial e.g. ‘because there is always something to discuss, and without the name, it should be searched from other contacts’ (# 4).

On search engines, it was stated that ‘Information search should at least cover intranet information sources’ (# 10). However, this can be also meta‐information for the users — it is more practical to know where data is stored than to try to trust the search engine.

Access to document bases

The access to documents was depending on restricted network access to raw documents (# 8), and on document replication. Often the solution was a mixture: external access to replicated document base. Both methods required mail support, telephone support and courses. The restricted access could be either through the network of focal company, or through the network of collaborator (# 8). There were also web pages accessible more widely, not including confidential information, but e.g. generic instructions. Documentation processes were described in many parallel documents. If the requirements were defined well enough, no help from information designer was needed (# 8).

There were so many document bases that for one person the information was impossible to follow‐up (# 12). Maybe exaggerated, one side note claimed that the lifecycle of the document bases was faster than the lifecycle of documents themselves. In some cases the customer saw only the documents, while in other cases they were using the system as well (# 13). This was depending on access rights.

Knowledge bases could be tied up with special interest groups. Those were utilising different techniques – news, intranet pages, groupware. They were dynamic – flourishing for a while, vanishing, and popping‐up again (# 14). That was already a sort of communication more than documentation.

3.5.3 Communication – from internal technology and meetings towards extended organisation

Communication seemed to be the most important application area of technology already in the first phase interviews. The interviewees claimed that everything is documented. However, the meetings were considered necessary as well. There must be tacit knowledge not available from the documents, or knowledge that is easier to find compared to construction based on information from various systems (# 1). Formal communication included the pre‐defined progress meetings, audits, and kick‐offs, and informal communication included emails, telephone calls, and occasional meetings that were arranged for solving a specific, urgent problem.

So there were many types of meetings. Some projects were working without actual face‐to‐face meetings, in virtual meetings and phone conferences (# 10). Also the recurrence of the meetings could be irregular. Some projects were meeting when needed (# 10). Both virtual and face‐to‐face meetings were arranged, but more face‐to‐face interaction was required, because all problems could not be solved via email. If there was a problem that required discussion or hands‐on approach, there was an evident need for face‐to‐face meetings.

When people got used to the virtual meeting system, it was not seen as an obstacle for communication. However, real meeting in the beginning of the project was important according to majority of the respondents (# 1, 4, 5, 13). Most of the interviewees stated that it was easier to contact a person after meeting him or her in person. Typical thought was ‘calling a person is easier after face‐to‐face meetings’ (#1). First contact must be face‐to‐face, and after that similar meetings either during project milestone changes or regularly. ‘Maybe a meeting every other month is optima l – in any case between one month and half a year’ (#13). So it was seen extremely useful that the people met at least once in the beginning of the project: this improved the communication significantly. Communication



180

was seen successful if there were enough personal contacts between the partners. People wanted to be familiar with the person they were talking or sending e‐mails to.

So there were various types of meetings, like kick‐offs, reorganisation meetings, regular meetings, confidential meetings and ad‐hoc meetings. The kick‐offs were in the beginning of the project. Reorganising was happening very often in the focal company. There were special meetings for that purpose, to share the knowledge when tasks are transferred to new persons (# 6). There was a lack in ‘knowledge transfer’, if the persons were not meeting at all. It was easier to adjust the level of future communication, if the participants have made some kind of mental model from each other. The real‐time communication by phone or meeting was creating common understanding. However, the written documents in communication were released from the simultaneous sharing of time, and written documents are memorised by both discussants. So the gap was between documents and discussion.

During the project there were regular meetings. Those could also be educational happenings where subjects were taught by persons that were not necessarily participating to kick‐off meetings. ‘Knowledge sharing’ between collaborators was depending on the role of the partners. As claimed in one interview, very deep information was delivered to subcontractors in specifications. The focal company carried the contractual liability, while collaborators had special know‐how from a narrow field (# 12). One generic distinction could be made between partners and subcontractors. Partners are normally sharing the risks of the project, while subcontractors are just doing their work. Therefore partners are also sharing more information (# 4) – not necessarily so deep information. Technically, partners used groups of systems in intranet, while externals used only separated applications through extranet. The quotations on ‘deep information’ should be noted, as this is one reason to an alternative view of data‐information‐knowledge pyramid.

Ad‐hoc meetings were needed because some interviewees thought that discussing face to face is solving a problem quicker than by any other media (# 5). This is depending on the problem, but could also be an emotional question – shared issues are lighter to bear. Informal feedback and networks for personal information exchange were considered of utmost importance. A technical problem could serve as a reason to call or to visit a person, and many other important matters were discussed during the call or visit as well. The interviewees felt that ‘sharing knowledge’ was a value at the focal company, and in practice, some managers encouraged people to share knowledge.

One good practice was that managers dedicated time to regular, e.g. weekly confidential meetings. More general meetings or steering group meetings were for accepting the status, while more confidential meetings could dig deeper in the substance (# 10). So, there was no uniform practice in sharing knowledge. Instead, the ways varied a lot from unofficial coffee‐table discussions to more official newsletter‐like forms of distributing the lessons learnt. The comment of the changing culture was complaining on the lack of sharing knowledge in the extended enterprise. ‘…it’s hard to find out about the experiences of the others, because we don’t meet these people and nobody writes down the lessons they have learnt…’

In the extended enterprise level the knowledge sharing and communication were seen quite process‐oriented. Normally the focal company specified requirements for the system, that the collaborator implemented (# 12). Or more iteratively, the focal company was giving their input what they need, and the partner was specifying requirements, that were agreed together. This rough process could work even better than internal, constantly iterative process, with no explicit customer relationship (# 13). It is interesting because some of the partners used to be part of the focal company, so they have got common history. In every case, the focal company thought that the process couldn’t be controlled as well as when the collaborators were part of the focal company (# 12). It looks like the well specified demand‐supply relationship feels safer than constantly changing needs – even in an extended enterprise.

Everyday presence vs. virtual meeting via communication technologies

Interviewees claimed that the communication is better when people are working in the same premises (# 5). One common reason was that ‘it is easier to ask questions when meeting people’ (# 2). This was not the whole truth. There could be disturbing interruptions when working in the same



181

room, and many kind of communication is easier through systems (as written in the subsection Collaboration by systems.)

The practical tips were commonly pulping up in discussions, if the discussants were doing similar work (# 4). So the informal information search was happening in most cases. Anyhow, the searching and sharing was based on human relations, which could vary a lot. Still most of the interviewees considered this as an efficient way to work (# 9).

The initiative of a discussion is most often a problem that must be solved. The common factor between lousy search engines and discussions is that some unexpected advices might be popping up. That is an advantage, but it does not happen always, and could also result just a waste of time.

All the focal company workers and some of the collaborators were able to participate to meetings via communication technologies. These technologies were becoming more commonly available.

Meetings via communication technologies here mean simultaneous group connections through technical networks, and referred to as virtual meetings. The practice has changed. ‘Most of the internal meetings, which used to be face‐to‐face meetings, are nowadays virtual’ (# 14). Based on the interviews, in most cases there should be a rationale which kind of meeting will be called together. For example, ‘document auditing can be done in meetings, by videoconferencing, by meeting via net, by phone or by e‐mail – depending on the maturity of the document’ (# 2). The document could be shared with computers, and the participants of the meeting via net were discussing simultaneously by phone.

Virtual meetings were becoming more common. ‘Virtual meetings using audio connections and shared content are nowadays most common, while the videoconferencing is not giving the value compared with the complexity’ (# 14). The added value of seeing the speaker was not so relevant compared with sharing the presentations and other content.

Also with virtual meetings there used to be a restriction for extended organisation – during the interviews the collaborators could already participate in the meetings (# 9). There were, however, difference between face‐to‐face meeting and virtual meeting – for example, the whiteboards were not so fluently used, compared with flip‐over pads (# 12). ‘Drawing’ however, was not considered the same comparing pen and paper with computer tools (see e.g. # 6). Virtual meetings were for sharing ready‐made pictures, while face‐to‐face meetings were for drawing the first drafts – making the first shared, explicit models.

With virtual meetings as well as telephone calls, the different time zones were making one hindrance (# 8). For example, when meeting people from USA and China, two meetings were needed.

One technical restriction with virtual meeting was the quality of voice. If there were too many participants, the quality of voice could be poor (but this could easily be approved). An interesting notion is that for some listeners, it does per se matter who is talking. ‘You filter the message based on the identification of the speaker’ (# 12). Is this subjective preconception necessary to filter the communication for knowledge construction? Should the knowledge be evaluated more objectively?

Correspondence via e‐mail and lighter communication channels

Even the engineering information sharing was not working with documents only. There must be interactive discussions to generate as common understanding as possible. However, this did not always mean face‐to‐face meetings. E‐mail and other non‐formal communication channels were giving added value between meetings and formal documents (# 1).

E‐mails were not bridging all organisational barriers. There was a gap in e‐mail connections also. The organisational structure was enforcing these gaps. Most people did not bother their superiors with e‐mail easily. However, this was also depending on personality (# 1).

Some of the respondents thought that e‐mail is even a problem – or, more precisely, the way how to use e‐mail. ‘The e‐mails should be written weighing up (the utility value for the readers), to avoid two meters long messages, where thoughts are to‐ing and fro‐ing’ (# 5). Also the selection of receivers was a problematic issue: how to select the right correspondents – without missing anyone and without making the list too large (# 12).



182

According to some respondents, it caused an additional workload when information was transmitted by e‐mail (repetitiously). For example, the message subject lines were becoming misleading and some (external) correspondents were omitting the headers altogether (# 3, # 4).

Communication quality vs. quantity was a problem in e‐mail. It was a question of information overflow. Because the receivers’ capacity was overrun, the decision what he or she should discard, was critical. ‘How we can know what to ignore?’ (# 7).

The main advantage of e‐mail was the traceability – the correspondents had evidence what has been agreed and when (# 10). One disadvantage was lack of sharing. Also people normally just archive their individual e‐mails, and the information is mainly forgotten (# 12).

Extended organisation must sometimes use e‐mail for reporting purposes – if there was no access to more suitable tools. The templates for monthly reports could be e‐mailed meaningfully, but it required manual transfer to right archives (# 9).

One real problem was the mail attachments (see e.g. # 12). It overloaded the mail, giving scattering pieces of documents, and a flood of less relevant information. The mails could be complex even without attachments. They were forwarding discussion, which was hard to follow, in reverse order. Often the attachments were discarded in some phase of the mail chain.

All the e‐mail overload was due to two basic reasons: First, the organisational changes that were causing delays in updating information about the right receivers. Second, the wrong usage of e‐mail (discussions, attachments, misleading subject fields or headers) was causing futile information sharing.

There was a common practice to send e‐mails to distribution lists. The problem with e‐

mail overload was caused also by lists of respondents: the sender wants to make sure that the message will reach the right recipients even he or she is not sure who the right receiver might be (# 2). In most cases the distribution lists were generating overlapping information delivery (# 4). However, in some cases some readers could achieve information that will not be known otherwise.

One result is that e‐mails should be divided to pull and push mails. A general copy should be released in a newsgroup, while the actual mail is sent only to the relevant persons. This, of course, could be done by sending only one e‐mail. A technical question was how to select the external readers.

The e‐mail publishing for a larger audience was the channel to deliver information that the recipient did not know how to ask. Based on that first publishing, the raised questions could have been answered in a more personal conversation.

Also chatting, phone services and usenet (news) were commonly used in our technologically‐oriented focal company. There could be any sort of culture in usenet kind of services. One common opinion was that newsgroups do not work if they are used as Q & A bulletin boards (# 4). In some cases, however, they could be used very actively. This might depend on the nature of the discussion. Interaction keeps up discussion.

Lighter kind of e‐mail was chatting service, where the discussants could be changing messages on‐line. However, they could have breaks within a session and so they are somewhere between e‐mail correspondence and conversation. Also this interactive communication interrupted work easily. Preferences depend on working methods.

One widely used communication method was short message service, which was used widely in the focal company. Not yet so official methods with collaborators are other interactive mobile device based chatting services.

One reason for telephone (voice) calls and the new interactive instant messaging services was the ‘full‐duplex’ nature of conversation, compared with ‘half‐duplex’ e‐mail correspondence. ‘There is always a delay with e‐mail’ (# 3). According to respondents, people could also ignore e‐mails, and must sometimes be contacted by phone. It was a little bit illogical that people, who were behind the gap to be contacted by e‐mail, must be contacted even more personally. As a benefit, e‐mails leave kind of traces which phone does not leave. Should the storing of chats be asked from the respondent, is an open question, but the risk in contemporary culture is that even the permission to phone recordings will not be asked for in practice.



183

3.5.4 Conclusion on most important observations

The collaboration with an information system is a different behaviour than communication through an ICT system, as the information species are presentation in the first one and communication in the latter one. This basic difference has been mostly ignored in the KM literature so far.

There is also a connection between authors and readers through information presentation. This communication requires the knowformation layer in the new model. Technically, knowledge objects in case III require XML. The nature of trying to make informal knowledge explicit by communication media has been perceived in many occurrences. Feldman (1987, p. 83 – 101) suggests that a totally new type of communication occurs in large organisations that use electronic mail frequently. On groupware tools, such as e‐mail and Lotus Notes, Conklin (1999) notified that they generally fail to create a coherent organisational memory. Knowformation include presentation in implicitly communicated short term memory of an extended enterprise.

It looked like in a web environment there was a problem of information overload, trust, reliability and validity, while in groupware environment the problem was dualistic – either no information or unique (right) information with no alternatives. So the conclusion is to use groupware application for intranet publishing.

The searching and finding are related to presentation, while ‘knowledge sharing’ is communication. The groupware documents were not always indexed with search engines. In those cases it was easier to ask from someone who knows the right document bases. Also there might be the same documents (different versions) in various document bases (# 12). One result is that the human memory is working so that the context of data is memorised as well as the actual data. For first times, when searching for a data, you memorise how are you searching, and the actual data is another issue. Of course that is depending on the importance of the data and the expertise of the user. Especially a newcomer in a project needs to ask from colleagues. When the right sources are known, the information system search capabilities are utilised. The document bases are easily classified to master data and replicas.

It looks like finding a ‘right’ person does not actually mean to find the right information, or – not necessarily even the right person. The reason why finding the person feels safe, is that he or she could give a feeling of shared responsibility for the consequences of the action. The searcher is liberated from the burden of undistributed determination, even though the formal accountability is organisationally on the decision‐maker alone.

Information and data can be normalised to third normal form in organisational as well as system level. However, knowledge is not even recommended to be normalised, because the worldview, interpretation and construction are depending on the people. As a result, one slight difference between information and knowledge is just the advantage or disadvantage of normalisation.

There are many reasons to see the data‐information‐knowledge triangle (presented in Part I Chapter 1.5, What are data?) differently. Deep information presented here is one reason for alternative view of that triangle. The differences between internal and extended enterprise ‘knowledge sharing’ is so different that it drives towards separating communication and process. This – together with other results – will be presented in the Part III.

CONTENTS

Part III – Modelling the results

Systems

1 FIN

content based onthe othe

Ca1.1

case studgeneral individua

and relatinformatsystems.present

began, wnoted sy

knowledfree‐formknowledOracle, DKnowledlike DAMrichness

between inf

DINGS FROM

Is the is realised, an the case ster was based

ase I – Mana

This sedy, analysis oresults from als in the foc

The fited productstion and kno. Therefore tin the backg

After twith knowledystems were

All the

Phasedge mappingm text. Throudge items canDB2 or MS‐Sdge mappingML or RDF, th of the mode

formation an

M THE PREVA

KM paradigmand how the udies. All the on human k

agement of d

ection mainlof knowledgthe researchcal company,rst phase of s. The clear‐cowledge manthe ‘universaground also inthe DB compdge managem:

• • • • • • • • • • • • • • • • • • •

e systems no

e 1 results co, in their mough file‐servn be archivedQL Server rag tools confohereby providelling langua

nd knowledg

ALENT USAG

m living yet asystems (ince case studieknowledge‐r

data from bo

ly presents ae based on dh field, as we, representincase I begancut result wanagement syal’ RDBMS syn case II andparison, a wiment applica

RDBMS, BuDesktop DaODBMS SQL DeveloOLAP ToolsData MiningData WarehData MartsInformationKnowledge CollaboratioKnowledge OrganisatioOn‐line TraDocument (Project & RReportingSearch engiNon‐catego

oted are liste

ould be summost primitive ver or, in mord and reusedther than prrm to a standing universge used, kno

ge

185

GE OF SYSTE

another pericluding commes consisted elated issues

oth the indiv

a summary ofdata manageell as case stng current inn with a wideas that ultimstem, and thystems were Oracle in caide analysis oations as the

siness Intelliatabases

pment Tools g house

n Retrieval Discovery on Mapping &

onal Memoryining System(content) mesources

ines orised KM re

ed in the App

marised baseform, assist re advanced d. Better tooroprietary redardised moal and unamowledge map

Part III – M

MS

iod of prospemunication)of two partss, so both sid

vidual and gr

f the results ement systemudy results fdustry. e market anaately, data mherefore the evaluated. Oase III as wellof the more main focus

igence

s

Ontology y m anagement

lated applica

pendix 2: Sys

ed on the sysin the captucases, datab

ols cooperatelational dataodelling langumbiguous accpping tools c

Modelling the

erity, now wwill follow us: one dealt wdes were ana

roup viewpo

from a reseams and applifrom the spe

alysis of all knmanagementcritical core Oracle and Dl. advanced daarea. The ca

ations

tems investi

stem categorre and visuabase technole with leadinabase managuage for expessibility. Decan codify re

e results

hen the valuup, could be awith the systalysed.

oint

arch project cations. It coecified views

nowledge mt is behind evof all levels

DB2 products

ata managemtegories of t

gated.

ries. KM toollisation of unogies, captug databases gement systeporting and imepending on lationships,

ue of analysed tems while

on the first overs of

anagement very of the s were

ment tools the initially

ls, like nstructured red such as ems. mporting, the rules and



186

constraints (selecting from an arbitrary or a fixed set) on the objects they model. Some tools can store explicit information that accompanies a concept or object.

Organisational memory tools are rather similar to knowledge mapping tools in that they often capture knowledge in a visual format. Their distinguishing feature is that they focus on unstructured tacit knowledge. Here tacit knowledge is interpreted on a common, organisational layer. This is due to organisational changes, after which stored documents remain in the systems but the persons whom they concerned are no longer there. Meta‐information on systems and content is relative. The implicit layer between information and knowledge has been shown to be of crucial importance as a type of organisational consciousness.

Owing to opinions, obliviousness, or other unknown reasons, the focal company considered that meta‐tools such as organisational memory systems did not offer enough value compared with the additional work they would cause; the actual systems were therefore replaced with other systems already in use, with the accurate content. The content side of the memory metaphor is more important than the systems dedicated to it on an extra layer. It is advisable to support files that store organisational memory information and models – just as any models – if the organisation, or more preferably extended enterprise, shares such tools. However, it should be emphasised that planning of systems demands learning and analysing of the memory content: systems require clarification of the implicit knowledge towards explicit information and data.

Information retrieval systems got to the heart of the knowledge content stored in a database or a file server by providing intelligent multi‐criteria search functionality for every word or object in the knowledge resource, based on elaborate indexing structures that supported a wide variety of storage formats. The file servers were not intended to host any applications or application connections to backend servers, thereby limiting security risks. The servers were dedicated to storing or processing, or to search as a meta‐process. The search functionality appeared to be crucial in all case studies.

For the IT professionals, the simplicity of the search was not the criteria, but the search with Boolean operators, and other methods combing the search parameters. The server side was also investigated. Since the mid‐nineties, optimised search algorithms and significantly faster hardware technologies have offered a very conducive development environment in which data mining tools can flourish. Since then data mining and knowledge discovery have been considered critical, as companies realised that it was humanly impossible to sift through millions of records of transactions or events in order to draw some conclusion – or at least establish a pattern or a trend, and subsequently make a strategic business decisions. Ad‐hoc reports and other data based knowledge are important for management. Extended enterprise executives are constantly collaborating in steering groups and other communities utilising this kind of information.

The phenomenal expansion and acceptance of the intranets, extranets, as well as the Internet and mobile networking have fuelled the development of collaboration tools that allow employees to work together across theoretically limitless distances. The disadvantage of different time zones could be minimised with delayed communication in organisational short term memory. Activities like work meetings and their capturing, white boarding, and document sharing are now feasible without time‐consuming travelling or paper mail. This independence from the constraints of time and space has been even more emphasised with the practicality of mobile devices. Mobile collaboration and multiple kinds of communications were used in the background of all the research cases.

In a similar fashion as in collaboration tools, the Internet's ubiquity and the user‐friendliness of the web‐browsers' interface with different devices was a boon to on‐line training tool developers, who sprang to capitalise on the opportunity to save people time and money as they tried to earn a specialised professional certification or a degree. Though it is true that no network or browser can substitute totally for an on‐site instructor, recent experience and the sprouting of on‐line universities show that when expectations are realistic, on‐line training tools have a major role to play in general and professional education.

There was one common denominator for any kind of platform: middleware. The ever‐increasing consumer pressure for richer and better functionality of enhanced web services, along with the variety of programming languages and database interfaces have spawned a perpetual vendor race



187

to develop high‐performance, feature‐rich middleware technologies that will, hopefully, capture the attention of the market and become the dominant standard. This study was concerned with the organisational viewpoint, which sees service orientation or business processes as more important than technical implementation. All the case studies showed that knowledge work relies on few combined applications in which data from several sources is integrated and displayed on a single browser interface in any device.

Case I phase 2 interviews showed that a company‐wide vision of data management systems is beginning to be formulated, even though different units used different systems. Applications and data are increasingly important for all units. Since the amount of data continues to grow, the processes, workflow, working methods, systems, programs and manuals must be revised constantly. Because the processes are more important than the technical applications, how these organisational processes function should be analysed first (being aware of the information itself) followed by the systems planning to support those processes.

The quality of data is the basis for everything, because results depend on the accuracy, completeness, consistency added value, interpretability, accessibility and other factors. If any data are incorrect, the reasons should be traced and errors should not be corrected to the reports or other outputs for decision‐makers. Knowledge based on the systems is the final reason for all data. If decisions are to be based on high quality data, there is a need for criteria on how to evaluate the quality.

The selection of data from operative systems must be well founded. The designers of the data and knowledge management systems must know the criteria for decisions – even though criteria may change. Business units may change their views on the data, but that should not be a reason to store everything in data warehouses or data marts. At least static data and meta‐data should be normalised on the organisational level, and dynamic data could be analysed on the basis of samples or summaries.

Normalisation is not a new idea. Rao and Goldman‐Segall (1995) discussed the idea of saving chunks of video‐data, with the user retrieving relevant chunks to interpret past events. If information about the same topics is stored in multiple formats – say in a database and multimedia format, users will need tools to re‐integrate and synchronise the memory. The point is that communication does not normalise anything, while process‐based data and documents can be modularised.

There are numerous applications for data warehousing, data marts and hence knowledge management processes. Product creation process applications include document management, metrics calculation, project management, reporting, portals and product data management applications in this context. Delivery and support processes are using more or less the same selection of applications. Applications run on different platforms. Some examples from the market are Informatica, Progress, Artemis, SAP/R3, Ingres, Documentum, Optegra, ClearCase, eMatrix, Hyperion Essbase, Business Objects, Cognos and Crystal products, but there could be plenty of Lotus Notes, Hermes, Oracle and SAS‐based tailored applications. It is always possible to find alternative applications with minor advantages, but two specific fields, not yet in optimal use, are data mining and generic knowledge management applications; only special interest groups are aware of their potential.

Product creation process data come in many formats, but resource and project management data in systems are common mainly on the business unit level, whereas finance and control budgeting and reporting data are normally used by corporation. A common conception is that production control systems data are in better shape than data in the rest of the systems, possibly because production systems have been developed quite independently over the past few years.

The quality of data is difficult to improve if the business representatives and support process do not work together. Quality managers could participate in improving the actual process data, even though they were not involved in the processes. Data models of the systems were not well known. Also business record archiving was in practice a little bit distant issue, although it was planned well. Both concepts are needed: archiving of memory contents in a long term memory, and the record as a system category between data and non‐structured information.

Knowledge management templates for different purposes exist, but they are not aligned or harmonised. Practical formalised knowledge relates to business programs: roadmaps, feasibility studies, plans, feature descriptions, manuals, and customer documentation. Information is entered

Systems

mainly bdate con

data stogrowing

to actua

exact. Mby the gopages, w

competeshould bdifferentmethodsuse it.

in a modlicences data min

some buunits do projects these sy

lack of awide conunits do need forresult ofapplicatibenefit f

Ca1.2

managemconverticlient aptwo stantheir feawillingne

administsystems,administ(personnformalitythe form

between inf

by assistants ntrol togethe

Futurerages will sp. Metrics obj

Businel reporting, a

In genMeta‐data molden mean:where nothin

Repore with each obe easy to cht rights and rs and not rec

Costs dern corporato other appning based o

Some usiness solutnot work tothat humanstems.

The pe corporationncept is not ynot aim to cr such an aimf this case stuions – due tofrom system

ase II – Mana

Documment tools. Tng, routing, pplications. Tndardisation ature set is, tess to put th

Case Itrators, and t, as well as atrative (HR), nel developmy of process mality of the

formation an

and is of mixer with manye plans inclupeed up gettijects will be cess planningas a part of cneral, there aanagement : data worth ng is archivedrting is needeother, whichange. Differeroles should ceptive to ne

during the aation should plications. Mn certain apconcepts arions that areogether from knowledge,

ersons involvn‐wide conceyet crystalliscreate knowlm. This issue udy is that fuo organisatiodevelopmen

agement of i

ment managThey promissearching anTo address thactivities hathe efficacy oe system intI included twthe second oa comparisonprojects (proment) and Invsubject or coprocess did n

nd knowledg

xed quality. y storing platde data inteing informatcreated inste will be empcontinuous pare many opis conflictingarchiving is d). ed more ofte is a disadvaent user intebe defined aew solutions

application dbe planned c

Metrics reportplications are not yet wee not yet awa the very beg decisions an

ved in the deept as a probsed as an expledge level cis discussed urther develoonal knowlednt and know

information

ement systee to offer thnd distributinhe sheer numave been initof documento use, integrwo phases: thone was a stan of five diffeoject managvention (innontent did nnot correlate

ge

188

There shoultforms. gration and ion all the tiead of reporphasised in thlanning, whien issues to g in differentnot recognis

en. Some datntage for buerfaces are nand manage. It is futile to

evelopmentcarefully to ating, meta‐de difficult toell known. Daare of its potginning of a nd actions de

evelopment lem. One reaplicit need. Toncepts for ton an extenopment of a dge – is needwledge creatio

in an organis

ems comprisee knowledgeng documentmber of docuiated. As comt managemerating it into he first one catistical comerent processement, withovations andot define thee with the m

Part III – M

d be greater

at least procme. All kindts. he future, asich follows abe solved. Ot systems. Arsed, so every

ta managemusiness units eeded for ded. However, o change a s

and use phaavoid risks inata managem trace in termata mining, ftential. Inforproject. It sherived from t

of business ason might bhe data levethe enterprisded enterprcorporate‐w

ded. A case ston.

sational unit

e a very largee worker an ts of variousument storagmplete as thent systems htheir everydconsisted of

mparison of twses, in order different prd patents) pre formality oaturity of th

Modelling the

r focus on do

cess wide plas of metrics

the need gr product's lif

Ownership ofrchiving of hiything is arch

ment and repusing both. Sevelopers anpeople are tystem, if the

ase often expncurred, e.g. ment as welms of cost alfor example, mation manhould alwaysthe data is th

unit level sysbe that the lael applicationse, althoughise level in cwide concepttudy like this

t

e portion of integrated s formats, froge systems aeir functionahinges on theday routine.interviews owo documenr of formalityroject substarocesses. Theof the procese content.

e results

ocument vers

atforms. Opeand reportin

rows for a sefecycle. f data shouldistorical datahived (cf. cas

orting produSystem archnd end users,tied to their e users do no

pand. The licwhen changl as data warlocation. could be proagement ans be kept in mhe final aim

stems do notack of the con owners of b managers dase III and tht for data mas gives a firm

knowledge uite of tools om and to a lnd client appality is and ase knowledge

of users, manntation manay: legal (i.e. juances), learnie main resultss itself, and

sioning and

erational ng are

amless view

d be more a is not donese II intranet

ucts itecture , and old ot want to

cense policy ging from rehouse and

oductive for d business mind in all of all of

t see the rporate‐business do see the he main anagement m basis to

for storing,large set of plications s rich as workers’

nagers and agement uridical), ing t was that vice versa:

e



189

1.2.1 Suggestions and results from case II, phase 1

For knowledge sharing or communication purposes, the case organisation needed to publish project results with a meaningful tool. It should be noted that there are many verbs interpreted as communication, such as delivery in case II. Together with formal delivery, more informal lessons learnt from projects should be published. Projects themselves needed a framework: instructions and templates for documentation. Both line organisation management and projects could utilise documentation management systems. Neither explicit knowledge nor documents vanish when a project ends. The knowledge (i.e. content here) written down on documents should be independent from the tools. However, a single tool for all kinds of documents was required. The case II respondent mainly stated that the tool should be the same for the whole case organisation, preferably for the whole focal company. But the enterprise content management is for storing the content that should be accessible with any client application.

Easy access to documents was a necessity for all knowledge workers. Organisational structure, groups and projects changed continually. Administration of documents must be easily done preferably by the users themselves. There must be a unique 'original' master version for each document. The document Management system needs to be flexible to manage deliverables and active documents.

Even though communication was required, electronic mail is not the place to store documents. Neither Notes nor intranet pages should be the master tools for documentation management. Projects must use a unified document management system. In communication, most bulky messages are references that can be replaced – only links to the documents should be delivered when messaging. File servers should be utilised with or without document management systems for their meaningful purposes.

A decision must be made on the kinds of files to be are managed with a document management system. Meaningful templates should be used more. Intranet searches must be improved: Also here we encountered the problem that some of the search engines were indexing wrong, and not offering rational Boolean searches. Documents need keywords and abstracts, and they must be logical. The keywords are for limited kinds of documents, while the abstracts are more important for new kinds of search methodologies (see Subsection 2.2.11). Synonyms and homonyms should be defined in abstracts or keyword lists, but not used in the texts of structured technical documents, especially when writing about mature data based concepts. Standardisation or recommendation might help.

Workflow was strict only with a few documentation processes in this case organisation. Document lifecycle and workflow are not disadvantages when used practically. Many regular documents can be modularised and some of them were already dynamic XML. These include project plans, handbooks and development discussion‐based work target and evaluation documents. However, compound documents, such as instructions, technical manuals and data based reports are more akin to results of knowledge management, not the knowledge work itself.

Too many systems for similar purposes were in use. The systems must be selected, but committing to one single document or content management system contains the risks that on one hand it could be rooted too deep into an organisation and on the other hand one system may be impossible for an extended enterprise consisting of many organisations. Even though centralised master documents management reduces the amount of redundant copies, version control is needed for different branches of document compilation and history. The problem with Intranet editors' tools might be the 'overdose' of futile pages, similar with search difficulties. The document management system must be compatible with e‐mail and office applications (independent on the operating system), and ready for XML.

The less conversion needed the better. Intranet pages also need version control and history. Version management was important, especially for ordinary end‐users, project managers and system administrators. The document management system must work together with web technologies including e‐mail here as well. External sources (e.g. articles via information services) could be integrated or tagged into other content. Personal, project and organisational unit (laboratory) ‐specific documents form one distinction of levels, the project level being most important. Co‐operation inside one country was working, but sites abroad varied.



190

Active contracts or other juridical documents should be neither only on paper, nor published. Some interviewees claimed that public database identifying competencies on project and personal levels will help to search for the right persons. However, the competition between workers remains. Organisational culture should not encourage futile documentation (especially e‐mails and intranet pages). Correspondence with external partners depends on the relationship. Inside an organisation, the document management system should become used only when it is agreed on and justified, and not dictated by information management. This concerns storage in the servers and not workstations of the users in general.

1.2.2 Suggestions and results from case II, phase 2

The requirements of Case Organisation (CO) are listed below in the order of importance concerning generic document management requirements. The features studied are suitable for any basic content management process.

Table 20. Importance of system features

Ref. Importance Requirement CO 16 3.1m Very important Web servers and mobile clients 87 % 7.1all Very important Workflow 80 % 2.1all Very important Document management system in general 79 % 7.2all Very important Lifecycle 77 % 4.1u Very important Search from document attributes 75 % 6.1m Important Different security levels 74 % 2.2all Important Integrated document management system 71 % 6.1u Important Notifications 68 % 4.1a Important Compound documents search ability 64 % 4.2.4u Important Representation about document relations 63 % 8u Important Customised document management applications 62.5 % 8a Important Standards 62 % 2.3all Important Compound documents 61 % 5.1a Important Manual version numbers 60%

We can roughly divide those requirements into process‐oriented and communication‐oriented. For example, mobile (thin) clients, which are especially important for the case organisation, belong more to the communication category, while workflow, document management system in general and lifecycle are definitely process‐based requirements. Process vs. communication will be a basic distinction of the model, as this division appeared in all repetitive studies. On the other hand, web servers, document attributes (and other meta‐data), compound (modular) documents, document relations, standards, customisation, and integration with any editor or the like, are saying that content should be independent from the systems.

As with the information species, the formality of the process gives the location on the process—communication scale. The above table does not indicate how important a process specific feature is; for example, that workflow is very important to invention process, while workflow is not very important for projects. As mentioned in Part II subsection Case II, phase 2 exceptions, such relations were analysed, but not listed here.

16 There was total of 35 respondents (N) but a varying number from different subgroups (users, administrators, managers). Therefore the N is not marked in the tables of graphs.



191

1.2.3 Case II future

In Table 21 the averages have been collected to show some comparable scores on general attitude. They are continuation to the results from the case II and their relevancy is better understood as case II recursive analysis. The table summarises how important the features are for different processes and user roles. It also shows the differences between groups. A comparison of managers of Innovative process and Learning process, for example, is meaningful, while it would not be meaningful to compare Innovative process managers and administrators in this context.

Table 21. Importance of system features in summary, compared between processes and roles

Process Managers Administrators Users Process average

Innovative 62.86 ‐10.02 46.15 ‐24.27 84.69 +12.11 68.52 ‐2.74

Learning 76.62 +3.74 76.37 +5.95 87.76 +15.18 80.19 +8.93

Projects 67.86 ‐5.02 71.43 +1.01 63.95 ‐8.63 68.20 ‐3.06

Admin. 64.94 ‐7.94 89.01 +18.59 79.59 +7.01 74.66 +3.4

Legal 85.71 +12.83 68.13 ‐2.29 52.04 ‐20.54 64.54 ‐6.72

AVERAGE 72.88 ↵ 70.42 ↵ 72.58 ↵ 71.26 ↵

This table has been placed in this case II future subsection, because it is significant for

selection of future strategy. We can see from the table that Innovative process managers’ and especially administrators’ attitudes were below average. Projects were more or less in the middles but learning toward the negative side. Legal process managers were very positive, and the negative score for the whole process average is due to users. Learning and administrative processes were very positive on average, but their managers were pessimistically realistic.

When making a conclusion on correlation between the importance of document management and the formality of process, we should remark an independent entity: content. For example, Innovative process respondents did not consider systems in general important. They were against compound systems and even integrated systems. A reason was that they already had got a process specific system that was working properly (but only inside one process.) The main reason was that their content consisted mostly of formalised patent applications. The innovative new knowledge was captured into mature process and managed as ‘frozen’ data.

For the Learning process the less important features (universal platform, Unix clients) were not critical for the process itself. All the fundamental features needed could be utilised as well. Therefore the Learning process told the importance of knowledge management system features as they were.

Projects formulated so wide a category, that there would have been many projects for using each platform. The main negative differences from the average were different security levels, workflow, manual version numbers, object as attribute, object relations, customized applications, and somewhat the lifecycle. We must remember that there were respondents from different kinds of projects, so their comparison was based on further analysis already in this requirements phase.

The irrelevant features of the Administrative process (cross‐repository searching and filtering, and somewhat manual version numbers and universal platform) were also out of the process, so the relevant features are based on real needs. Also, they dealt with quite a critical data, and therefore they were not involved in the system testing.

The lowest points from Legal process (Unix clients, standards, and checked out document view option) were all second class issues. The main point with their process was that all the features

Systems

that gavhighly fo

implemeAdministcrucial tawhen m

results othe qualin the inpractice Howevereliabilit

communformal thwhom thmanagemknowled

go towathe more

Ca1.3

creationinformatessentiaintroducinformatparadigmto make

well. Theand fast‐organisainformatwithin thsupportiand to su

technoloinformatorganisabetweencompanthe prod

between inf

e most profiormal and re

In thisentation (Dotrative procearget of caseoving from d

Whenon internet‐pity of the coformation spas importanr, the conteny of quite sta

One cnication) is vhe subject ishe informatioment systemdge.

Univerds organisae important

ase III – Man

Inform and managetion. Howevel in collaboraced theories tion servicesm of researchthe memory

Duringe keywords w‐changing coational memotion and knohe project, toing knowledgupport the r

The cuogical practiction systemsational point n organisatioies, if we decducts and ser

formation an

t out of the squired this ks set of reseacumentum) esses. Howeve III. The usagdocument mn the documepages researcntent, usabilpecies) and unt as the contnt requires satic data is mrucial notionarying indeps, the formalion is meant.ms requireme

rsal platformtional or evethey’ll beco

naging the kn

mation proviement. The ser, the sociaation. Even tabout the sos for engineeh as well, evy managemeg the researcwere global pompetenciesory. The goaowledge creao other projege manageme‐using and urrent interpce was preses. Some cultuof view. A so

ons: traditioncouple knowrvices that an

nd knowledg

systems werkind of contearch cases, pwas the nexver, these wge of the exianagement tents are pubch (Pembertlity, the ‘fresusage speed.tent specificystem featumuch more cn is that the fpendently froity of the pu As mentionents should b

ms correlate en extended me, as will b

nowledge of

ded by systesharing of inl context is cthough sociaocial construering as an exen though a ent model stch, knowledgproduct deve, strategic pal of KM in enated in produects and to cment were exsharing of expretation of cented. Literatural aspects oocial viewponal methods wledge from n organisatio

ge

192

re consideredent managemprocess analyt step for a sere out of thsting systemtowards the blished on theon 2004). Thshness’ of th. In www con reasons – wres, not vice crucial than tformality of om the naturblished conted, the crucbe set from t

with thin clieorganisation

be seen also i

f an extende

ems and servdividually cocrucial for orgal is possible ction of knoxample. Consresult is tharategically wge managemelopment orartnership nengineering wuct developmcollaboratorsxpected to stxisting knowcommunity, ture based oof knowledgoint is more iare the leastinformation.on creates re

Part III – M

d important.ment accordiysis for commselected projhe ‘knowledgms (including web contene web, we cohe trivial rease content (qntext the syswhen there isversa. For e

the novelty tthe process re of the protent dependsial factor is cthe viewpoin

ents and connal communin the next c

ed enterprise

vices form a constructed knganisational without socwledge in anstructivism tt informatio

working. ment becamerganisations,etworks, virt

was to facilitament projects. In this studtore and deliwledge. system and

on a sociologe sharing wemportant wht able to man. Engineeringeflect organis

Modelling the

. Therefore, tng to procesmon documeject(s), togetge sharing’ qNotes) was

nt. ould refer tosons why useuite the samstem specifics not any othxample, in tho the reader(on the scalecess. It means on the targcontent. In thnt of content

ntent indepeication fromase.

e

common basnowledge creknowledge.

cial constructn ultimate coturned out ton must be ca

e more impo requirementual teams, aate the retens by passing dy, the informiver product‐

organisationical context ere observedhen dealing wnage the knog gave us a gsational kno

e results

the legal pross. ent managemther with Leguestion, whiconstantly o

o Steven Pemers return tome as noveltyc requiremenher specific mhe legal procr. e process — ns that no mget group audhe future, tht – or ultimat

ndency – the sheltered p

sis for knowleates organi This is partitivism, this caontext: usingo be a philosategorised in

rtant in engints for diversand improvetion and reuit on to engmation syste‐specific info

n concepts inwas adaptedd from a critiwith collaboowledge of sgood case stuwledge.

ocess was

ment gal and ich was one observed

mberton’s o a page are y specified nts are in motif. cess, the

matter how dience, to e documenttely

e more we’llrocesses,

edge isational cularly ase g ophical nto species

neering as se, multiple d use of ineers ems ormation

n d for the cal, oration separate udy because

t



193

1.3.1 Results of case III part 1 – extended organisational cooperation

Theories from sociological context are well suited for information services and even for engineering. The problem so far has been that engineering information has been seen too much as systematic, process‐oriented data, but the knowledge constructed on various information species is more important. Organisations must be flexible in responding to volatile markets. A dynamic organisation needs different management of knowledge workers than historic large vertical organisations where mainly physical resources were managed. Case III showed how organisational culture shaped the way in which knowledge of diverse repertoires or routines were integrated, while organisational history differentiated the way new knowledge was created. The collaboration of individual experts created new knowledge, which could only partially be transferred to information services. This subset of organisational knowledge transferred to information services was more effortlessly available also for cross‐organisational collaboration. Thus the constructed knowledge includes other information species, even though epistemic knowledge is one subset of information presentation.

Collaborative organisations should be seen more like communities of practice, as technological knowledge is never pure. Information and communications technology as the only mode in virtual working can be questioned from the viewpoint of sharing knowledge. (This is another context where epistemic knowledge should be distinguished from the word knowledge as used in everyday language.) In any case, collaboration and thus communication is necessary between dispersed teams within an organisation and between corporations. Working apart and possibly asynchronously underlines the significance of availability of communication and collaboration technology, the richness of the media, and easy access to shared information. The content might not be pure, but the technology itself could be.

The sociological context gave a well‐formulated interpretation of collaboration between different organisations where organisational culture played an important role. (This was based on the observations of case III Finnish, French, German, Greek, Norwegian and other organisations, not as a research objective itself). No matter what the geographical background is, some basic rules of knowledge management should be part of the organisational culture. It is the only way to confirm credibility between employees, and to support the active search philosophy of knowledge workers. Nowadays knowledge management results should absorb into social culture among organisations, instead of being artificially taught. Knowledge of the ICT system or any document does not mean that it will be understood similarly by the members of the organisation or any other actor. Another issue is tacit knowledge, which can only be transferred by moving people. As the complexity of the knowledge to be transferred increases, the ability of documentation alone to convey that knowledge decreases. The correct mix of documentation and interaction is needed for human understanding. This is especially valid in external, international collaboration.

Reconstructing tacit knowledge in a new context is difficult – if not impossible. The borderline between tacit and explicit can be viewed as always shifting (Hedlund & Nonaka 1993; Polanyi 1962). Tacit knowledge can only be shared through shared experience. Apprenticeship, for instance, is a common way of adopting and sharing tacit knowledge. However, the concept of tacit knowledge does not cover only learnt practical skills, but also intuitions, assumptions and beliefs not expressed publicly. This does not conform with the original information species. Therefore, as the case study interviews demonstrated, it is an advantage to distinguish not only explicit but also implicit knowledge from tacit knowledge, because no distinction is normally made. If knowledge is not classified as explicit information or tacit knowledge, it is something in between. The concepts around implicit knowledge will be further refined.

Nowadays organisations collaborate with different partners in a value chain, consisting of standardisation associates, subcontractors and other component or service providers, and customers. All suitable methods of management and research should be utilised for this kind of social culture of extended organisational knowledge. Mutual obligation is needed between workers but also between employers and employees. In collaboration, mutual obligation of partial information and knowledge sharing should extend between organisations. One challenge is that knowledge sharing should lead to positive results – retraining and career achievements within the organisation: rewards should not be



194

artificial. Knowledge sharing with collaborators should be seen as extended organisational knowledge, and treated in analogy with the management of internal knowledge. From a constructivist and communication viewpoint on knowledge and information, ‘sharing’ is a smaller risk than open data sources.

It is challenging to study technological change through a non‐deterministic model. This research suggests that technology as knowledge, system or function (inside a process) is a vivid distinction in organisational knowledge. The process could also function between non‐formal users, such as a community. In a community of practitioners individual membership is not based on disciplinary commitment. There are normally multiple traditions of practice within multiple communities in the technological hierarchy. Communities of practice for communicative collaboration are close to ‘knowledge sharing’, but experts who have solved problem do not participate in a community without another reason not connected to solving the problem. Sometimes knowledge can not be captured simply because people do not understand how they become successful. These are real problems in extended collaboration, where informal communication is rare.

Technology requires integration with the community and organisation. Invention and entrepreneurship have a wider perspective on the sociotechnical and organisational system. Integration of multiple expertises implies a complex organisation. If an information system reflects the organisation, it should be emphasised that people are learning, while computer systems lag behind. People’s networks outside the corporation are not so often represented in internal information systems. Thus collaboration is the crucial component of dynamic organisation. Future information systems for cross‐organisational or inter‐organisational knowledge are rooted in the history of people and thoughtful communication belongs on the top of process based information. Systems themselves do not learn, while the content and process could reflect the learning.

Meta‐data on skills

If knowledge is interpreted more than justified true belief, as in current vernacular, actual knowledge can not be managed, but only supported with communication technology. However, meta‐knowledge can be somewhat managed. For the model, we should define knowledge as a subset of information that could be managed, while skills are something other than knowledge.

Meta‐data on each person’s skills does not work in a big, co‐operative organisation, if data is updated by human resources or another kind of official department. A list where each person updates his or her personal data is giving – if not a comparative – at least a personalised image of each worker. The information system could be used to give personal background meta‐data on knowledge, which is normally delivered during introductions in face‐to‐face meetings. Of course this is personally edited, lacks tacit knowledge with its impressions, and does not have any other qualifications except publicity. Introverted real experts could be underrated in comparison with extroverted colleagues.

When searching for the right information, ‘detective work’ would be minimised, if there were a person responsible for the defined documentation entity (subject of expertise or skill). This leads us again from people management towards systems through communication.

Conclusions from case III, phase 1 – users’ viewpoint

The inter‐organisational network was designed to support a flexible product creation process, which was divided between the focal and partner companies so that the main responsibility for innovation remained in the focal company. The methods of collaboration were clearly defined, which was a critical factor for the success of work distribution. The IT supported tools of knowledge work, however, had to be chosen carefully, with a good balance between documentation and communication.

The work processes, ways of working and even the extended organisational cultures in this case study were closely related, creating a sense of a virtual organisation and common way of doing things. The factors that enabled and facilitated virtual collaboration included direct personal contacts and meetings between persons in the early stages of the projects, which requiring extensive virtual



195

communication between distant sites. The results also brought to light new dilemmas that emerged in networked product creation. People pondered on where and when to work when colleagues and collaborators were dispersed in time, location, and organisational structures. Great flexibility on the part of the partners was needed when the focal company’s processes, products, and projects changed during product creation. These flexibility issues in the organisational structures were reflected in the tools of knowledge work. In addition, designing flexible and modifiable system architectures became increasingly important.

1.3.2 Results from case III, phase 2 – KM systems

The phase 2 outcome is in part related to the case II document management systems. A key advantage was that the tailored case III system (Web‐Enabled Information Services for Engineering, WISE) offered a proactive retrieval of documents that related to what the engineers were working on, via knowledge applets.

But there were many systems behind the meta‐level WISE application. Briefly presented are the findings concerning information systems in the networked knowledge work, and the results of a comparison of two groups of interviewees: the employees of the focal company and the partner companies of an extended enterprise. The common factor between these two groups was the attitude to information systems in knowledge‐intensive cooperative work, and little if any difference was found.

Extended organisational networks between collaborators required new kind of management. A holistic view of processes was necessary, not only in the focal company, but among all collaborators as well. The information systems supporting knowledge work easily remained unchanged and rooted in the practices as they were before the outsourcing. The holistic view of the process was based on the fact that collaboration worked best when systems were not the only channel for information and knowledge sharing. None of the respondents denied that certain knowledge was dynamic and generated during discussions – even when the topics were basically work process related. To some extent informal communication based on personal and organisational relationship was needed, since information systems and documentation were not enough.

During the phase of knowledge creation, interaction was more important than documentation. Even later, in a formalised process, at some point documentation – without supporting interpersonal relationships – began to contribute more to confusion than comprehension. Quoting some of the interviews: ‘When we were working in the same room with designers, the unofficial information search was easy.’; ‘Informal knowledge sharing is crucial in the beginning of the project’; ‘In principle everything is available in the system, but if you have something to ask, it’s easier to contact the person directly.’; ‘Draft documents are commented on discussions, in on‐line meetings or by e‐mail.’

Discussions were often more than information sharing – they generated new knowledge. However, information systems could be used to optimise the correct mix of documentation and interaction to convey understanding, because their role was twofold: they were used for documentation storing as well as for communication. All interviewees of this research mentioned e‐mail as a crucial supporting channel of documentation process. A document management system is used for storing formal releases. The basic features, like version management, workflow, document lifecycle, check‐in and check‐out, compound documents etc. were considered important, but only for the inspected documents.

Before the formal documentation process, more informal storing and sharing methods were used. These included file servers, web servers, wikis and groupware. Nowadays these phases depend even more on communication technology and culture. These more informal document sharing platforms – as well as the formal systems – need a responsible key user: a person who knows about the technology but also about the content of the document bases. Of course there can be standard roles for the users, such as administrators, managers, authors, designers, readers, editors or depositors, but the specification is not clear cut when talking about a temporary, working storage used in collaboration with



196

external partners. The roles are not formal, the users and their roles are changing, and the user can act in multiple roles.

It is crucial that information systems support dynamic roles when organisations and processes change continuously. Otherwise, static roles only reinforce the practices used when the information systems were built and new process innovations have to struggle against the system. Hopefully these impressions of user roles will influence the designers of process‐oriented information and knowledge management systems, which currently can not handle emergent processes very well.

When selecting the right system for each collaboration case, one problem was that multi‐purpose systems had the best features for one purpose only: a groupware system was not the best document management solution, while the document management system did not have good enough support for communication, and so on. Even within the document management category there were discrepancies: differences between inadequate new XML‐based systems and ‘too mature’ SGML based systems, neither of which is best for PDF or other publishing purposes. Therefore it was not possible to define a unique new system for engineering documentation management in cooperative work. For example, some of the respondents considered audit trails or tracing important for a new system (in case III), but none of them said that this feature was used in the existing systems – even though it was available (case II).

The selection of the right system also depends on the phase of the project and process, as do other supporting systems, including project management, meta‐data management and reporting. In summary, the networked characteristics of collaboration were reflected in the information systems on three levels:

1) Both traditional documents and new document formats could be managed well in collaboration, after they had been inspected and imported to the formal product creation process. This included also other documentation specific features, like reporting and publishing.

2) Groupware, meta‐data, project management; portals, dynamic file servers and other extranet servers were placed as a layer between 1) process‐oriented storing and 3) communication. This layer could include data on the persons and their roles in the project as well. This new layer supported some generic results discussed later.

3) Systems for communication were needed in every phase of the collaboration. This included e‐mail and on‐line meetings, mobile communication media and ‘web 2.0’ kinds of applications, but also some features of the document management systems, such as notifications. Informal relationships gave another means to support cooperation between organisations, and should not be ignored.

All systems needed some amount of communication. For example, when a product update was released, the learning curve was significantly lower if personal training was supported by self‐studies and on‐line exercises. The communication was very critical – and most difficult – between collaborators with a different educational background. Time restrictions and linguistic and cultural gaps also existed. However, these should be seen as challenges and enrichment of collaboration. All in all, the question was not to select documentation or interaction exclusively, but the right mixture of process documentation and the right kind of interactive communication to create extended organisational knowformation. That is not mature information to be managed as data in the systems, but yet communicated through the systems. Interactive information is a wrong word. The right word is unfinished communication: knowformation.

Storing and retrieving information

In most cases the information management was handled with a shared storage for project documents to which all partners had access. In the projects studied, the information or document exchange was highly regulated between the focal company and the partner(s). As a detail, version



197

management was found to be problematic, and there was no way to verify whether the document had been read or not. In addition, the information in the documents was often outdated. For example, according to one interview ‘We do have a shared place for the project files. But because it depends on the people that upload the documents on the server, it does not always work as it should. The documents and information may be a year old, or even worse. You never know what you’ll find there. Or then the most recent document version will not be found there at all; it’s just lost in the e‐mail with the other trash mail.’ The various responsibilities for updating and managing documents caused ‘detective work’ especially for project managers who were trying to find the information needed in the current project. In some cases, project members informed the others via e‐mail when there was a new document or a version containing bigger changes, and thus, worth reading. In most cases people seemed to be used to asking whether a certain document was updated or not (usually by e‐mail). Improving this situation required the master data management in a formalised process, even with collaborators.

The most common procedures and means for information exchange included e‐mails, mailing lists, shared databases, and circulation of regular project reports. Due to the fact that people did not save the documents in the shared system, the documents and information had to be acquired from several places or persons. It was difficult to get people to use some of the information systems. For this reason the newest information was not always available, and the documents could often be outdated. There clearly seemed to be room for improvement in version and change management. Of major importance was that either the updated version of a document or the meta‐information about the new version would reach the reader. One user accessing the document management system ‘never knew what (documents) she would find’ but she had an assumption whether the content was valid or not. It should be underlined, that these were problems observed in document sharing with external partners. Also the content, product development, is quite different than more mature administrative processes.

Communication media and ‘knowledge sharing’

Information systems were given a central role and used for sharing, storing and exchanging documents that contained the information and knowledge created in the project. However, the systems did not function properly and people did not use them as expected because of outdated information and unsatisfactory version management. It was a chicken or egg situation, typical of many new systems: people did not use or update information in the system because the information in the system was not up to date. The loss of knowledge value between competitive employees could be underlying reason. The financial issues of the new systems sometimes posed a problem, especially when the justification of budget was based more on the number of users than savings.

In addition, too many systems were available, which was confusing; a few systems or a couple of portals or portlets would have been enough for the interviewees. Some of the comments were: ‘At the moment we would not be able to work without the computer systems. They are basic tools for us. But they could be improved… and people should start using them as they are supposed to…’ ; ‘…they are rather fragmented. …and what we hate is that the systems change each year. Just when you have learnt to use the system someone tells you it will change…There should be some integration and then we could build interfaces through which you can move information from one system to another.’ The interviewees also acknowledged that the information systems were not used for sharing informal information and knowledge, such as lessons learnt during the project execution. Those interviewed felt there were too many tools, the tools did not function properly, they were not user‐friendly, and they were being changed too often. The interviewees wanted a centralised tool or portal through which information and knowledge could be easily accessed for finding relevant documents or people. The result based on these and other cases observations (together with various further discussions) is that content should be independent from the systems. Even though storing could be done with one system, access should have been possible with any client application and new versions. For example, a co‐authored document, the result of brainstorming and a series of electronic mail messages may have been created as part of a project planning. Yet, this important memory was only available through the respective applications, and not utilised.



198

The biggest barrier to ‘knowledge sharing’ (in general, communication) seemed to be that key people were (or at least told they were) too busy to share information and knowledge and to answer all the questions asked. This, of course, depended on the project and the project phase. Another problem was that there were too few personal contacts between people working in different organisations. Sometimes company borders, foreign language, and physical distance could be seen as obstacles, but not as serious as the lack of time and contacts. Additionally, people could not cope with the information overflow and messaging.

The constant lack of time, as well as the lack of personal contacts, was the major barrier to knowledge sharing. Differences between cultures could hinder efficient sharing of knowledge; it became increasingly difficult to share knowledge the more ‘distant’ the cultures were. Some interviewees commented: ‘…it is not about the people being unwilling to help. It’s about lack of time. They just don’t have the time to answer the questions. The lack of time is the biggest problem.’; ‘A foreign partner is a more complicated case. The cultural difference makes it… A good example is the way in which people from different cultures admit that something does not work properly. …it is a challenging task to find the problems.’; ‘I feel that the cultural differences are the biggest (barriers). Language and culture are things that make it more difficult. …But I don’t think that the geographical distance affects it; it’s the same if you speak on the phone to the neighbour building or to the other side of the earth.’ Also, changes in the economic trends and business situation affected the willingness to share knowledge; when the market situation was challenging, people wanted to keep knowledge to themselves, i.e. they did not communicate much in general. Moreover, when extensive organisational changes occurred, the personal networks could fall apart, which complicated the sharing of knowledge. Therefore the importance of data grew when moving from organisational unit towards extended enterprise.

On the other hand, the movement of people put a value on existing knowledge. The physical value of information decreased, but the organisational level of ‘awareness’ increased, which requires a new level between information and knowledge. The limits were fuzzy, not only knowledge vs. information, objects and subjects, but also between ‘emergent processes’ and dynamic communication. In case III system some persons were even defined as ‘knowledge objects’ (see the following figure).

Figure 43. Knowledge object – an example from ‘who knows what’ in case III



199

‘Who knows what’ tool create links between persons seeking specialist knowledge and

key experts who can help them, and use e‐mail profiling to create a user profile. They do not, however, normally treat people as documents or other objects…

The factors facilitating ‘knowledge sharing’ included personal contacts created in face‐to‐face meetings. This enhanced the common feeling of trust, or ‘knowing the colleagues’. These contacts between people collaborating in a product development project facilitated knowledge sharing significantly as they formed a network of resources that could be utilised in different phases of the project. Secondly, good project managers with their own reasonable ways to coordinate collaboration were recognised as a factor improving knowledge sharing. Naturally, the procedures agreed on between companies were seen as facilitators.

Moreover, the current organisational culture, including constant change, facilitated or enforced knowledge sharing by allowing people to ask questions and work together: ‘…nice to meet them face‐to‐face so you know that what kind of a person you are working with. …it is easier to contact them if you have a problem.’; ‘If you know the people, not only from the mail discussions, it helps. Then, if the tools worked right, it would be easier without the extra problems.’; ‘It is easier to communicate in Finnish with Finnish colleagues. You don’t ask if you have to do it in English.’ However, distance seemed to affect the success of knowledge sharing: it was easier to share knowledge when people were close geographically. In addition, it was of course easier to communicate in the native language.

Systems

It is not meaningful to construct systems for systems’ sake. Requirements‐derived system and process are the basis for a working system, which simplifies knowledge by means of information analysis. Case III system was evaluated before tailoring, using System evaluation framework, and technically, as seen in the following table. The detailed reports are excluded from this thesis.

Table 22. A sample from technical evaluation reports

The focal company was so big that the decisions concerning information systems were not suitable for every level of the corporation. Common sense was required when adopting the systems – otherwise adaptation was driven by systems, resulting in the right systems for the wrong use.

In the simplest form of content sharing, a file server or any technology with a clear folder structure was sufficient for teams with fewer than ten persons. File servers are not intended to host any applications or application connections to backend servers, thereby limiting security risks. There are also many new kinds of editable media such as Wikis intended for more public use (although the number of editors is in fact quite limited). For sharing, and sometimes for storing, e‐mail was used – most often as



200

the wrong kind of substitute for a knowledge system, especially with collaborators. One interesting finding was that the less advanced the systems in use were, the more the respondents thought about knowledge related issues; an indication that there is a real need for knowledge management with communication systems.

In terms of communication infrastructure, selection of the communication channel could be based on how quickly the content changes. Draft documents that need many changes should be commented on with more interactive media. For example, inspection or audit of documents can be done in meetings, by videoconferencing, by phone or by e‐mail – depending on the maturity of the document.

The comments in two of the cases showed that e‐mail was not the right media for argumentation. It was a misuse of organisational memory, caused by the unchanged data in the information systems. The analogy is human memory, based on emotions, vs. the physical technological memory of information systems. E‐mail argumentation could result in endemic misinterpretation that was difficult to heal. Even simultaneous on‐line media (e.g. chat) was better, although it could be captured as well. Conversations were preferable, and they could be divided into confidential discussions, and more public telephone conversations from group meetings. The system observations covered the following communication media:

1. text messages 2. multimedia messages 3. e‐mail (html or text) 4. e‐mail with attachments 5. news (usenet, listserv), wikis, blogs, widgets 6. simultaneous media, instant messaging, chatting 7. telephone conversations 8. meetings via network 9. videoconferencing 10. meetings 11. discussions 12. collaboration 13. social networking 14. other kind of presence

Personal vs. public communication could have all variations: one‐to‐one, one‐to‐many,

many‐to‐one, and many‐to‐many. Internet communication tools could be utilised in the intranet only. Another distinction was whether people communicated by their own name or in their own blog. They could refer to existing sources or just throw in anonymous provocative opinions. It is closer to knowledge in the epistemic form, if the content (including references) speaks for itself. Otherwise validity will be evaluated by first impressions and opinions – whether it suits the reader’s values or not. This guides us to a common question of knowledge diversity.

The diversity of communication and knowledge sharing is crucial. A basic question what do we mean by knowledge: is it justified true belief or the common misinterpretation that knowledge includes various kinds of information from communication and presentation models. There is always informal communication that should not be formalised (that is not knowledge). People can communicate via technology, even though it is not formal knowledge in a system. If the communication content is formalised and controlled, then the informal communication moves immediately to an alternative channel. A basic dichotomy is required: knowledge must include a justified true belief as such, or it means something vague as communicated in contemporary standard language.

1.3.3 Case III – final conclusions

In this case study, the inter‐organisational network was designed to support a flexible product creation process, which was divided between the focal and partner companies so that the main



201

responsibility for the innovation remained with the focal company. The ways of collaboration were clearly defined, which was a critical factor for the success of work distribution. The IT supported tools of knowledge work, however, had to be chosen carefully, and there had to be a good balance between process‐based documentation and communication.

In this case study, the work processes, ways of working and even organisational cultures of the focal company and subcontractors were, if not the very same, at least very close to each other, creating the sense of a virtual organisation and common way of doing things. The factors that enabled and facilitated virtual collaboration included direct personal contacts and meetings in person at the beginning of the projects, requiring extensive virtual communication between distant sites. The results revealed examples of the new dilemmas that emerge in networked product creation. People pondered on where and when to work, if colleagues and collaborators were dispersed in time, location, and organisational structures. Great flexibility was demanded of the partners when the focal company’s processes, products, and projects changed during product creation. These flexibility issues in the organisational structures are reflected in the tools of the knowledge work, and designing flexible and modifiable tool architectures becomes increasingly important.



202

2 BUILDING THE MODEL – INFERENCES FROM CASES STRUCTURING THE MODEL

This chapter combines the principles and practices of knowledge management with observations from the cases by constructing a model for extended organisational system strategy concerning actual KM systems. It does not concentrate solely on internal or closed administrative systems. Evolution has reached a phase, where information systems are no longer built for internal managerial use only — even the administrative knowledge management systems are becoming global or organisation‐wide intranet applications, shared with extended parts of the organisation.

This research phase is crucial. The internal validity refers to this analysis phase, where causal relationships are established between the cases (Yin 1989). Causal and descriptive elements are highly interrelated and internal validity and credibility require finding common themes and differences.

Model building is always based on existing models, and has both intellectual and practical sources. Even though this is a theoretical study, it is based on real world experiences and shows the current state of the subject, as well as a model with practical value. The normative instructions finally lead us to a new strategy. The cases compared had the following variants:

Table 23. Variables of the cases

Variable CASE I CASE II CASE III

Level Data Information Knowledge

Scope From individuals and groups to organisational units

From organisational units to focal company, including processes

Extended enterprise, processes between companies

Number of systems Many systems Two systems Unique system (WISE)

Numerous observations were made by comparing the cases. The category of data,

information or knowledge categories was not important. Of critical importance were the content output and how they affected the organisation. It did not matter what the level of content was, but how the content was dealt with – i.e. what was the result of new data, information or knowledge that systems gave to decision makers or any experts. Case II demonstrated the differences in information management based on documentation. Administration, Human Resources, Innovations and Legal procedures are all defined processes on an information management level, even though the subject or content of the processes varies. Case III then, represented engineering (which would seem to be process‐oriented in content) but the content was more knowledge‐oriented. However, there was no difference between those cases level‐wise, compared with the first data management case. The difference comes when comparing new and old content – the ‘knowledge creation’ in communication, including innovations and inventions.

The case studies were compared with each other from both holistic and detailed perspectives. For example, the case I (including data mining), and case III tailored KM system (WISE) were linked through knowledge discovery tools in two aspects. Firstly, analysing large numbers of records and identifying patterns and trends was not a focal point in WISE, but keeping track of who stored or accessed what was a requirement, and several conclusions were drawn with regard to user‐penetration and the effectiveness of the WISE product by studying usage records.

A second aspect of WISE concerning data mining tools was a technical one: data mining and knowledge discovery tools are frequently at the leading edge of technology in terms of database connectivity, data analysis and format conversions. Several ideas and guidelines were gleaned by looking at solutions and tricks adopted by vendors of those tools. These logical analogies were discovered empirically.



203

Although WISE was all about collaboration within large engineering companies, it focused on off‐line rather than on‐line collaboration. Therefore, the support of meetings, discussion groups or whiteboarding was outside the scope of WISE. Collaboration tools offer good insights on the type of documents and knowledge that circulate inside a company or project team, as well as the roles that are involved in the process of carrying out a project or task. However, the generic tools for communication were shown to be very significant in case III interviews.

The high‐level results of the case studies can be summarised briefly: information management and ‘knowledge sharing’ are in active use in the studied network of partners collaborating in product development projects. However, the tacit dimension of knowledge is not given enough attention. The principles should be different. The information systems and other tools supporting knowledge management are focused on storing, sharing and searching documents and other explicit knowledge, while not‐so‐easily articulated tacit knowledge is shared in immediate communication meetings in an ad‐hoc, unstructured manner, creating new dynamic knowledge. Support for sharing tacit knowledge should be improved, for example, by encouraging the creation of unofficial communities (also on an extended organisational level) and by providing technological platforms for them. The systems between information and knowledge manage neither information nor knowledge, but something in between in the short term memory of an organisation.

The organisational differences were minimal between partners that shared a history of collaboration. Instead, the processes, ways of working and even organisational cultures were close to each other. The challenges of collaboration resided in multi‐cultural co‐operation: language and different meanings attached to work could be obstacles for efficient project work.

The factors facilitating ‘knowledge sharing’ included direct personal contacts between employees working in partner companies, and face‐to‐face meetings in the early stage of projects, which required extensive virtual communication between distant sites. Information systems also facilitated knowledge sharing by communication, but they needed to be improved and integrated in order to support easy access to this shared cognition.

ICT support for re‐using existing information and knowledge is necessary. Most of the information produced in previous or parallel projects is needed in order to be able to plan and run the new project efficiently. This also includes a learning dimension: the lessons learnt during the projects may be passed on, and the documents, templates and project plans may be used as models that can be modified according to specific project needs. The organisational memory remains, even though the persons may change. It could be claimed that organisational memory is the only social phenomenon that retains knowledge in an organisation, since individuals and collaborators come and go.

In organisational learning it is important to recognise the shifting borderline between implicit and tacit knowledge – non‐explicit knowledge is much more difficult to reconstruct specifically for tacit, and not only implicit knowledge. ICT can be used to refine implicit knowledge for sharing. ‘Deeper’ tacit knowledge is possible to make organisational only by making people collaborate in practice. This implicit consciousness will from now on be referred to as knowformation, as it is known and memorised throughout the organisation in information systems and individual knowledge processing.

Sharing the information created in the product development process is vital for today’s engineering work. ‘Knowledge sharing’ seems to work better with collaborators, who are in a win‐win situation, than with internal colleagues, who are constantly competing with each other. Reusing information and knowledge created in other projects enables the engineers to avoid extra work and to build on existing knowledge. A mere document management system is unable to support the exchange of informal and often tacit knowledge based on lessons learnt, experiences and new practices that have evolved during a project. Instead, an information system with an integrated approach to knowledge work is needed. This platform or solution, however, should be a combination of systems in an extended enterprise with different user requirements. Knowledge creation and ‘sharing’ is different from processing (while both are based on ICT), and therefore communication, systems and process are independent entities in the model. The new meta‐systems function on the level of pragmatic information. These systems contain mainly memorized organisational consciousness as the basis for knowformation, while persons are changing.

Systems

knowledthat allostandardengineersystem, another should bcollaborcoherena positiv

The resumotivate

for suppsolution;make mapromiseDavenpoclassifiesthings w

managethat theydeeply ethan justcommunin proceusers in not work

To2.1

reuse of memoryintensive

between inf

It is prdge assets in ws interconnd. Ideally thering assets, fa well‐justifinew informa

be dynamic foative work, mt practices ove change in

ults showed ted by the fol

• All ccom

• Casefocu

• Integ• Info• Info

freq• The

line • The

Variouporting know; an extensivanagement ss concerningort & Prusak s objects, onwhile using th

Commd. Most of thy are only abembedded intified true benication and ss‐oriented aless formativk as systema

owards exte

Organf individual ay, intelligent e operationa

formation an

robably advia semanticanection withe knowledge for instance, ied KM introation systemor changing more emphaon how to usethe mindset

that moving lowing reaso

ase studies immunication e studies conused) systemgration of inrmation glutrmation in buently updatcost of infor

support cost

us kinds of tewledge managve change is successful. Mg the process1998). Whee operates whe system. munication ishe current ITble to procesn its context elieves, and memory. Foand communve processesatically as an

nded organi

nisational mend corporateknowledge mal task with a

nd knowledg

sable to expally rich mannh a variety of mapping proa form of RDduction plan

m. System deorganisationasis must be e KM system.

from managons:

indicated the(pragmatic infirmed the is because offormation ist is avoided wooks and othted and mainrmation depl

ts should be

echnologies gement. As nneeded at thMoreover, ths of knowledn seeing systwithin as syst

s creative, dyT tools for KMss informatio(Nonaka et apart of knowor example, tnicative systes. All in all, a organisation

isational me

emory is the e knowledgemanagementall the inform

ge

204

lore standarner that conf knowledge ocess shouldDF/XML Schen is needed tvelopment sns. In additiolaid on form

ms and porta

gement of in

e critical dimnformation)mportance of the contents about solutwith more mher printed mntained activloyment sho

minimised w

and informanoted earlierhe behavioure informatiodge creationtem like in catem of possib

ynamic, ‘knowM are in fact on, which is ial. 2001). Thiwformation. Tthe role of adems. One socrucial distinnal memory.

emory – infor

core of knowe and lessonst services ac

mation neces

Part III – M

dised modelforms to a ribases and usd lead to a prema. Furthero avoid comshould be baon to creatingmalising, trainls, because h

formation to

mension of co of universal (t and systemions, not proeaningful infmedia is ofteve documentuld be minim

when custom

ation systemsr, however, tral, cultural aon systems min processingase III as a sobilities. Com

wledge’ — pinformationnherently diis knowledgeThe division dministratorlution is to rnction is that

rmation syst

wledge manas learnt. Arratively providsary and use

Modelling the

lling languagigid set of enser interfaceroposed newrmore, whenplicating thesed on existg viable systening and creahuman adapt

owards know

ontent matur

(not applicatms should be oducts formation, ben out of dattation mised, by pro

mising and up

s have been technology aand organisamay have failg, not commociotechnicamunication

pragmatic infn managemefferent frome, however, ishould be fos shown in aeplace admit individual h

tems as the

agement. It sanged arounde the user weful for fulfill

e results

ges that can rnforceable rues that adoptw standard fon implemente engineers’ ing communems supportating guidelintation requir

wledge and c

rity – from p

tion or platfoindependen

based on rolete, because o

oviding infor

pdating infor

developed aalone does noational levelsed to fulfil th

munication (cal entity and enables to d

formation, wnt tools, in t

m ‘knowledgeis mostly somoremost betwall cases will bnistrators whuman mem

nucleus

supports shad organisatioworking on a ing this task

represent ules, and t the same or modelling ing such a work with nities, but it ting nes and res time and

ontent was

rocessing to

orm‐nt

es and tasks.of

mation on‐

rmation.

and adoptedot offer a s in order to heir cf. one o creative

which is not he sense e’ that is mething elseween be different ith key‐ory does

aring and onal knowledge‐.

e



205

Organisational memory assists in all of the knowledge management activities: from identification to preservation.

A society or community, such as a self‐steering organisation, must continue to receive a

full flow of three kinds of information: 1. information about the world outside 2. information from the past with a wide range of recall and recombination and 3. information about itself and its own parts Constructivism takes a critical departure from objective knowledge and the search for

universal truth, and argues that there is no universal foundation of knowledge, only the agreement and consensus of the community (Barabas 1990). Since all of these sources of information are subjective, it is crucial that everything is stored in memory.

What is then objective? We can answer this from a practical point of view. Objective data in the corporate memory is stored somewhere in the skeleton of information systems. One unplanned feature of all three cases stood out as a common denominator: case I included a comparison between Oracle vs. DB2 and other databases; case II included comparison between: IBM Lotus Domino Document Manager vs. Documentum and Oracle behind it; in case III tailored KM system, Oracle was used as a database. Some of the data stored in relational databases are objective from the viewpoint of the individual interpreter.

Autonomy in the long run depends on memory. As early as 1966 Deutsch stated that the facilities for memory storage and particularly the circuits for recall, recombination, new storage and reapplication of memory data are critical. The will of individuals or groups can be paralysed by destroying their stored past information or by disrupting its flow into the system. This discussion could be linked to sociology (Durkheim, Giddens, and Berger & Luckman).

In 1979 Weick wrote that if an organisation is to learn anything then the distribution of its memory, the accuracy of that memory and the conditions under which that memory is treated as a constraint become crucial characteristics of organizing. If knowledge is packaged in the mind of one individual presumably the organisation will unfold in a different manner than if the memory is housed in a set of committees with different interests. Furthermore, the organisation's usage of its retained interpretations will also be affected by whether that memory is placed in files, books of rules, or on computers, and how much of that information the organisation admits to.

Functionally, organisational memory is defined as the means by which knowledge from the past is brought to bear on present activities, thus resulting in higher or lower levels of organisational effectiveness (Ackerman & Stein 1996a). Information science has developed many methods and techniques to describe requirements and concepts, and even found a normative theory related to organisational memory.

Organisations store only a small fraction of their knowledge in storage media that are shared facilities and are freely and automatically accessible. Experts may be afraid of losing their expert power (Beerel, 1967). On the other hand, all important knowledge can not be coded.

There has been some discussion about decay of memories and forgetting (Argyris & Schön; Nelson & Winter), but nobody has included the 'vanishing' of knowledge as part of the maintenance function. One reason may be that organisational memory is filled up in every case. The 'turnover' of memories, however, should at least be measured in organisational memory systems.

The other missing link in the OM literature is effect of system usage. Because the organisational memory systems are difficult to use, unfamiliar new work phases force people to learn often unnecessary meta‐information (e.g. content conversions). Yet at the same time this enhances the learning of actual knowledge content.



206

2.1.1 The most important organisational bin is information systems

One result of organisational memory is that information systems are the nucleus, while the OM system is only a conceptual layer above. A dedicated organisational memory information system works after other information systems formulate the organisational memory bin.

Wijnhoven added computer‐based information systems to the bins defined by Walsh and Ungson. In 1998, these had, for the most part, highly formalised memory contents and were increasingly important (cf. Stein & Zwass 1995). All of the retention media have opportunities and limitations for storing knowledge and differ in speed, reliability, physical degeneration and availability (Wijnhoven 1998 p. 36). The article ‘Supporting Granular Reuse of Knowledge Elements in and Organisational Memory System’ (Selvin 1998) presented an approach to creating and populating an HTF‐based organisational memory resource in such a manner as to maximise both the efficiency of the population activities and the likelihood that elements of the memory resource would be retrieved and reused in multiple contexts – already at the beginning of the research. Since then, the web‐based systems, including communication, have become common.

In the current mobile working era, where organisations renew every now and then, the core of critical information is stored in systems. Communication creates new knowledge, but dynamic living knowledge (as communication of pragmatic information) can not be managed as data or information in the systems. However, the something between the two is the already familiar knowformation.

Table 24. Retention media and memory contents (based on Walsh & Ungson 1997; Wijnhoven 1998)

Media Memory Content Systems Planning and decision systems based on data management; document and data

combined in record management, process control systems; groupware; content management, knowledge‐oriented systems; administrative systems

↑ Interactive meta memory model ↓

K N O W F O R M A T I O N

Individual Professional skills; evaluation criteria and results; explanation of procedures, decision rules; personal ethics and beliefs, performance criteria; individual routines

Culture Schemes; stories; external communications; cultural routines; norms base Transformation Tasks; experiences; rules, procedures and technology; patents Structure Task divisions; hierarchy; social structure; formal structure; communication structure Ecology Layout of office and surroundings; building architecture External Client and market characteristics; competition profiles; list of knowledgeable people and

organisations; technology of competitors During this research it was observed that a separate Organisational Memory (like any

knowledge management) system does not work; a new system layer, including other systems below, is the right solution. Knowformation layer exists, whether dealing with data or information management systems.

To survive in a rapidly changing environment, the organisation needs to forget intentionally. Otherwise the memory content vanishes, fades out or is muted. The stability of organisational knowledge is considered as quite self‐evident. It has been written that some organisational level phenomena form emergent and stable patterns that will persist regardless of individual organisational participants. Thus, although all the original members of the organisation may be replaced over time, certain patterns can be expected to endure. Individuals may leave an organisation or an organisational group, but the knowledge of the organisation or the group does not necessarily vanish with them (see e.g. Jelinek & Litterer, 1994 p. 6; von Krogh et al. 1994, p. 60; Walsh and Ungson, 1991, p. 61; Pirttilä 1997, p. 68). The managed renewal of knowledge is argued to be more critical.

Advantageous forgetting, an important function of memory, can be considered in a number of ways. In an information system, forgetting is often considered analogous to a technical

Systems

failure inrepressiointeractiand mut

remembthis resethe orgaavoidingmemory

managem

extendeenterpri

normallythe contplan andwhetherand the

began to(M3.exe

St2.2

The mainsubchap

1 m

between inf

n the storageon of somethion betweenting of inform

Most bering too mearch is that anisation andg organisatioy of an organ

The foment model

Figure 44. M

Thesed organisatiose specific p

Anothy require addtent, processd construct syr the result (sreal content

In theo introduce a), which is ba

tatements on

n statementpters.

Memory is tmemory is thmeans (such

formation an

e or retrievalhing unpleas internal andmation couldscientists couch but misasubjective knd its units. Thnal amnesia isation. Thusollowing figuof an extend

Mnemonic fu

e functions aron, the functrocesses andher result is tditional (planses, applicatiystems againsystem) will t is transferre context of ka developingased on the

n a memory

s based on t

he right methe only conch as systems)

nd knowledg

l of informatsant from a pd external ind be somethionsider forgeapplying knonowledge is he archiving and misunds we come tore illustratesded enterpri

nctions for m

re adaptabletions must bd working prthe side‐effenning, designons and othn. This has pbe valuable ed into memknowformatig concept, thfindings pres

y manageme

he case stud

taphor. Evencept including).

ge

207

tion into or opsychoanalytscriptions (Eng positive ietting as disaowledge fromalways the fand deletionerstanding. To the divisios the mnemoise, including

memory mana

e no matter we generic enrocedures. ct of the (van, testing) woer organisatroven to be or not. Orga

mory from thiion, memorye Memory Msented in the

nt model of

dies are in th

though all ag memory co

Part III – M

out of memotic perspectivngeström etnstead. dvantageousm the past (sefinal aim; it dn of informatThe archivesn between sonic functiong archiving a

gement mode

what the actnough, even t

luable or futork, which coional memoran added vanisations leais interactivey media and Managemente following s

an extended

e following l

aspects of knontent (such

Modelling the

ry. Forgettinve, or a breat al. 1991). W

s. However,ee Conklin 1depends on ition are the ks can be seenhort term anns or processnd deletion

el of an exten

ual work prothough there

tile) work. Inould affect lery related isslue in any caarn while thee knowformamnemonic ft Model of ansections.

d enterprise

ist, and expl

nowledge canh as human k

e results

ng can also bakdown of thWith OM, the

the real pro999). One condividuals ankey success fn also as the nd long termses of a memof content:

nded enterpri

ocess is. For e might be u

formation syearning. Peosues while thase – regardley are buildination. unctions wen Extended E

e (M3.exe)

ained in the

n not be manknowledge) a

e a he e vanishing

oblem is not onclusion of nd thus on factors in long term

m memory. mory

se

an underlying

ystems ple learn hey must re‐ess of ng systems,

already Enterprise

naged, and memory



208

2 Content can not be managed with a single system. There must be an extended enterprise‐wide set of information systems, and a service‐oriented architecture on top of those systems.

3 Information systems should be integrated to offer a role‐based, web‐enabled access and on‐line visibility. Data must be normalised on an organisational level to the third normal form.

4 The knowformation concept is required between information and knowledge. From the knowledge workers’ perspective, there is no difference between data, information and knowledge management systems: What is crucial is the meaning of the knowformation.

5 Knowformation is mainly an organisational short term memory interpretation of information in the systems.

6 Communication can not be managed, only supported. 7 The bliss of ‘knowledge sharing’ is dubious. 8 Culture expands to all collaborating companies, and information shared between collaborators

is purer than internal ‘knowledge sharing’. 9 Knowledge always has context, while information can be modular even in documents. The

creation process should have an end. Knowledge classification is required here, too. 10 References should be used on the presentation layer. 11 A sociotechnical architecture for new kinds of search and other future technologies will come. 12 Communication technology will offer a ubiquitous access to information storages to construct

knowformation. 13 Extended enterprises should interact for learning, and use standards for collaboration. 14 Even though referenced, it is quite impossible to adapt the SECI Model and the ba concept in

this context. They may be valid for one kind of knowledge creation. A new memory management model is required.

2.2.1 Memory is the right metaphor

In knowledge management we should admit, that utilisation of both individual and shared, organisational knowledge entails the definite existence of some kind of organisational level ‘memory’, where previous knowledge and experience can be stored and retrieved (see e.g. Pirttilä 1997, p. 70).

One new contribution of this research is to show how organisational memory adapts to extended organisational knowledge. The processes of organisations are too specified as general results, and therefore the acquisition, retention, maintenance, search and retrieval, but also archiving and deletion, are the only universally applicable functions inside processes. The value‐creating but quite unmanageable function added here is communication, including pragmatic information applied in practice.

Memory is also the right metaphor because it includes memory content as well as memory means. There can be no knowledge management, content management, information management or whatever without management of the memory where the content substance (physical information) is stored. Also, the memory functions are the same, whether it is a more specific process or not.

The memory metaphor also includes human knowledge. A sufficient context to information systems is needed for knowledge workers to interpret the data correctly. All information systems per se, culture, transformation, structure, ecology and external information sources, support the worldview of individuals. Still, personal worldviews are always constructed somewhat differently – a fact that some engineering managers find hard to accept. This should be seen as a strength and opportunity instead of a weakness or threat. Objective information is still the backbone of knowledge for all actors of the community, and diversity should be seen as a main enabler for successful innovations.



209

2.2.2 The set of information systems should be extended

The traditional division of systems into operative (e.g. enterprise resource planning) and reporting (e.g. business intelligence, data warehousing) is not the future solution. Forthcoming operative knowledge will be initiated by communication, and reporting should be planning for the future instead of factual statements about history only. Operative goes cooperative.

A hybrid solution is more than applications. If formalising knowledge requires too much effort, it should be left informal, and processed on the basis of human communication. Information and communication technology can improve the quality of a person's decision‐making and problem‐solving, for instance, by providing relevant informal knowledge in the actual work context. Thus, the machine amplifies human knowledge (Abecker et al. 1998, p. 41). Also, the planning itself creates and clarifies knowledge in a shareable form.

The difference between the cases studied does not stem from the data, information and knowledge; the group, organisation and extended organisation are different. It does not matter what data, information or ‘knowledge’ make up the content of the systems; the complexity is directly proportional to the extent of the organisation. The crucial point here is that this short term memory or consciousness of the organisation is based on the knowformation from systems, and it is the factual base of peoples’ mindset – the implicit mixture of objective information and subjective knowledge.

Knowledge can not be managed with a single system. There must be an enterprise‐wide set of information systems and a whole conceptualisation, like service‐oriented architecture on top of those systems. An extended enterprise should not plan content management with a single platform.

The memory allegory is definitely right when talking about current technology and applications. Traditional information systems are effective when applied to an appropriate task: storing, sorting, computing, comparing and visualising data.

The construction of knowledge (even in a wide sense) is based on learning from information. Organisational learning is not enough. Current corporations must learn from their partners and other collaborators. Extended organisational memory requires information systems to support that kind of learning. This also involves meta‐learning: learning how to learn.

In the wide sense, extended organisational knowledge is the collective experience and understanding of an organisation's processes for managing both planned and unplanned situations. Its management is the process whereby information seekers are linked with sources, information is transferred, and knowledge is reconstructed. The knowformation concept is better to describe knowledge in this wide interpretation. The extended set of information systems will help to manage the meta‐data on personnel’s skills, and support collaboration with the communication infrastructure.

2.2.3 Information systems should be integrated to offer role‐based, ubiquitous access, and on‐line visibility to normalised data and information

One person might have many different roles in an organisation; hence the relationship between persons and actors is one to many, and various user profiles are needed for information systems. One basic procedure – searching – is difficult to profile as long as the user does not know what is not known. The default set of information sources could be based on role and the personal questions arising from that kind of communication.

Ubiquitous access generally means web‐based technologies including mobile devices. However, mobile devices can deal with knowformation, as discussed in a later chapter. End‐to‐end, on‐line visibility will help to organise the model of the extended enterprise. To go a little further, the extranet could be seen as the actual occurrence of the extended organisation. Ultimately, an extended enterprise could organise itself on the basis of shared information.

Data must be normalised on an organisational level to the third normal form. Methods range from star models of data marts to hubs links and satellites of data vaults. The concept here is more generic: the extended organisation (instead of individual systems) is seen as the data storage. Replication is not the right solution even if the data is ‘universal’. One set of master data and modular



210

structured information are preferable for future management. Using modular content management in all structural documentation will partially solve problems in version control, as in cases I and II.

The information architecture resembles object‐oriented classes more than entity‐relationships. Modular documents need some level of de‐normalisation. However, normalisation of data on the enterprise level will already be a big advantage. Communication during research, knowledge creation or content acquisition is thus something completely different. Most knowledge should be regarded as communication, which is only supported, not managed. An individual view of data, information and knowledge, on the other hand, could specify any of these as knowledge: part of epistemic subset of information presentation.

2.2.4 The knowformation concept is required between information and knowledge

Before the new knowledge is classified it is an independent entity. It can be in communication systems, and it is partially constructed from the existing data and information sources. But as for human beings, classification is the main method to memorise also on the organisational level. Knowformation is somewhat unclassified and stored mainly in a short term memory.

Managed data is part of knowledge, while the individual and organisational viewpoints differ. Parts of the data exist in the individual consciousness, while all of the managed data is included on the organisational level. The content that is stored but not actively known about anymore is tacit organisational knowledge.

The memory content which is easiest to remember is emotional. If an issue is of sentimental value, it will be summoned up, even though details may no longer be remembered accurately. What has been missing from the information species is this emotionally remembered, modern unclassified knowledge, supported by information and communication technology. This concept is named knowformation in systems between information and knowledge.

If we interpret knowledge practically – like a generic term from KM literature, instead of an epistemic justified true belief, a crucial distinction between information and ‘knowledge’ is that knowledge can not be detached from emotions, while information can. The emotional issue is more closely related to cognitive science. However, affects can be utilised primarily in memory and knowledge, as the emotions have a fundamental effect on learning. Information systems – in parallel with content – can also arouse emotions. If the planning or use of a system gives positive feedback, it motivates a person to analyse the content of the system, or to use the system. That leads us again to semantic and episodic memory on an organisational level, while individuals’ emotional remembering also covers implicit memories. For example, semantic memory in systems can be authored content, and episodic include historical trace.

The critical question is what knowformation will be outlined so that it will flow towards information and data, and finally belong to categories that are organisationally remembered in a long term memory. As mentioned, any ideas, thoughts, research or the like are not organisational before they are communicated. The question is what content of communication will be stored and shared.

Observed from the case studies, there was no difference between data, information and knowledge, what comes to management systems and their meaning for the users. That is mainly because the data is interpreted in any case. However, as part of the reality construction, the data is more easily seen as objective, the meaning of which is similar for all actors. But in the context of enterprise strategy for management, data, information and knowledge management systems are easily seen as similar entities. This should not be the solution for the future.

The distinction between objective data, information and knowledge, and subjective knowledge is essential, and should be used in the model and strategy. Based on the categorisation presented in Part I subsection 2.3.2, Information species based on qualitative interpretation, the main distinction could be as presented in Table 25.



211

Table 25. Species of data, information and knowledge – reflecting management systems

COMMUNICATION PRESENTATION PROCESSING

DATA Knowledge‐related Epistemic Data, Meta‐data

INFORMATION

KNOWFORMATION

KNOWLEDGE

Absolute values, norms Modal Data‐derived information

Assumptions, moods, intuitions, beliefs, questions, orders…

Doxastic Communication technology, Knowledge systems

In practice the table means that data management systems should include only

normalised, unique, valid and reliable (mostly) numerical facts that are objective and commonly agreed on, other information systems deal with non‐structured data that are mostly truth‐valued, process related and positivistic in a sense. Quite the opposite, however, is the knowledge layer, where opinions are communicated, argued and discussed, and where new knowledge is created.

There are some systems between communication and process‐based storing, such as blogs and other opinion based sources of publishing. The content on such sites is not necessarily valid and reliable, even though publicly available.

If we define data as a systematisation of existing information, information systems aid knowledge management in analysing organisational knowledge. This modelling provides learning and makes the data and information a part of knowledge. The knowformation concept is needed as a layer where organisational knowledge exceeds individual knowledge. It is something shared, but not traditional information or knowledge.

Nonaka’s SECI is based on two dimensions of knowledge creation: epistemological and ontological (Nonaka & Takeuchi 1995: p. 57). However, the actual ‘spiral’ is not a particularly good description of the content. Therefore the axes of the model are also used in a new way here.

We presented the role of language in local communication when discussing semantics and semiotics in subsection of Part I. The value of information in terms of reliability is not necessarily reduced as it moves from human knowledge towards data. Figure 45. Reverse Knowledge Axis presents the idea that knowledge levels or forms move upwards from tacit knowledge towards information. Tacit knowledge resides inside the personnel with different levels of cognition, and thus on a varying organisational summary level. Explicit knowledge, in the organisational context, consists of, for example, operating procedures, strategy systems and usage of assembly lines. An automated project management system and computer automated order processing system are examples of explicit knowledge on the information level, which can more easily be managed and computerised. The other axis describes where the knowledge is: within individuals, groups, the organisation, or extended organisation.



212

Figure 45. Reverse Knowledge Axis

The reverse knowledge axis represents the idea that the importance of different maturity levels of information is variable, depending on the scale of the bracket. I prefer the term ‘importance’ instead of ‘value’, used normally in such figures (cf. Figure 3). On the other hand, the importance or ‘power’ of the individuals, teams and groups also vary. If we see the organisation as an impersonal sociotechnical actor, this axis is valid especially in collaboration. The data should be divided into two categories – according to the physical worldview and the cognitive science or information systems. The data around us include unknown entities or objects, while data classes in the systems were analysed when the model was created. This, axis as well as Figure 46 deal with data within our awareness.

At the beginning of this millennium, it was predicted that knowledge components such as judgment, design, leadership, better decisions, persuasiveness, wit, innovation, aesthetics, and humour would become more valuable because information systems could provide information more easily. After the last few decades, when many commentators argued that information was the ultimate object of every organisation, the value of more knowledge‐intensive skills has become more widely recognised. The premium that companies paid for them reflected this valuation. Though we have seen a tendency — especially among vendors of software — to reductively define knowledge management as moving data and documents around, knowledge management grew out of an understanding of the critical value of these other, less digitized factors, and the clear need to devise ways to support and benefit from them (Prusak 2001). This kind of interpretation of information and knowledge requires a new kind of model, where knowledge is presented otherwise than classical data, information and knowledge pyramid. Values also vary when comparing individual and organisational levels.

The content of the systems is on all levels of data, information and knowledge. The classical data, information and knowledge pyramid is only another (but misleading) way to see the hierarchy. The opposite of this is classification of knowledge as one form of information, i.e. a subspecies of information. The right interpretation, however, is depends also on the scope: individual vs. organisational. In the individual cognition, data hardly plays a minimal role, while in organisational consciousness the data is a basis of many further constructions. Figure 46 includes the knowformation concept. Knowformation from the systems point of view covers knowledge based on integrated content across applications, all document types and Internet content that is independent from systems and possibly includes tags, references and annotations as meta‐data. In the future, users will be free from application and location‐centric computing, and dealing more and more with extended organisational knowformation. This development is already happening.



213

The triangles show that concepts of data and information in the systems are already processed once when the systems have been planned. The context is information from physical surroundings that has not been processed. The cognition of the individual, however, includes much more knowledge and information than unprocessed ‘data’ or physical information.

KNOWFORMATION

Cognitive knowledge

IndividualInformation

DataOrganisational

Data

Extended organisation

Context

Figure 46. Knowformation in individual and organisational triangles

Knowformation is a recursive spiral between information and knowledge. It is partially

based on information; on the other hand, the information is interpreted on the basis of deeply memorised knowledge. In common models, data represents observations or facts out of context that are, therefore, not directly meaningful, while information results from placing data within some meaningful content, often in the form of a message (e.g. Zack 1999/2). Knowledge as a ‘justified true belief’ includes a mood of trust (which is a crucial part of human communication). In practice, the truth‐value of knowledge and information is not required, and the concepts of data, information and knowledge differ from individual and organisational points of view.

On an individual level, data and information accumulates to knowledge through experience, communication, inference and actions, but similarly, most data either never becomes cognitive, or vanishes from consciousness because it is not needed constantly. On an organisational level, memory is in a different position, because content is mostly explicit data in systems.

Knowformation is based on the fact that although explicit data and information and objective knowledge can be articulated into words or mathematical formulae, these must be tacitly understood and applied both on an individual and organisational level. The results are always constructed somewhat differently. Compare the social or organisational DIKW hierarchy with the individual viewpoint described in Part I Chapter 1.5.

Firstly, objective information is a part of knowledge, but subjective knowledge is not the same as tacit. Secondly, subjective and objective division is not binary. Thirdly, there is organisational tacit knowledge in the sense of forgotten documents or even systems (Cf. Wicramasinghe 2003). Knowformation is a mixture of all of tacit, subjective and objective forms, it closest to implicit short term memory or corporate cognition that can be viewed as an ongoing phenomenon, and is shaped by social practices of communities.



214

In the sense of knowformation, the case studies showed that there was no significant difference between data, information and knowledge management systems – they could be all considered as content management when the aim was to create organisational output for decisions. The individual memory does not include much data, while organisational memory consists mainly of data. (Data management systems have also become ‘business intelligence’ applications or the like.)

The cognitive processes of individuals do not deal with masses of data, but individuals refer to their information and knowledge in their thoughts. On the other hand, organisation is dependent on the shared data in their systems. Content covers all layers, while context is sociotechnical for the collaboration systems of an extended organisation.

2.2.5 Knowformation is mainly organisational short term memory

Everything is information, except disinformation. There are reasons to define implicit knowledge as being between tacit and explicit knowledge, and to clarify missing with knowledge meta‐information. Implicit information containing some features of knowledge has already been called knowformation. The case studies showed that if collaborators or colleagues did not know what was to come or how co‐operation would change, they could not ask relevant questions. The eternal division of knowledge is omnipresent.

Table 26. Knowformation and meta‐information

KNOWN NOT KNOWN

META‐INFORMATION Explicitly known knowledge Things to learn

FORMULATING META‐INFORMATION

Knowformation – Implicit awareness of information, deeper understanding of context missing from organisation

NO META‐INFORMATION Tacit knowledge Unknown

Tacit knowledge here could be a smaller category compared with its definition in current

articles. Knowformation comes between tacit and explicit knowledge. Meta‐information tells what is required for successful work.

Some paradigms have claimed that knowledge is not the object of management, but knowing is. For example, McDermott (1999) defined knowing as a human act and wrote that knowledge involves humans who do the knowing. On the other hand, case III included the concept of ‘Knowledge Objects’ in the system (see Figure 47). This discrepancy is solved in our information species and individual and organisational triangles, as the epistemic knowledge can be stored in systems, but when it is ‘managed’, it becomes communication, and the common (not epistemic) ‘knowledge’ concept is more like knowformation.



215

Figure 47. Knowledge objects

Knowformation is midway between two extremes. Knowformation is the part of

knowledge that is recursively constructed based on information in the systems, which is managed as objective information.

Knowformation is substance for communities of practice and other informal organisations within the extended enterprise, when explicit information should be re‐interpreted, re‐created and appropriated to contextual, often tacit knowledge about organisational practices and processes (cf. Wilson et al. 1994; Swan and Newell 2000). One critical factor in knowformation is that implicit trust is included as meta‐information. The organisation knows the fuzzy level of credibility based on the information itself. However, the procedures depend on the knowformation despite the opinions of individuals. The objective part of that knowformation is the basis for a shared understanding of extended and organisational worldviews.

To further clarify the knowformation another method (in addition to the DIK‐triangle) has been used to describe the concepts: circles ranging from a smaller data circle to the widest knowledge circle. That is also misleading. Data and knowledge are both subsets of information. Figure 48 suggest an alternative way to describe the sets of data, information and knowledge.

Figure 48. Alternative sets of data, information and knowledge

KM literature mixes information management with other sciences; thus the information

species are mixed as well. Here information as the main class covers everything. Entropy is used here,



216

but not in the sense of physics. Entropy is a better word than disinformation as we are not talking about misleading information, but just the opposite of information. Knowledge consists of justified true beliefs, and not everything that the KM literature imposes. However, it does matter do we speak about data‐derived knowledge in the system or epistemic information in presentation. Systems store information so it is actually the same thing as long as we do not speak about communication. Thus knowledge can be stored (specified as presentation) in the system, and the belief is only an explicable point, as computers can be programmed to tell whether the information is true or justified, whereas the only believer is the user who reads the knowledge.

In this model knowledge per se can not be tacit; seeing it as contradictory to explicit is due to a lack of understanding or knowing. Therefore, know‐formation is a suitable new concept as both parts can be derived from verbs. Data has caused errors in the prior models because the meaning is ambiguous: we can talk about data when we mean information that is already inside the system or planned to be put into the system. In that sense data is already planned and therefore more refined than knowledge. Data is ‘given’ to the system (Latin verb dare). In English, content is already inside the system (even though an etymological misinterpretation from Latin con‐tenere still exists).

Case III’s analysis of the use of editors was one example of experience‐based knowformation. An author should know which SGML or XML editor to use for which project, because the features of the editor result in quality differences in the output. The same unofficial information could be found when reading various instructions. This kind of knowledge that is neither tacit nor explicit. The category of knowformation is more descriptive. The presentation layer of the organisation can be used for storing knowformation as a short term memory.

Already Polanyi saw tacit and explicit knowledge as two dimensions, and wrote that ‘all knowledge is either tacit or rooted in tacit knowledge’ (Polanyi 1966: 7). However, the knowformation layer describes the practice better the two strict categories in Nonaka’s SECI‐model.

In the systems domain, the future trend is that this knowformation will be gathered as collective meta‐knowledge — tags, comments, descriptions or alike — into the knowledge object itself, instead of modelling the object another time in some other system. Knowformation will be formulated on the basis of information in the system, as the organisational systems are most likely planned and used by different persons and the ‘explicit’ information in the systems is interpreted by every person differently.

2.2.6 Communication can only be supported, not managed

Information parts should be normalised to one set of data in the systems. This is based on a different philosophy than knowledge. Knowledge can not be normalised, nor should it be constricted. Pragmatic communication includes mainly expressive information, not the factual data of information systems. Information and communication technology is supporting knowledge transfer only on the basis of communication. Process‐based information should have a lifecycle, but communication does not have any managed features. Versioning is also different: new data and information replaces the old — this is not the case for knowledge.

When people are connected, the ‘knowledge’ level of DIK‐triangle is closer. Communication, however, is dynamic and active, including expressive information, such as moods, intuitions, beliefs. The only thing systems can offer is the media for communication.

Information technology offers media ranging from one‐way to simultaneous two‐way communication (publishing, e‐mail, instant messaging, chatting). No matter what kind of communication technology is used, organisations need a short term memory between information storages and communication.

One clear indication from cases II and III is that e‐mail should not be used as a delivery channel for documents. It results in large numbers of futile documents and some important contacts missing from the lists. The best solution is personal lists of recipients of links to documents.

Cases II and III also show us that interaction and communication are needed in document management. First, the author saves the documents in a common document base. How many people notice that, and how many people actually read it, is the kind of feedback that should be delivered to



217

the author. The numbers alone are not enough. Actually, the statistics could be misleading if the hits are measured. Communication and interaction are the only valuable kinds of feedback for the author. Of course, the original author must be notified about possible other authoring. This is especially true with extended organisational knowledge, when the documents are stored in different organisations.

When information systems are expanded with communication technology, the level is raising from information towards knowledge (in the wide sense). This ‘knowledge’ can not be managed, only supported, because cutting edge of this information is always changing. In this communication context we should be critical, is any of it actually epistemic knowledge – and does it really matter for the communicators.

The real task is to connect people to people. This can be done partly by mixing and alternating tacit knowledge sharing (e.g. meetings, discussions, job rotation) with explicit knowledge sharing (e.g. documentation) in the right proportion, scheduled correctly. The learning can be supported by mixing the instrumental information exchange (replicating, imitative learning) with knowledge creation.

Some systems, such as blogs, come between communication and storing. However, information on them is not truth‐valued, or valid and reliable – as it mostly is in data and information management, even though it is publicly available. Real‐time, on‐line meetings – including brainstorming, ad‐hoc meetings and decision‐making are communicative kind of collaboration, the results of which can be memorised by current meeting tools.

2.2.7 The bliss of ‘knowledge sharing’ is dubious

The mantra of ‘knowledge sharing’ should be criticised. First, subjective knowledge is not shared as is, but reconstructed, and the value of objective information is decreasing. One new finding is that information sharing between collaborating organisations works in a more formal manner and thus more effectively. The extended organisation’s ‘knowledge sharing’ is purer than internal, the reason being that collaborating companies reach an explicit common target, while colleagues inside a company also communicate all kinds of value‐laden opinions in a competitive situation.

When managing a network of people, dynamic knowledge is based on communication. This kind of ‘knowledge sharing’ is valuable. However, it is actually not knowledge or even information sharing except in the sense of communication. The communication is fuzzy, and all the concepts, sentences, and mental impressions – from semantic triangle to emotions – are constructed by the listeners according to their personal worldviews. There are seven boundaries to ‘knowledge movement’: temporal, technical, social, political, geographic, cultural, historic (Orlikowksi 2002) and they are valid also for ‘knowledge sharing’.

If ‘knowledge sharing’ is to be valuable, we should always concentrate on knowledge that is important – not only for the organisation but for the people. Learning communities are based on topics. Developing infrastructure for communities of practice is wasted effort if they are not used. The best way to benefit from communities is to identify the topics where leveraging knowledge will be useful to the organisation as well as to the community members.

Communication can not be totally controlled. Management can not control what information should be shared and what kind of communication is needed for organisational knowledge construction. However, the communication infrastructure itself needs to have a management to administer the media. It is obvious that the more general the information is, the more people need to know it. The deepness of the information, however, is another dimension. Therefore, it is recommended that narrow and deep expertise is channelled so that it will not obstruct general, shared knowledge, and vice versa. In practice, one way to do that is to manage subscriptions or user accounts for deeper expertise documents and keep the general information available to everyone.

When objective information is shared, its value (and hence that of the people who have the information) decreases. On an organisational level, business costs are to be minimised and reduction of personnel will help the companies to make more profit. On an individual level, objective information sharing is risky and mainly disadvantageous. Organisational policy is challenging: enterprises should encourage people to make themselves futile.



218

Knowledge management requires a strong leadership that indoctrinates a culture of information or ‘knowledge sharing’. Whether knowledge management is implemented in a centralised fashion or in a more decentralised system, the systems require information sharing, as they have ever since Communities of Practice. That is one reason to modern trifling ‘social networking’.

Organisations could try to use different incentive systems to make sure that information (or ‘knowledge’) is shared. One solution is to give people better positions in organisations. Another solution is to give people more free time for being innovative and hence give added value to the company. Observations during this research indicate that such sharing incentives are either impossible or of marginal benefit, if they are not strictly implemented in company policy at all organisational levels.

The types of incentives and the ability to measure contributions to knowledge management are generally contingent on the level or function in the organisation and the particular application to which the knowledge management system is being put. Inevitably, however, incentive systems are based on measurable activities.

Whether such incentives actually foster a cultural framework where employees feel it is in their best interest to participate actively in knowledge management systems remains to be seen, but this research does not indicate this. Clearly, however, such knowledge management systems benefit corporations that take advantage of the technology. As enterprises being driven toward knowledge management systems to meet competitive pressures and create value, they are increasingly finding that these systems can facilitate reuse of exiting knowledge and create new knowledge in an effort to allow better decision‐making. The ‘shared knowledge’ is based on the experiences of making such systems.

The presentation of existing information is political in nature. People who are eager to ‘share knowledge’ most often express the thoughts of other people, and their motive is vested interest. If you are able to show, that instead of presenting new information you are able to create it, your value will be evaluated by the content: how well you can create information that fits the enterprise’s strategy. The difference in value between people who create information and presenters is often hidden from the decision makers.

Sharing requires motivation. According to Peter Drucker (2000) the historic shift to self‐management offers organisations four ways to best develop and motivate knowledge workers: 1) Know people's strengths, 2) Place them where they can make the greatest contributions, 3) Treat them as associates and 4) Expose them to challenges. These management guidelines are still valid, but the ‘knowledge work’ has also drifted towards one sort of communication. Management support for ‘knowledge sharing’ could lead to a situation where there is non‐valuable communication to non‐targeted audiences, with knowledge being lost in disinformation.

Later it has been claimed that managers need not only specialised knowledge, but wisdom, which is interpreted as ‘the ability to weave knowledge together and make use of it’ (e.g. Mintzberg & Gosling 2002). However, modern leadership could be somewhat different. Specialised expertise is not needed as much, and the real management of expertise is based on organisational autopoiesis. If the knowledge worker actually discovers who he or she is, the worker can learn to manage him‐ or herself. The main policy is not to force employees into artificial ‘knowledge sharing’ but to renew the organisation regularly so that employees can give their input when adapting themselves to new collaborating situations.

Knowledge sharing is dubious also in the terms of the extended organisation and management. Managers may share knowledge with each other somewhat and try to look like they are communicating openly with their subordinates, but the fact is that most information is classified. With an extended organisation, knowledge sharing limits are even stricter, and information is delivered from only a small subset of contracted content.

In contrast to memory content sharing, intelligent organisations are often developed as an organisational memory means for retaining, processing and filtering out essential external information for transforming this into policy‐relevant insights (see e.g. Westney & Ghoshal 1994). Knowledge is power, and workers’ – as well as collaborators – knowledge sharing is encouraged only to inflate their value and keep management aware of their results.

Knowledge can not be transmitted or received as it is. The ‘sharing’ in KM so far must be interpreted as the communication in which the actor is the presenting person. The coexistence of



219

process and communication is still valuable, and a fraction of objective information could become common knowledge.

2.2.8 Knowledge always has a context, while information can be modular; the creation process should have an end

Knowledge always has a context, while information can be modular. Their creation process should have an end. Knowledge creation should be defined on the basis of a process that has an end, and knowledge classification with context is also required here. The term ‘knowledge creation’ is as vague as ‘knowledge sharing’. In case III, the focal company did the innovative work related to product creation. Partners produced very limited and pre‐defined parts of the product according to the requirements and specifications of the focal company. ‘Innovative work’ was only a part of formal product creation process.

Both knowledge creation and sharing are connected to organisational knowledge construction. Communication does not necessarily lead to knowledge creation. Research is closest to knowledge creation, even though some (mostly low‐value) knowledge is created accidentally in discussions. Why is the knowledge creation concept used instead of research and why is the knowledge sharing concept used instead of communication? The answer is obviously that these old processes are given new names in order to unify them under a paradigm of knowledge management. In any case, the creation process is not an eternal loop; it has an endpoint. New knowledge is based on old, but then the process changes recursively.

The term knowledge creation is definitely too wide if it is not based on knowledge itself. As mentioned, knowledge creation in the form of research is different from innovation, and communication can create some new information, mostly moods, intuitions, beliefs, which have no explicit truth‐value. A better term (related to systems) is knowledge capture. However, organisational memory acquisition covers both creation and capturing. Everyday tools for ‘knowledge creation and capturing’ could include MindMaps, SmartBoards and Office tools, but communication applications are still more interactive.

As seen from the analysed information species, the physical species is important only in terms of syntax and semantics. Nowadays the physical model is based on the quantum perspective, and is valid for this research only as a possible technology enabler of communication and information management. Possible quantum computing can make a rational assessment of an almost infinite complexity, but does that make sense to people depending on the success of scenario‐driven ‘wisdom’ instead of fact‐driven decision‐making? (See e.g. Tissen et al. 2000.) If the physical worldview has been furthest away from intuition, emotions or empathy, why would it utilise them on top of rational analysis in the future, even though information processing technology should change towards social computing and more human communication?

The classical data management categories, such as (business) events, configuration and master data, as well as their dependencies, are for technology only. For human work, such categories could be used for an analysis of knowledge, but knowledge is not independent from context.

Storing additional contextual data together with information proper helps when information is used as a resource for action and reinterpreted according to the needs of the new situation. In the early stages of this research, Kuutti (1998) constructed a framework of contextual information needed in collaborative reinterpretation. He showed that the 'why‐level' is important in contextual data, and that both the organisational memory and ‘common information space’ type of support systems would need a similar knowledge structure to facilitate perspective making and taking.

The sociotechnical paradigm emphasises the importance of work entities. A self‐steering group or individual needs whole entities of work, not meaningless parts of tasks. However, the responsibilities must be shared as well. In case III inspection is one example of a duty that the engineers of the focal company do not see as their primary job.

All process‐based information should have a lifecycle. The versions are also useful, but in the case of knowledge, new information is not just replacing old. Therefore the model should be divided into processing vs. communication on a higher level.



220

Knowformation could be considered as information from a system with a human context. For example, case III application Knowledge Objects require measurable personal relevance to be included in the appropriate context (see Figure 49).

Figure 49. Relevant knowledge objects in personal context

The ‘knowledge worker’ needs meaningful entities, procedures or whole processes to

work with – are they considered knowledge or not. Modular information pieces are for systems only. How to combine them in a mature model is a critical question to answer.

2.2.9 Culture expands to collaborators and knowformation

Organisational culture (Schein 1999) has been defined as a property of a group of people, which includes the shared assumptions, attitudes, values, beliefs, and simply the common ways of doing things. Most of the culture of an organisation is invisible; it is a shared mental model or a mindset that manifests itself in the daily actions and behaviour of the members of the organisation. Organisational culture can be understood through a three‐level model incorporating 1) the invisible, basic underlying assumptions, 2) the values, strategies, and goals, and 3) the visible artefacts, such as the organisational structure and processes of the organisation. The culture significantly affects the ways people act and interact in the organisational contexts (Schein 1999; Davenport & Prusak 1998). In virtual work the perception of organisational culture may become extended so that it incorporates the whole virtual organisation, resulting in a feeling of cultural ‘globalness’ among the employees (Burn & Ash 2000).



221

McDermott has claimed that ‘if a group of people don’t already share knowledge, don’t already have plenty of contact, don’t already understand what insights and information will be useful to each other, information technology is not likely to create it’ (McDermott 1999: 104). But the case studies have shown that information technology provides the motivation (fake or not) to work together and thus ‘share knowledge’. Another level is the communication infrastructure which is supporting all kinds of collaboration. A sharing culture could be critical for the success of knowledge management, but the background of that culture should no way be artificial carrot‐and‐stick methods that are creating artificial role‐playing among workers. Enterprise‐wide content management with scientific references in selected systems accessible to collaborators could be a solution to presenting the right information instead of quoting hype from each other’s files.

There can be different cultures of systems; one might concentrate on specific types of users. They may include identifiable groups, like scientific communities. Another culture could be a task focused system, like decision support. These categories will be reflected in the collaboration and nature of the behaviour.

Information is meaningful mainly to the community members that are creating, sharing and using it. For an extended organisation and collaboration between enterprises, knowledge construction is the key issue in sharing information and knowledge. The ways to share information and reconstruct knowledge should be as multidimensional as the content is. Although the planning of systems and content develops knowformation from communication towards processes, shared knowledge construction should not begin with the use of information systems alone. Communication technology is used for communication, while information storing is based on more mature process. People should not be forced to use the tools. The culture of passive learning could be a solution for information feed, that is, publishing, but information sharing (in a way that organisation will remember) is to be done in a channel that supports knowledge construction.

In practice this means that information system must be related with communication technologies to deliver active documents and content – knowledge in a wide KM sense — and not only information. The channels could be face‐to‐face, information systems with communication abilities (whether asynchronous or synchronous), speech, audio, written communication, even audiovisual media. This will support dynamic collaborative thinking and not just static presentation of information.

These findings are extremely valid for inter‐organisational collaboration. There must be trust between organisations – legal contracts as the core. To encourage trust between individuals, there must be daily work, common problem solving, and not just contractual obligations. In this way organisations learn to create a culture of common understanding, including information systems and tools, approaches and solutions.

2.2.10 A presentation layer is needed

The actor‐network theory does not describe the daily practice of organisations as concretely as the presentation layer of information requires. Various sharing tools have been used by organisations for many years (Hayes 2001). In the focal company, these particular systems between information and knowledge have gained increasing importance as the network of organisations has extended. The actor‐network perspective covers more communication form of knowledge instead of presentation. Memory is critical factor also in this communication: if new knowledge does not have meaningful context, it is only information overflow.

As networks of people are connected by information and communication technologies, dynamic knowledge requires information visualisation, which should be independent from the content formats. Separate presentation and communication layers are needed for various information systems including visualisation. This is included in a new application function. In addition to actual data normalisation, this gives a good opportunity to use references better – also in collaboration documents. The independence required in case II means that the content can be edited with any tools, but the back‐end system for storing the files could be unified throughout the enterprise. This of course is a challenge for extended organisational collaboration – so the content must be independent also between corporations as well.

222

Information visualisation (towards communication) and presentation (towards storing) should be different layers than the information itself because information objects can not be independent if the presentation format is part of the object. This will make storing independent from communication and also documents independent from languages. The XML‐format documents could be thus compiled from modules in any language.

In XML there is a distinction between data and document management and a combination of them both. See e.g. the following List 1. Case III XML code sample.

<?xml version=‘1.0’ encoding=‘UTF-8’ ?> - 

- <xs:schema xmlns:xs=‘http://www.w3.org/2001/XMLSchema’ xmlns=‘http://www.ist-wise.org/Person0.xsd’ xmlns:wdt=‘http://www.ist-wise.org/Datatypes.xsd’ xmlns:wko=‘http://www.ist-wise.org/KnowledgeObjects.xsd’ targetNamespace=‘http://www.ist-wise.org/Person0.xsd’ elementFormDefault=‘qualified’ attributeFormDefault=‘unqualified’ version=‘0’>

<xs:import namespace=‘http://www.ist-wise.org/Datatypes.xsd’ schemaLocation=‘Datatypes.xsd’ />

<xs:import namespace=‘http://www.ist-wise.org/KnowledgeObjects.xsd’ schemaLocation=‘KnowledgeObjects.xsd’ />

- <xs:element name=‘Person’> - <xs:complexType> - <xs:complexContent> - <xs:extension base=‘wko:KnowledgeObject’> - <xs:sequence> <xs:element name=‘Title’ type=‘wdt:string’ minOccurs=‘0’ /> <xs:element name=‘Note’ type=‘wdt:text’ minOccurs=‘0’ /> <xs:element name=‘URL’ type=‘wdt:string’ minOccurs=‘0’ /> <xs:element name=‘Role’ type=‘wdt:string’ minOccurs=‘0’ /> <xs:element name=‘Email’ type=‘wdt:string’ minOccurs=‘0’ maxOccurs=‘unbounded’ /> <xs:element name=‘NickName’ type=‘wdt:string’ minOccurs=‘0’ /> <xs:element name=‘Org’ type=‘wdt:ObjectID’ /> <xs:element name=‘BDay’ type=‘wdt:date’ minOccurs=‘0’ /> <xs:element name=‘knowsAbout’ type=‘wdt:string’ minOccurs=‘0’

maxOccurs=‘unbounded’ /> <xs:element name=‘OrgRole’ type=‘wdt:ObjectID’ /> - <xs:element name=‘Telephone’ minOccurs=‘0’ maxOccurs=‘unbounded’>

…

List 1. Case III XML code sample

The same kind of analogy could be reflected in this study: case I as data management,

case II as document management and case III as data and document management. It seems quite obvious that in the near future all applications and content will be XML based, for example groupware, word processing, portals and even messaging. This enforces the idea from organisational memory, that memory means and content are collateral. This could be seen as an initial idea of web technologies.

Presentation species relates mainly to storing of information, for example through the structure and internal organisation of data. However, it can be indirectly related to communication in practice, for example through the representation of data in the user interface.

Presentation from case I follows the most common classification of data structures, but we can identify at least five information structures in organisations with different chances of redundancy and valuations of redundancy (and hence normalisation). These data structures are flat, hierarchical, network, relational and object structures.



223

Memory structures are far from static entities in which memory elements are retained and retrieved; instead they are constantly updated and refreshed. This requires that the structures must also be modelled in a dynamic way, and that it must be made clear under what conditions the memory content can be updated. That is far from communication information species, and requires independent presentation.

For years already client and server side presentation layers have been separated from the business logic behind the enterprise information systems. So the presentation concept appears in technology also – in slightly different semantics, though. Here the presentation layer will enable epistemic knowledge to exist in systems.

2.2.11 A sociotechnical architecture is needed for new search and retrieval and other future applications

Search methods already enable employees to find key knowledge held within databases, shared areas, websites and repositories (documents). In future, search technologies should be based on another technology in order to ‘intelligently’ support real knowledge work. The search platforms changed a few times during the research, and many new possibilities of search were tested in case III (see Figure 50. Case III advanced search options).

Figure 50. Case III advanced search options

Cases II and III in particular indicated that the search is a critical issue in knowledge work,

especially when the process is not strictly specified. Searches can give results for the subject that users do not yet know about, and retrievals give results that the user knows on a meta‐level. Case II showed that the indexing on the current tool was not correct, and the Boolean searches were not used enough. However, the evolution has gone in a different direction. The search platforms of the focal company are already internally under heavy duty –more than 135 000 searches are made every day. There have been plenty of products, from AltaVista and Inktomi to Autonomy Ultraseek and Verity Enterprise Search. The challenges indicated by case III are searches based on content, such as in semantic web.

The semantic web is an extension of the World Wide Web where different types of content objects can have machine‐understandable descriptions associated with them. This expansion of the Web will allow computers non‐ambiguous process content based on information properties. In



224

practice this results in increased automation of content management processes and the ability to find the relevant content using a semantic search instead of a keyword‐based search.

The keyword search seems to be a legacy, but language support is in daily use. For the extended enterprise, it is challenging to decide what will be shown as search results. On the basis of this research, public abstracts are mostly harmless. Abstracts can be even beneficial, if they arouse interest in learning. Cases II and III indicate that organisations should encourage co‐operation between units, and to trace searches. Meta‐information on what the knowledge workers are interested in is recurrent on the portal design.

Cases I and II showed that Lotus Notes search capabilities are limited, partially because of security features. The policy should be changed so that everything is published inside the enterprise, except when otherwise signed. Even though departments and teams are related to the rest of the organisation (and beyond) and need to share data and work, departments and teams also need distinct boundaries for security and maintenance of identity. Therefore, the team’s department needs to control its level of memory sharing with the rest of the world. This is especially true in the case of new extended organisations.

The other Web ‘2.0’ technologies could be used for more open services in extended organisational networks as well17. For example, Wikipedia technology is as easy to install to extranets, as it is in the intranet or the Internet. Blogs and RSS feeds are in constant use, but they are far from presentation of epistemic knowledge, whereas the idea of Wiki is to function as collective intelligence where no‐one is an owner of the content, but anyone can edit anything (as interpreted by one interviewee). Then there are some group specific wikis, and extended organisational wikis also exist. As the content of wikis does not have a responsible editor, it was supposed that the information quality will be low. However, the main problem so far was not wrong content, but that people did not give input to wikis.

More rationalistic solutions behind Web 2.0 include Web Services as a collection of integration technologies that are able to expose core business processes to the network. This architecture has been delivered in a bottom‐up fashion where the transport layer has been specified and deployed (SOAP). The service description layer (WSDL) makes it easy for people working with different programming languages to access SOAP accessible services. The service repositories (UDDI) are now becoming more relevant as the number of available Web Services increase and developers can publish and find relevant services from a centralised repository. More and more technologies are now being standardised on top of this solid foundation that addresses security, policy management, business process modelling and ‘business process orchestration’.

When this research began, some organisational memory systems were already designed for stand‐alone use, i.e. the design assumes that the user will start the application, perform some tasks with the OMS and then close or move on to another application. Other systems were designed to run on top of or as an ‘add in’ to existing systems and data. There were also portals like in case III. The future, however, will more likely be in autonomous resident sociotechnical applications, separated from real data.

The use of search agents was not discussed much in the interviews. It looks as though alternative findings are true: developers should not expect users to manipulate agents. Cognitive theories can state what are the benefits of modern architecture and what is required for knowformation derived from it. In principle we should socialise computers and not mechanise users (See Watt 2000). It will be best for the user if the interface is unique or invisible, and the epistemic knowledge comes from relevant applications and databases that they can count on. The web portal is current mainframe.

17 Web 2.0 – as well as Office 2.0 and other version updates – are actually marketing buzzwords, compared with the initial www ideas. It seems as though their real meaning is not mature enough to be described in actual word.



225

2.2.12 Mobile communication technology enables ubiquitous interaction to formulate knowformation

KM has adopted an old management requirement: right knowledge to right people in right time. Nowadays it should be more specific: The right management information should be available all the time anywhere.

Except for management, mobile technology can also be used for learning purposes. As an example, e‐learning by mobile communication media could include personalised services, not just a simulation of a classroom. Together with mobile news and chat, there could be bulletins with push technology about engineering product information. These mature technologies are already in use.

Text and multimedia messages could be used for expert services, to deliver series of questions and answers. The reply time could vary from instant messaging to hours, but still be interactive. More information can be delivered by e‐mail and other messaging media. Other services could be comparisons with other students’ answers, and a list on hints and favourites. Voice data is easy to use for delivery of content and best practices, including product features for a reseller. Another application is to record students’ own comments.

However, one further conclusion of the case studies is that the ubiquitous knowformation is clearly a requirement. This could be achieved to some extent with, for instance, Lifeblog, SenseCam, What Was I Thinking, MyLifeBits or alike. They are early examples of ‘lifelogs,’ (aka Cyborglogs or ‘glogs’): systems for recording, auto‐archiving and auto‐indexing all ‘life experience’. Like Wikis, these can be used to more epistemic purposes as well; depending on the content provider’s expertise as such. Mobile devices will enable the storage of everything that has been heard or said, and then everything that we have seen. All this storage, processing, and bandwidth will enable us to network in new ways. When collaborative filtering is added this data begins turning into knowformation. The retrieval, however, should be selective as human memory. Even though knowledge in systems is stable, the human memory constructs memorised information every time differently, and learning happens based on existing knowledge. From systems point of view this bulk of content will formulate towards knowformation when the receiver has an active part in the process.

2.2.13 An extended enterprise should interact for learning and use standards for collaboration

Organisational learning means two things: First that the consequent projects can learn from history, but also, that parallel processes can learn from each other simultaneously. There are two levels of learning: learning the content and meta‐learning, or learning how to learn.

The new memory management model for an extended enterprise (M3.exe, see Section 2.3) will impact on organisational learning in four ways: It is expected to enhance lower level learning by providing a history of standards against which incremental improvement can be made. Then M3.exe will offer a place to encode the results of higher order learning. Third, it will promote higher order learning if information from the past is made the basis for review and modification. On the caveat side, M3.exe may hinder higher level learning if members fail to question underlying assumptions and paradigms.

As a traditional example from airplane manufacturing, observers noted that building the second airplane of a given type took considerably less time than the first one, and the second airplane had fewer defects than the first. In other words, it was proven that workers really did learn from experience. Already in the fifties, the Rand Corporation began to analyse and codify observations of this type. The phenomenon was given its classic expression in economist Kenneth Arrow's 1962 article, ‘Learning by Doing.’ This airplane example suits case III, where another industrial partner related, for example, that every now and then the retired employees tutored the organisation, as their tacit knowledge was the only way to retrieve certain solutions. This knowledge relates to design science, where the knowledge is ‘scientified’ through everyday experience, based on classification of research in an article ‘The aim and structure of applied research’ (Niiniluoto 1993).

Learning by doing is definitely working in the design of information systems. The content and processes of the organisation will be analysed and the subject learnt – no matter how successful the actual systems ultimately are. In many cases relatively new technology has been given a more significant role and emphasis than it deserves in the discussion of problems connected to the design and utilisation



226

of information systems. It should be realised that the difficulties observed in many empirical studies of management information systems and information management in general are not necessarily a result of inadequate or inappropriate design of these systems. These problems should be approached more broadly as difficulties in cognitive capabilities, i.e. in the whole process of information acquisition, interpretation and learning, for example, in managerial and expert work.

Individual learning can be reflected on an organisational level. Novices frequently solve problems by following step‐by‐step procedures, while experts solve problems in a holistic way. They typically develop a theory of potential causes based on their experience and test to see if the theory is correct, often testing the least complex or expensive theories first, rather than the logically correct ones (Konrad 1995). This could become a way to work on the extended organisational level also.

Communities may realise that they are doing somewhat redundant work. For example, when information is collected from external sources, it is easily done multiple times in an organisation. To avoid this there should be a common ‘library’ and someone responsible for documentation management dealing with external collaborators.

Learning is enhancing the competence of the individual as well as the organisation. Competence refers to a set of capabilities that rely on knowledge, enabling to proper action. Usually learning and practice is necessary for a competence to emerge and develop. (EKMF) Competence is defined as the sufficiency of qualification, or the ‘individual part of qualification’; capacity to deal adequately with a subject (Oxford English Dictionary). Competence is described as know‐how that includes a certain kind of reflection. A competent actor not only adapts to the rules but also reflects on their applicability and eventually changes the rules accordingly (WISE D1.1).

Competence levels refer to the levels of ability to practice competence. The different levels can be labelled as novice, advanced beginner, competent, proficient and expert, as defined in case III (WISE D1.1). Core competences can be seen as a set of skills, experience and attributes recognised by an organisation as critical to success in KM – for example information literacy, sharing culture or the like (see CWA 2004). According to Prahalad & Hamel, core competence refers to the company’s distinctive (technological) expertise and skill set at the leading‐edge of its industry or field that enable it to provide benefits to customers that rivals can not match. Core‐competence is at the heart of an enterprise's ability to compete. Core competences are important for competitive advantage, and they create added value that the competitors can not duplicate (Prahalad & Hamel). It could be questioned how the systems corresponds to real expertise. Dreyfus argues that skill based upon rules belong only to novices and advanced beginners, while true experts acts by intuitive intelligence (Dreyfus 1986). However, the action‐guided rules and information in the systems are combined with intuition in knowformation.

This research shows that an extended enterprise’s core competences are not technological skills. The real core is organising the work, while technology offers a self‐sustaining backbone for information storing, processing and communication. In the comparison of cases II and III there were a contradiction in terms of the publicity of competencies. The best solution is to have project level competencies publicly available, but the competency of a knowledge worker is stored only in HR databases – and the employees’ work results are indirect indication of their real competencies.

Nowadays, the trend in the Web is towards wider collaboration. Traditionally, standards were individuals, books and journals, data, information and documents in shared databases, telephones, taxonomies, directories, classrooms, conferences and e‐mails. In the first generation of Web we moved slightly more toward the public world, including communities, web sites and portals, online presence and instant messaging, teamrooms, forums, web conferencing and expertise profiles. During Web 2.0 other tools could become common. These include platform across enterprise and beyond, online reputation, RSS feeds and ‘social networking’ with blogs, wikis, social bookmarks, tagging, voting, podcasts and videocasts. To get an added value on top of Web 2.0 social networking portals, such as LinkedIn, Xing, Facebook, MySpace, Bebo, Hi5, the real knowledge related services must remain in the background. As new sources and methods to ‘share knowledge’ appear, none of the old disappear.

For an extended enterprise, one advantage of information systems is globalisation. The complexity and volume of global co‐operation today is unprecedented; the number of global actors, products, and distribution channels is much greater than ever before. The speeding up of all elements of global trade — mainly because of information technology — and the decline of centralised economies



227

have created an almost frenetic atmosphere within organisations, which feel compelled to bring new products and services to wider markets ever more quickly. This combination of global reach and speed compels organisations to ask themselves, ‘What do we know, who knows it, and what do we not know that we should know?’

The main answer is that to begin with, we know objective definitions from the everyday world. Standards are therefore the most useful practices for collaboration between corporations. Another point in the extended enterprise is related to ‘knowledge sharing’ and the value of information. The previous sentence gives the crucial distinction. Knowledge is considered in the sharing context as something that consists of personal opinions, beliefs, and the like. Therefore this kind of sharing does not deflate its value. Information is not shared as openly as this kind of bulk ‘knowledge’. Other future issues (not discussed here) include the question: if some valuable information is shared, how will we solve the problems related copyright laws?

2.2.14 The SECI model and ba concept are valid for one kind of knowledge creation

Nonaka’s knowledge creation makes the same distinction as in knowledge sharing: knowledge is not information, and it is easier to create and share something value‐laden than a ‘justified truth’.

The SECI process is also relevant when concepts have to be actualised into action and practice. This research, however, approaches the SECI process on an organisational level. If this model is used at all, it could be presented as in Figure 51, which utilises all meaningful concepts from literally ten years of research – from OM conference (the 13th of December, 1998) to Trends in Global Knowledge Management presentation by Hackney (the 13th of Noember, 2008).

Figure 51. Sketch for an expanded SECI model 18

18 The researcher can try to find and compare best candidates with new models, and often existing models are extended (March & Smith 1995). This sketching is one result from this proportion. Models, systems, practices, etc. have been evaluated constantly, and only fraction of these results is written here.



228

First of all, the SECI model is based inconsistently on both constructive and positivistic views of knowledge as it assumes that ‘explicit knowledge’ is same for all, and only tacit knowledge is subjective. The knowformation concept is based on the fact that information — as well as this wide interpretation on knowledge — includes personal and cultural connotations (based on beliefs, viewpoints, obligations, attitudes and intentions) no matter if it is in computerised form or not. There is no clear‐cut limit on tacit and explicit knowledge. Also, the truth value is fuzzy. Data can be binary coded, but even that meaning is interpreted subjectively. Everything is relative and tied to context or situation‐specific thinking.

Instead of dividing knowledge into tacit and explicit as in the SECI Model, we should use the constructivist approach of dividing knowledge first into objective and subjective knowledge, and then to trisection of tacit, implicit, explicit as follows:

Table 27. An alternative to tacit vs. explicit dichotomy

Subjective, constructed knowledge Explicit, objective data and information Tacit knowledge Implicit, subjective knowformation

We can see from the table that only objective information can be explicit, not as in the SECI model, where only the cycle or process is constructivist and the structure is positivist. Explicit knowledge is originally objective and positivist, while tacit knowledge is subjective and constructivist: According to the concept of knowformation, however, only constructivist objective data and epistemic knowledge as a subset of information presentation could be somewhat explicit. This seems to agree with reality based on this research.

Organisational knowformation is based on the idea that systems include more information content than any individual in the organisation has. The meaning of knowformation is based on the communication, as human beings interpret presentations. The position of each information item on the axes between process and communication is subjective. The crucial thing is what the organisation remembers – first in the short term memory and later in a more explicit form. Knowformation is a fairly stable grey area including data from the systems as well as interpretations based on that data.

Innovation and research are related mainly to design in product creation – this applied design is referred to as ‘knowledge creation’. This knowledge creation (innovation and research) is dependent on a continuous process of dynamic interactions between tacit and explicit knowledge, with the knowformation layer coming before communication. Such dynamic interactions are shaped by shifts between different modes of knowledge conversion, instead of just one mode of interaction.

Information accessing and capturing are different from communication, because they are done in any case, and not necessarily in simultaneous interaction. These, together with information search and retrieval give input for planning and forecasting in research and innovation. Problem solving and explicit data‐based decisions are easier when the right facts are obtained.

Nonaka, Toyama and Konno define four different types of ba, each of which is classified according to the type of interaction and media used in such interactions.

• Originating ba is defined by individual face‐to‐face interactions, as this is the only way to share certain types of tacit knowledge.

• Dialoguing ba is defined by collective face‐to‐face interactions. Selecting the right mix of specialists is the key to management of knowledge creation in dialoguing ba successfully.

• Systemising ba is defined by collective and virtual interactions. This type is meant to offer a context for the combination of existing explicit knowledge. Information technology, such as on‐line networks, groupware, documentation and data bases, offer a virtual collaborative environment for this type of ba.

• Exercising ba is defined by individual and virtual interactions and offers a context for internalisation, for instance through documentation and simulations.

Systems

practice;communknowledspecific chand, thinformattechniqu

dealing mcodified has showcommunmatter w

A 2.3

knowledimportancontent.and techmanagem

19 The simdimensioreality as

between inf

It has ; but not quinities, as it isdge is develoconcepts ince ‘story tellition is scientue of storyte

The pomostly with informationwn that thosnication is ouwhat the kno

memory ma

The cadge is a cruciant as the pro. Communichnologies. Thment and sy

mplified modeonal (dependin shown in Figu

formation an

been claimete. Ba is alsos often on theped. In this rclude somethng’ traditiontific. The simlling is learntositivist, cogdata. Such ITn (Clark and Se systems arut of that scoowformation

anagement m

ase studies sal componenocesses and tation makeshe following stem infrast

el is a result frong on classes aures 52 to 55,

nd knowledg

ed that ba sho about create boundarieresearch, thehing more th of communplified format. gnitive modeT tools can bStaunton 198re especially ope. Systemsn constructed

model of a a

showed that nt in the practhe knowleds informationextended orructure man

om all the casand relations,, and still keep

ge

229

hows some siting a shareds between dere appears han the subtyities of practat is suitable

l of the SECI e useful for 89; Swan andadaptable as are for stord on that dat

an extended

transformatctice of all oge is construn and knowlerganisationalnagement.19

ses and the w, cf. Bunge 19p it reasonabl

Part III – M

imilarities tod context forifferent comno evidenceypes of westtice could bee for handling

and ba are acodification,d Newell 200s an organisring the datata might be.

enterprise

tion of data, rganisation lucted partialedge organisl memory ma

hole research67), there aree.

Modelling the

o the conceptr participantsmmunities tha whatsoeverern commune called into g complex si

adaptable fo, storage, ret00; Hayes 20ational mem and ‘explicit

information evels. Commly on it and psational. Thisanagement m

h. Even thoughe means to ad

e results

t of communs from multiat interestingr that the culnication. On question whtuations, if t

or informatiotrieval and tr001) but this mory, and t’ informatio

and (fractiomunication ispartially on ts is true also model cover

h it could be dapt it to refle

nities of ple g new lture‐the other hen the he detailed

on systems ransfer of research

on, no

ns of) s as the memory for systems s content

drawn multi‐ct complex



230

Figure 52. Memory management model of an extended enterprise (M3.exe)

The key components of the model are: all information systems (platforms and

applications), knowledge of personnel as well as content of the systems, organisational and inter‐organisational processes, and communication. This model can occur in different domains – the extended organisational memory defines the subject – e.g. what process or systems we are planning.

The subjectivity of the memory content rises from objective process‐based data towards subjective fuzzy communication. The organisational memory is extended in the sense of dynamic information in the short term memory shared by collaborators. The process manages epistemic data, while communication is dependent on trust between interlocutors. The process includes objective information and data that are shared with external collaborators, while the communication deals with unmanageable subjective internal knowledge creation inside corporate communities.

The border between communication and content acts as a filter: to select the parts of the dynamically created documentation that will become the official content of the systems. Otherwise the amount of content would grow exponentially. The presentation layer will help in the other direction: presentation below communication includes visualisation, to define what content will be published and thus officially communicated. In memory concepts: communication is normally more episodic, while content is more semantic.

The information systems model and cognitive paradigm used to be tied to repetitive action, replication, and standardisation of information (Swan and Newell 2000). This new model will separate processes and systems as totally independent entities. The obstacle to change has been removed, and new knowledge, which will change static routines, is only beneficial. As the cases of this research have shown, rapid change and technological discontinuity are in contrast with each other. The new model makes the elements independent, and a dynamic balance – as the fine line between exploration and exploitation proposed by the expanded SECI model and ba – can be achieved.



231

To summarise, the domains of the Memory Management Model are:

• Content domain – Methods and tools for modelling knowledge, information and data requirements, and the available information in ICT systems, and changes to the data offering. Also, evaluating organisational expertise based on who is editing content to, or utilising content from the systems. Data, information, knowledge and meta‐data from the extended organisational point of view. Document entities like deliverables, service documents, product specifications, status reports, production schedules, issues lists, standards, meeting minutes, tips, improvements efforts, instruction manuals, lessons learnt in active short term memory updates to the communication domain through the knowformation layer. Explicit content is the basis for the organisational structure, which manages processes. Relates to the system domain by information containers.

• System domain – what systems are to be used? Methods for describing what kind of platform and application landscape the organisation and its collaborators have, and their roles and dependencies. In addition to computer (information) systems and applications, this also includes the technical infrastructure and platforms. When independent from content and processes, the system architecture will be easy to change, as required in the case studies. System domain links to the communication domain through communicating technology and processing is in accordance with inter‐organisational processes. The maturity of knowformation containers between systems and content varies, as described in Figure 54.

• Communication domain – how the information is communicated, and how the knowformation is supported before converting to explicit data (or tacit knowledge and archives). The information presentation layer is near to the system domain. It includes individual level communication between engineers, product managers, etc. but also team level communication between product development teams, product launch teams, management teams etc. and organisational communication between external process consultants, standardisation or quality consultants, marketing to customers, etc. This domain is even more important for new knowledge creation: it contains communities of knowledge workers, where interactions and sense‐making, modes of enquiry between people, as well as innovations occur through research.

• Process domain – what are the extended organisational processes or procedures? Methods and tools for modelling the activities and roles. The process domain is linked to communication through learning processes, that is, mainly the parts of knowledge than can be somewhat managed on a meta‐level. Other information and data processes are easier to manage. From a philosophical point of view, the objective elements of knowledge have an impact on the process. The mapping and characterisation of organisational memory – content and systems – should be done here. Awareness of free form expert communities and collaborative intelligence (in the communication domain) could be noted in inter‐organisational collaboration. The mnemonic functions of organisations appear here as the process phases.

The extended organisational content domain should first be analysed by further

modelling each content source in such a way that it is possible to measure its usefulness, its weaknesses and its appropriateness within the organisation and with the partners. Information analysis is a necessary step for managing knowledge. Modelling and acquisition techniques can be used in knowledge analysis.

On implementing meta‐data, the major result is that the meta‐data repository does not have to be an integrated solution – it must be an easily integrateable solution. The meta‐content of service‐oriented architecture should be centralised, while the actual content of the systems should be totally independent and as distributable as needed. One crucial reason why KM systems have failed is that people are managing documents and systems and not the content, which is independent from the files.

After a data, information and knowledge content analysis, organisation will be able to make future plans. The organisation will now be able to develop a multi‐year knowformation plan that



232

defines how it will develop its knowledge resources, train its human agents and develop knowledge‐related information systems to support human resources.

The systems between information and knowledge manage content, but when content is formulating into human knowledge, it becomes knowformation. One crucial feature of knowformation is that knowledge workers do not question it. It is shared, organisational information, as well as part of individual knowledge. Individuals might not think of it is as the ‘final truth’, but organisational knowledge work is based on it.

The management of knowformation should be based on the actual substance – what it means to the users; otherwise the process would be a sort of file and unstructured data management, like technology‐based KM has been so far. On the other hand, following the constructivist line, all knowledge is contextualised by its historical and cultural nature (Agger 1991). This applies also to the content in information systems and knowformation.

Figure 53. Knowformation and the systems between information and knowledge

As in the picture above, the information species can be included in knowledge, because of constructivist scholarship, and because the data, information and knowledge triangles could be organisational or individual, and include the knowformation concept, which extends the information in systems to human knowledge (Figure 46). This is a result based on both KM literature and the case studies concerning systems, even though the generic division between knowledge and information should be preferably as in Figure 48. The systems can also be other than the categories listed, but these system categories are normal pillars of contemporary knowledge work.

The case studies showed that data is considered so explicit that it is not constructed as knowledge even though it is interpreted. However, the decision‐making and other further processes based on data are very political in nature. Data management affects all layers of data, information and knowledge. One reason for the term knowformation is that categorisation of information or knowledge categorisation is rarely done in the normal work environment. The truth values in communication or information presentation are not considered important if the knowledge is ultimately subjective. Human knowledge is subjectively determined by many factors, including pedagogical, socio‐economic, cultural and psychological issues. Language and context also affect to the construction of knowledge. Most of these factors operate unconsciously (Kuhn 1970), even though the value of information is a crucial element, and knowformation is floating also on the level of trust (its position between presentation and



233

communication is variable, and the truth value is not binary). The long‐rooted species of information and knowledge were already presented in Part I Chapter 2, Information and knowledge species, it is time to categorise the systems.

The systems between information and knowledge include a variety of data, information and knowledge management systems. In practice, most generic categories are document management, content management, record management, and archiving. Record management is a result or Cases I and II, while archiving is a more generic issue. The knowledge management systems also include access systems: portals or portlets; collaboration systems: team tools or meeting tools; and communication and presentation media: like e‐mail or some systems used in ‘social networking’. The systems are independent from content itself, but handle ‘containers’, such as those in Figure 54. The arrows on the right hand side present a recursive loop from containers back to knowformation.

Figure 54. Maturity of containers and knowformation related concepts

These categories also relate to the knowformation in individual and organisational

triangles (Figure 46), because data is the basis for everything, but it is also a core of knowledge. Data as the input or content is mature, while outside systems it is not; in human cognition the meaning of data is opposite to organisational sharing.

Record management is between data and document management. In the early literature on organisational memory, document management was most often associated with a set of documents (artefacts) (see e.g. Burstein et al. 1998). In the early part of this research, more forward‐looking, active design documents were investigated by collaborators of case III (see Boy & Dürstewitz 1998, 1999).

Although document management has generally been and will be the core of organisational memory, in the future the context, which gives the document meaning, will be more important (see e.g. Conklin 1999, p. 9) and nowadays the document management concept has mainly been replaced by content management.

Some of the systems’ content will be archived, while the living content is on the nearest peak under knowformation – in the individual construction as knowformation, and in the organisational memory as content. The recursion from archives, as well as from every container category, makes the containers part of knowformation.

Communication is needed between persons but it can be computerised and also independent from time and direction – i.e. it can be something other than interactive communication. The presentation is related to, but not a part of communication. Communication, like discussion of communities, is self‐managed action that exists across the extended enterprise and requires time and



234

technology. Even though the communication is supported by systems between information and knowledge, it is partially based on interpersonal relationships, respect and trust, which are critical.

The new content processes cycle consists of six mnemonic functions deployed to the extended organisational memory. It includes:

1. Acquisition • Creation, Learning, Hiring, Research, Innovation • Knowledge and information acquisition and input, transfer of knowledge

to the system 2. Retention

• Capturing, Storing • Memory contents retention; retention and structuring of organisational

knowledge 3. Search and Retrieval

• Memory contents search and retrieval, including indexing and classification • Knowformation recursively constructed on selected information,

application systems and containers. 4. Maintenance

• Reconstruction, Reuse • Memory contents maintenance, long term storage and integration of new

knowledge 5. Application

• Processing, dissemination and visualisation towards publishing. • Presentation of existing information in containers.

6. Archiving / Deletion • Intentional version management to the end of the data lifecycle • Archiving and deletion decision as the part of lifecycle

In the model, these should be considered as functions, instead of processes, which is the

wider scope. An extended enterprise should also have active co‐operation between the internal processes offered by solutions and services. This non‐measurable co‐operation is crucial for enterprise‐level benefits. The traditional knowledge processes, such as generation, representation, storage, transfer and transformation (Hedlund 1994) can be adjusted to extended organisational memory processes. As the retrieval is a crucial part of memorising, there could be background phases before search and retrieval, for instance, indexing and classification.

The seven memory bins for extended organisation are now including information systems and also external archives:

1. information systems supported by communication technology, 2. individuals, 3. culture, 4. transformations that occur in the organisation, 5. organisational structures, 6. actual physical structure or workplace ecology of an organisation and 7. internal and external archives. The role of information systems has become dominant after the communication

technology became part of the infrastructure. As mentioned in Part I subsection 2.4, Practical knowledge can not be passed by, memories are mainly semantic and episodic. Examples from the content of semantic

Systems

help orgon the hresult orFor exampreferaborganisafrom a uthe revegovernadecided

Ad2.4

in theorysystems knowledof any knknowled

the layerand taxonetworkarchive; Howevefor storachosen oand humretrieval

involves organisepossess operatioinformatjustified construc

not usedquote thknowledrecursive

between inf

memoare damy firas motalkingOrganindividfor co

The ovanisations muman actorsrganisationalmple, reportsble situation ational memoubiquitous sherse K‐I‐D pyrnce is only vsomewhere

dapting the

What y and practicthat deal wi

dge and infornowledge obdge is suppor

The trrs rise from konomy generk; document and e‐learnir, the infrastage, retrievalon the basis man‐resourcel and mainte

The ina person wh

e informationinformation,onal routinestion and knotrue belief ccted result is

The syd as an epimnhe heading bdge is too widely. For exam

formation an

ory are: ‘I waata on explicrst day at schotoric memog or knittingnisational meduals, but thnstructed kn

verall intentmake optimas’ subjective l long term ms that are preis one whereory in their ohort term meramid) turn ioting indiffe else.

model to sy

can systemsce. Thereforeith objective rmation systebjects that harted by commraditional knknowledge mrators to learmanagemening services, tructure for kl and dissemof the abilitye assets mainenance. nterpretationho knows. Thn efficiently, , but to use ts, skills and kowledge, howcan be storeds most often ystems couldnemonic layey Brown & Ddely interpremple, with co

nd knowledg

as born 8. Mait memorieshool I wasn’t ry and primi. Priming is lemories and he overall mixnowformatio

ion is to prodl decisions aworldview k

memory dataepared for ste decision mown opinionsemory. This wnto expert derently for de

stems betwe

s do with knoe, the strategknowledge‐ems is the coave been stomunication aowledge mametrics, awarning concepnt services, ssuch as courknowformatination of my to support ntenance, an

n of content he question iit does not bthat informaknow‐how, awever, is notd in a systemknowformatd be used forer, the questDuguid 2000:eted). Knowlommunicatio

ge

235

ay 1972’, ‘Pa. Episodic menervous at ang. Motoric earning frominformationxture of diffeon.

duce results nd to apply tknowledge tha from the syteering meet

makers includs, intuitions awill make thedecisions throecision by for

een informa

owledge? Kngy should be‐related dataontent itself ored in a comand collaboranagement ureness, objepts, such as cuch as librarrse creation,ion could be

memory elemorganisationnd their abilit

is usually suis belief. Evebelieve anythation. On there fulfilled in dualistic. Th

m, but the inttion. r capturing ktion is ‘how t: 2, even thoedge can be on technolog

Part III – M

aris is the capemories are all’. But thermemories am unconsciou systems areerent memo

by creating their know‐hhat is constrystems has otings are pole objectivelyand beliefs, e limited humough knowformally accep

ation and kno

nowledge mae to maximisea and docum(not conten

mputerised (dation technotilises the coct sharing, ccommunity sry services, d, course delive a combinatments. Infrastnal integratioty to suppor

bjective. Knoen though a Khing itself. Ae organisation the actual phe result is thterpretation

knowledge, bto capture knugh it is an ecaptured am

gy, it is fairly

Modelling the

pital of Francconsidered e are also imre motions leus sensory see mixed withries and curr

new high‐quhow in practiucted in a cuonly secondalitically presuy explicit datand can retrman memoryormation. Otpting the sug

owledge

anagement se the benefitents. Betwet managemedigitised) forology. ollaborative iollaborationspace, solutiodocument lifevery and curion of humatructural arraon, adaptatiot memory ac

owing (not kKM system cAnd ultimatelnal level thisprocess. Thehat knowledgrequires an

but if interactnowledge wiexample frommong other ieasy to crea

e results

ce’, ‘9 * 7 = 6explicit as w

mplicit memoearnt, like wensual data. priming onlrent data is o

uality informice. Decisionurrent situatiry effect on upposing a dta from the erieve the infoy data capabtherwise thegestions tha

systems are nts from inforen the comment systems),rm. Subjectiv

infrastructur and converson space or eecycle, publiriculum mann and techniangements son, goal achicquisition, re

nowledge) aan store, sorly, the aim iss means that limit betwege as presenactant and t

tive communithout killingm KM literatnformation –ate a record o

63’. These ell: ‘During ories, such walking,

y through one reason

ation to s are based ion. As a decisions. decision. A extended ormation bilities (as in e t are really

not similar rmation municated , consisting ve

re, where sation tools, expert shing or nagement. ical means should be evement etention,

always rt, and s not just to t en ntation of the

nication is g it’ (to ure where – of the path



236

taken to reach key decisions in meetings, rather than recording meeting ‘outcomes’ or final protocol. If the communication deals with other than epistemic knowledge, is a less relevant question.

Figure 55. Lifecycle of knowformation in M3.exe

Figure 55 describes recursive lifecycle of knowformation. This includes the same entities as M3.exe. Knowformation formulates when information moves from one species to another by learning, knowing and informing (or whatever verbs are used) in the context of organisational memory. Organisational semantic and episodic explicit memories are mixed with implicit individual memories and thus interpreted as knowformation. Pragmatic and communication species are closer to episodic memories, while epistemic and presentation information species are closer to semantic memories. These are constructed in the context of individual implicit memories into knowformation.

Knowformation also mutates somewhat every time knowledge is interpreted, and moves from one entity to another. For example, pragmatic knowledge‐related information is communicated, but it is mixed with expressive information. Systems process the information, some of it in a positivistic manner, some of it constructed by the users. Processing is based on input, produces output, and ultimately relates the whole lifecycle. The time symbol reflects measurable mutability or change in time, based on activities, procedures or other smaller functions. Processes as well as the whole lifecycle should have an end. Processing covers symbolic interpretation or model of knowledge, together with epistemic entities with little context. A subjective vs. objective interpretation is relative, and varies every time knowformation changes from one species to another. Processing itself could be quite epistemic, but the man‐machine‐interface results in autopoiesis or other kinds of construction. The objective vs. subjective measurement is related also to semantic (objective) and episodic (subjective) memories on an organisational level, and mixed with knowformation while constructed in the context of implicit memories.

Dynamic communication requires freedom, and communication is not just the delivery of information. The value of capturing is also a question mark; the value of information is inflated, when



237

the information is shared, and clear ownership is vanishes. Inside the corporation and between trusted collaborators certain information sharing is valuable, but if the consequence is redundancy of employees, the benefits and re‐employment must be planned beforehand. Poor staff retention causes a loss of human knowledge and personal skills.

The real benefit comes from creation of valuable information. Knowformation includes both information based on data, documents and other vast amounts of systems content, and the mainly implicit organisational thinking affected by memory, intuitions and other fuzzy factors. Only a fraction of the knowformation (presented in Figure 46) of individual human beings is objective. All the rest is constructed, subjective knowledge. On the organisational level, more data is commonly shared and independent from individuals. This transformation from individual to organisational goes through communication and refinement from knowledge towards data.

Information systems can support communities of practice on an organisational and extended organisational level. Most companies have used information systems to support individual work, and the individual should have sorted the information, decided what is important, clean and organise it. But on the organisational level knowledge refinement could be done more efficiently, and on an extended organisational level it is the only possibility.

In some instances of communication systems, community coordinators or core group members can judge what is important and useful and enrich information by summarising, combining, contrasting and integrating it. That is more or less like a work of an editor or publisher. However, this research established that the power is on an organisational level. Coordinators should simply give their time to the community and not try to obtain power for themselves. Quite often technical professionals and engineers consider this a glorified librarian role of organising and distributing information.

The value of engineering and other professional practices is derived from turning information by action into solutions. It is the art of finding a purpose for information based on the question: what is not yet known? It does not matter whether this is knowledge creation or discovery. The purpose is to determine what information to focus on and what to remember.

The engineering, science, administration and other practical professions studied showed that experience is needed to construct solutions to current situations. There is seldom a solution from history to copy and paste onto the current problem. The best practices and lessons learnt only give examples, and should be adapted to current practice.

Epistemic knowledge can be a presentation of any justified true belief. However, knowformation is constructed in a particular situation based on data, information and knowledge. On an individual level, knowformation could require thinking and experience. It is generated from the experience we are making sense of and testing against other’s experience. It is experience that is informed by learned theory and facts in relationship to a field or discipline. Memory is always present.

The challenge is communication of knowformation. The technology branch used to be (and is still) based on positivistic worldview. Many experts were introverted technology believers, who required one truth, which is why the SECI model was well accepted – and the diagrams were easily interpreted according to one’s own preferences. For them it is impossible to accept the worldview, where nearly everything is subjective and open to interpretation. Therefore ‘knowledge’ is not shared and the systems are not living or dynamic, but technology for data management. In that sense, knowledge creation and discovery are different – creation results in whatsoever new, while discovery is more like trying to reach and tell the objective truth. The context of SECI is creation, although it is partly positivistic.

Systems between information and knowledge should help to produce and store information, to capture data at the source, transmit and analyse the data, but most of all, to communicate the new knowformation based on the data and information, to those who can transform the knowledge into decisions and action. This communication and collaboration happen on both the micro and macro levels. On a micro level an employee’s significance to colleagues might be evaluated, while on macro level the value of the collaborating company might be measured.

Systems

3 STR

implied t1991). Rregardinsignifica

managemmain reamanage Informatrealised the com2001; ZainteractiNewell 2

extendeworkers services organisabased onan exten

(describe

• Cc

•

• A

• c

involves knowledhorizont

Co3.1

epistembased onplain stoto descrtacit knogenerati

communbecome

between inf

RATEGY BASE

Traditto strategy fesources areng the transfence (Kanter

Despitment projectasons for theknowledge tion systemsby appropriaplexity of maack 1999). Thion between2000). Extend

Knowd enterprisewith other con top of thational short n informationded enterpr

In added in the pre

Connecting pcommunicatMaking the iincluding an Analysing orknowledge gImplementincontent in a

In add

moving fromdge‐driven firtal communi

ommunicatio

Thereic knowledgen communicoring and retibe memory owledge. Theon, as dynam

In termnication techpossible to d

formation an

ED ON THE M

tionally in KMormulation ie idiosyncrater and replic1999). te strategic sts have failee failure of Krather than ds for manageate strategieanaging orgahe culture sh those who nded enterpriledge can no view, wherecommunitiese various daterm memon systems) crise, but onedition to the evious chapt

people with tion technoloimplicit and organisation

rganisationalgaps in the mng systems bnormalised

dition to thesm the traditiorm emphasiscation.

on creates k

is no such ae, the systemation betweerieving of docontents ane knowformamic knowledms of an extehnologies hasdistribute th

nd knowledg

MODEL: LON

M literature, in such a waytic and scarcecation of kno

significance, d to meet thM projects hdocuments oement of knoes, policies, panisational kould encourneed the newise depends ot be managee systems wos. The knowleta and inform

ory, in whichcan be used a solution is fstatements er), the exte

other knowlogy that can explicit datanal short ter knowforma

memory manbetween inforare number

se, the organonal ‘stable’ sing the valu

knowformati

thing as a coms mainly deen people. Tocuments or nd means, noation concepge is construended enters on its part e work not o

ge

238

NG TERM SYS

the resourcey that resoure – difficult towledge, thu

approximateheir objectivehave been thor systems (sowledge are procedures, pnowledge (Grage the orgaw knowledgeeven more oed with any ork in paralleedge workermation manaparts of conas part of thefor a third pafor memory ended enterp

edgeable peenhance the, informationm memory ption and skilagement moormation andr of informat

nisation shouhierarchicale of knowled

ion for syste

omputerisedeliver other inThat can be sother artefao matter whapt and commucted and norprise strategcontributed only between

Part III – M

STEMS AND

e‐based theorces are seento transfer os making kno

ely seventy pes, and over he lack of strasee e.g. Ruggimportant eprocesses, stGarvin 1993; anisation to ce and those won the culturcurrent inforel with otherr needs commagement sysstructed knoe communicarty to keep tmanagemenprise knowled

eople in the weir work n and knowlpresentationlls constantlyodel. d knowledge tion systems

uld adopt a t industrial odge, concurr

ems

d knowledgenformation,synchronous acts. The knoat the positiomunication arot explicit. gy, the develto the creatn different o

Modelling the

FURTHER RE

ory of compen as core comr replicate. Towledge ma

per cent of aa period of tategic prioritgles 1998, Mnablers, but tructure andPurvis, Samchange its wwho can prore. rmation systr systems, anmunication astems. This reowformationation. This bthe distributnt model fordge manage

working envi

edge availab layer y and interac

to enhance s

transformatiorganisation trent working

e system. Desand the widor asynchro

owformationon between re crucial in t

lopment of inion of new fooffices but als

e results

ESULTS

etitive advanmpetencies (There are banagement of

all knowledgeten years twoties and the alhotra 2005their value c cultures thabamurthy &

ways and encoovide it (Swan

tem alone. Wnd ‘tribes’ of and presentaequires an n (memorabiecomes a chted systems a an extendedment strateg

ironment, an

ble anywhere

ctively, to ide

organisation

on process, wto a networkg practices an

spite the stoer knowledgonous, but it concept muexplicit inforthe phase of

nformation aorms of worso between

ntage has (Grant rriers f strategic

e o of the need to 5). can only be at recognise Zmud ourage n and

We need an knowledge ation

lia on facts hallenge in available. d enterprise gy involves:

nd with the

e, anytime,

entify the

nal memory

which k and nd

red ge concept isis not just ust be used rmation and knowledge

and king. It has different

s

Systems

companorganisaor virtuaManheimcollabor

activitiesthe interknowledand extr

commenprocedulife, but insights

content needed tknowled

part of tnew meexisting another organisajust in ti

organisaculture, for the inof clear t

driven ainfluencewhich emknowled

Ta3.2

should bthe focabe part ohave all

enterpri 1. Extend

• To

• T• A

between inf

ies. People mational basis.al work requim 2000). Herators will ap

The ems, while at thrnet‐based mdge betweenranet on a m

Accesnting and conres, analysescurrent systethat are can

Large alone can noto constructdge by chang

Thereo the ‘tribesmber of the corporate knand it flows ation is too sme, within e

Learnations, the fovalues and bndividual, actargets, com

The fopproach anded by Nonakmphasises crdge aspects m

actical applic

The mbe stressed thl organisatioof ‘normal’ mbeen respon

Whense, the critic

ded OrganisaTrust and staobjectives anThe core of uArchives and

formation an

may belong t This createsires specific tre the commpear, while nmergence ofhe same timemobile techn co‐operatinobile infrasts to documenversation as, lessons leaems can alsonot exist wielectronic liot be spread organisationging the orga is no knowle’ of knowforcommunity nowledge. Inthrough prolow. The curexpanding neing is respecocal organisabehaviours. Tchievement ammon modulaocal compand deep integrka’s thoughtsreativity, innmentioned.

cation of the

memory manhat knowledon. From the managementnsible for man implemential factors are

ational Memability requirnd strategy bunified strated tacit knowl

nd knowledg

to several teas complexitytechnologies

munities are cnetworks thaf these tribese the virtual ologies mayng organisatioructure. ents should bre as trustwoarnt, maps, Co include intethout its conbraries of be knowledge,nal knowledgnisational stedge in the wrmation, or awill give his n organisatioofessional corrent requireetworks of pected in the fotion’s KM phThe leadershand continuoar processesy has unofficration rathers on the knowovation and

e memory m

agement moge managemoutset, the ct practice. Linanaging knowing the meme:

mory and conres a commobetween theegy of extenledge as a so

ge

239

ams or work in the work s providing sconsidered aat are open es, or any virtway of work offer one soons or collea

be ubiquitousorthy as fantCAD or illustrerpretationsntext. est practices , no matter wge. The ‘contructure reguwide sense wany communor her fracti

onal changes,ommunities. ment is self‐eople. ocal companyhilosophy hahip‐based corous learnings and commocially adopter than on advwledge‐crearenewal. Ev

management

odel can be ument has notcompany hane managemwledge. mory manage

ntext on culture of e organisatioded enterprource for rese

Part III – M

k groups thatenvironmenupport for eas ‘tribes of kequally to alual organisaking makes Kolution for suagues (cf. Sch

s. Active doctasy. Informarations, and , half‐formed

and other kiwhat are thenecting’ hereularly. without contities. Organion of new kn, knowledge If changes ne‐organising p

y. As with ms been greatrporate valumeet the facon managemed a knowledvanced technting companven revised S

model

used as a frat been consids emphasise

ment, solutio

ement model

clearly definns ise are contrearch for lea

Modelling the

t more oftennt. This kind efficient comknowformatil could be clotions also de

KM more comupporting thhein 1999): t

cuments withation systemother informd judgement

inds of know systems. Coe also means

ext, either. Ksations are fnowledge, anis inherited eed to be repeer product

ost pioneeritly influencees of customct‐based ma

ment principledge strategy nology. It hany, i.e. creatiECI model co

mework for dered as a seed that manan manageme

l (M3.exe) in

ned, concret

racts with tharning organi

e results

are formed of complex, munication (on’. Tribes oosing. emands imprmplex than be communicthe basis for

hout updatinms can delivemation with ats, ideas and

wledge manaonnecting pes transferring

Knowledge isfull of knowlend learn a mfrom one gedesigned, thtion. Learning

ng knowledgd by its own mer satisfactinagement coes and tools.based on a hs been someng an organovered only

designing a separate initiaaging knowleent and each

n an extende

e and achiev

e collaboratisation

on an inter‐distributed (Watson‐of

roved KM efore. Here cation of r intranet

ng, r statistics, a long shelf other

gement eople is g deeper

s always a edge and a

mass of eneration to he g happens

ge‐driven corporate ion, respect omponents . human‐ewhat isation those

strategy. It ative within edge should h employee

d

vable

ors



240

• The extended organisational memory is itself a living organism surrounded by culture. Even though it does not need to be viewed through the actor‐network theory, it has a sociotechnical context that is affected by all actants dealing with the enterprise

2. Communication

• People are connected through communication with the supporting technologies, and communication can include knowformation from systems between information and knowledge

• Direct and frank communication from management. Systems supporting communication and collaboration, like message‐boards, on‐line meetings, common calendars, wikis and blogs, news, and instant messaging can be used, but the content matters

• Information receivers must be active ‘searchers’ and perceivers – i.e. with the mindset of researchers or students. Belief that sustainable innovation means investing in research and development

• A presentation layer between processed content and communication, supported by the system infrastructure. In the presentation context there could be visualisation where the selected editors with secure access could publish content, while the communication infrastructure supports everyone

• Personalisation, localisation and regionalisation for communication depending on expertise of the audience, languages for different communities, country‐specific culture and so on

3. Systems • Clearly organised database management systems are the core for all information. On the other

end, extended enterprise content management as shared resources and electronic asset management enable reuse

• Structured and modular documents as information management in the middle. Portal, portlet and browser based templating and editing control, e.g. DTDs and schemas with restriction of changes to gather new extended organisation‐wide consistency

• Fast, efficient visibility above the presentation layer to deliver knowformation as the content between information and knowledge. Presentation form of information with application integration, portlets, or other similar technology

• Interactive, dynamic, ubiquitous communication systems to generate discussions, tagging and other social methods to improve knowledge. Dynamic content environment also covers container related features, menu systems and link management

• Intuitive, simple tools with common portal and portlets on top of every technology. Caches and other performance enhancements with scalability geographically

• Active search technologies to get the relevant, up‐to‐date information for the correct people • Role and scope‐based user and group management with integration to existing systems.

Customisation providing options to user to enable custom experience and (pre)viewing the content in context on presentation layer

4. Content • The data, information and knowledge available to an enterprise dictate what the organisation is

able to do • Clearly defined responsibilities for content providers and publishers. There should be distributed

authorships, as the number of authors and deployment stakeholders – as well as geographical distribution – are increasing. Information sharing must be integrated into job responsibilities and processes. The knowledge of each individual is unique and the divergence should be respected and seen as a resource. The more innovative communication is inside communities – and that should be officially supported



241

• Tight integration on data and information on extended organisational catalogues, lists and other objective content. Content management should also include scheduling and lifecycle management with content tagging and meta‐data management. Meta‐data will lead to actions

• Learning material normalisation with other content – e.g. the same information can be used for manuals and tutoring

5. Process • If there are no other more detailed processes, then there are at least memory management

model process phases • There should be workflow in basic editing and approval processes, including reporting on the

actions with log file analysis functionality. However, this concerns the basic process‐based systems, not any upper level towards communication

• A notification availability and tracking of changes is needed in different phases of the process. This is also valid for the basic level processes

• Recognition and reward to support learning. A context‐sensitive feedback mechanism together with version control in the process of being developed

A knowledge strategy based on the memory management model begins with an

information and knowledge analysis. An organisational information analysis is the key component of the enterprise information architecture. It contains a representation of information entities, structures and relationships. It captures domain knowledge and states it precisely, for example by using UML class diagrams and textual descriptions. The content domain covers the information and knowledge analysis.

A tactical requirement for applying M3.exe to strategy concerns knowledge gaining. Rather than actually changing their processes, extended organisations could simply study their processes in order to learn. This does not mean that an organisation would not change their process if it made sense. What it implies is that it is not a matter of simply changing the process but rather the deeper understanding of how their organisation works that is profitable. The content can cover process recursively in the model. Thus enterprise architecture planning should start from content instead of processes.

The company’s knowledge strategy is not separate from its overall business strategy. It should be seen as a key element in ‘normal’ management practice. The responsibility of ‘managing knowledge’ in the KM literature sense belongs to line management and every employee.

Measuring the efficiency of knowledge distribution within the focal organisation was a key to creating the learning organisation. The knowledge efficiency index was defined as the organisational distribution of new or useful knowledge as a function of the time and effort which is required to distribute the knowledge to the right individuals or groups. The focal organisation discovered that from twenty to thirty per cent is a typical value for this knowledge efficiency index. The company’s goal arose to increase the index value to eighty per cent.

The focal organisation’s transformation from a traditional manufacturer into a world‐class high technology company bears witness to the benefits of managing knowledge. Recent upheavals in the global communications market have only reinforced the focal company’s belief that knowledge is the key global competitive differentiator. Continuous new product development based on turning tacit knowledge, know‐how and experiences into value‐adding corporate knowledge could have a memory management model at the heart of the organisational success. It requires communication that enables new knowledge in epistemic sense and also creative knowformation sharing with meaning.

The focal company encouraged the implementation of managing information in small steps, integrating initiatives with other organisational change activities and programs. The human aspects of managing knowledge were the biggest challenges. The focal organisation has continued to focus on change management in organisational culture, attitudes, working methods and processes.

Information sharing requires new corporate values and new forms of rewards and recognition. The biggest challenge is to understand at a much deeper level how an organisation manages human‐based individual and organisational tacit knowledge. For an extended enterprise, the

Systems

extendea part of

St3.3

on puttinexchangvalue of and the collabor

of the syknowledevolves f Systems

• C• C• S• S• •

platformrapid dewill evenubiquitoessentia

bookmasearch pcontent

differencexamplethe wron

knowleda tacticain the prsupport made avplatform

informat

between inf

d organisatiof the strategy

trategy and t

As a sng it into acteable and recontent is dexchange vaators and cu

The knystems, wherdge maturity from tacit in

between infCreate conteCreate moduStore, updatSupport the Publish the cDeliver the s

The mms, web‐basevelopment ontually work ous access. Thlly a recollec

Web‐brking and tagproblems. Rebetween inf

Howeces betweene, content beng persons. A

Whendge informatal result, but revious list. Sthe requiremvailable as inm implement

The fotion and kno

1. Syco

2. Thsy

formation an

onal memoryy of informat

tactics

ummary, hetion. Systemseliable conteependent onalue of the inustomers. nowledge mre the indepare on the sformation a

formation anent independular, compoute and managcontent lifeccontent in ansame conten

means to realed applicationof web‐basedvia browserhe pioneerinction of contbased architgging will dimcursive tags formation anever, obligation introvertedeing copied fAnd most of n the architecion and contbased on vaService‐orienments of prodependent sation. SOA wollowing kindowledge: ystem impleonsequencehe system mystem migh

nd knowledg

y managemetion systems

re is the stras between innt for the ben the rarity onformation. A

aturity betwendent contscale from prnd fuzzy idea

nd knowledgdent from thund documege the contecycle and wony media neet in different

lise the indepns and mobid technologis in any clienng enterpriseent and assoecture offersminish the loare includednd knowledgons for sociad and extroverom private the social necture is plantent) insteadluable contented softwarocesses and uservices thatwill also help ds of conting

ementationes may not be t be incong

ge

242

ent model is s for knowled

ategy in lists nformation aenefit of the of the informAll of these v

ween processent is managrocesses to cas to explicit

ge should hee managements when it ient in a reliaborkflow fromeded t languages

pendency beile clients. Knes. This will nts, includinge will define ociated infrass new tools foads of e‐mad in knowfore. al networkingerted professpresentationetworking coned, it startsd of processeent, the valuee architectuusers. Resou can be accestrengthen

gencies may a

n may not be

successfullygruent with

Part III – M

the key factdge manage

related to prand knowledgextended en

mation, its novalues are in

s and commuged and delivcommunicatit and structu

lp to: ent system ais meaningfuble way m planning an

and presenta

etween contnowledge macontinue to g mobile devan omnipresstructure anfor communils. Meta‐knmation, beca

g and hence sionals. The ns to blogs wontent is far s from corpoes. Service Oe of which isre defines thrces on a neessed withouthe indepenaffect implem

e initiated,

y implementhe organis

Modelling the

or for succesment

revious chapge should ofnterprise andovelty to the the context

unication defvered. Data, ion. The lifecred data.

and authorinul

nd creation t

ation format

ent and systanagement hdevelop so tvices. Mobilitsent knowledd systems. ication. Sociowledge in tause they co

publicity areworst side ewith the resufrom epistem

orate knowforiented Archs measured bhe use of lootwork in an Sut knowledgedence of conmentation o

resulting in

nted; users sation

e results

ssful coopera

pters, and tacffer valuable,d its sharehoreader; its aof the corpo

fines the clasinformationcycle of know

g tools

o archiving a

ts when need

ems are XMhas been fuethat all user ty is a big addge web that

al networkintags is also honnect peopl

e also generaeffects can inult that creditmic knowledormation (inchitecture (SOby the factorsely coupledSOA environe of their undntent and syf systems be

n political

might resist

ation, and is

ctical notes , applicable, olders. The pplicability oration, its

ssification n and wformation

and deletion

ded

L, universal lled by the interfaces vantage for t is

ng, elping le through

ating clude, for t is given to dge. cluding OA) could be s presented d services to nment are derlying stems. etween

t, or the



243

3. The system contents may not be maintained and used, because of lack of motivation to contribute

4. The system may not be used properly, if improper routines are invoked, or entrenchment is autocratic

5. the system may not fulfil its potential, because of the culture, values and management support

In addition to being independent from the media, the content of organisational memory

could be somewhat independent of languages. Modular documents can present the same content translated. This agrees with the philosophy of presentation, while cultural semantics are on another level of communication.

Systems between information and knowledge should offer a systematic method for storing, locating, and keeping track of content. Acquisition, retention, maintenance, search/retrieval, application, archiving and deletion are the mnemonic functions of processes in organisational memory to which all information containers as well as content belong.

The system classes are: data management, record management, content and document management, archiving together with infrastructure supporting communication, presentation and search of knowformation. Content management and document management should be seen as different levels: document management is included in content management, but only content management includes modular documents, independent from systems and media.

The key characteristics of the systems are: the ability to manage information, to collaborate when creating information, to distribute the information, and to allow secure access. On a high level, the main functions supporting management are therefore creation and distribution of knowformation – which are included on the communication side of the model.

As the whole knowformation concept implies, content will finally be targeted at the readers – no matter if it is explicit knowledge presentation, or other kinds of data visualisation. The learning, understanding and memorising of information should be taken into account from creation to delivery. Actions based on information produce concrete results. Information is not valuable if it is not available in the short term memory and understood when making decisions.

There are four possible impacts of the M3.exe memory management model on decision‐making, three positive and one caveat: First, M3.exe may increase confidence in decision making based on the ‘depth’ of its offering. Second, M3.exe may reduce resistance to decisions made in light of past documentation. Third, M3.exe may lower transaction costs of implementing decisions. On the other hand, M3.exe may limit the range of alternatives considered in decisions – if the communication side (including research) is not taken into account as new memories soon enough.

The philosophical difference between objective and subjective information (epistemic knowledge vs. knowformation) should be kept in mind constantly: constructivism is applied to human knowledge and positivism should be applied only to objective epistemic knowledge based on data and information. The subjective part of content is active and dynamic. Information acquisition and creation are different from information retention and maintenance. The key distinction is the role of communication. Objective data and information should be normalised on the extended organisational level: one set of information for the content and end‐to‐end on‐line visibility for the media. These will guarantee the right information at the right time for the right users. Also, the content should be in selected formats, in a ubiquitous network, accessible through different technologies.

Systems between information and knowledge should be based on real requirements. For example, system requirements determine whether a separate system is needed, whether it should be integrated or just accessible through a portal, whether it should manage files as well as modular documents, and whether the objects should be stored as attributes. Technical architecture addresses questions concerning the kinds of clients (operating systems, mobile devices, web servers, etc.) to be supported and the types of unstructured information available in practice.

Requirements management poses various challenges from the perspective of systems in the memory management model. Products are often delivered as incremental releases. Therefore, organisations can maintain information in requirements management system about the features that



244

have been supplied in earlier releases or are planned for implementation in future releases. Also, organisational memory in a requirements management system is crucial for sharing knowledge across organisational levels, product lines and functions. When requirements and new product features in a requirements management system are related to milestone plans and work effort estimates, organisations can have an experiential basis for improving project and release planning. Organisations should not be too dependent on immediate access (such as face‐to‐face meetings) to the expertise of individual experts. Organisational memory in requirements management systems has traditionally reduced this dependence (See e.g. Käkölä 11 March 1998).

Search and retrieval requirements answer questions on what kind of methods should be used: Search from meta‐data and contents simultaneously? Search from extended organisation wide repositories? How to search from modular documents? How should the results be presented? What kinds of filters should be applied?

Library services are quite common for any kind of content: Should it be possible to view checked‐out or draft documents? Can the version numbers be updated manually? What is the difference between working and public versions?

Control is especially needed when security is considered as a part of architecture. Notifications, audit trails and different security levels should be ranked for systems, and in terms of communication, control is not restricted by operating systems features.

Whether to customise applications depends on the business approach, standards, process and vendors of systems. One level – consumerisation – is obvious. This is more detailed than tailoring or customisation. Also open source will be applicable to many more systems – especially client applications. Both are critical for knowledge management, as planning and programming require explicit knowledge – to do is to learn. Open Source has an aura of amateurish enthusiasm and the communities of developers consider their work as a hobby.

Some features of managed information (not communication) should be ranked with any requirements management system, including information workflow and lifecycle. An extended organisation should at least use organisational memory function specific requirements, if those can not be specified according to more case‐based processes.

If the Memory Management Model’s Process is divided into the mnemonic functions of the extended enterprise, the most important common features for the systems between information and knowledge are based on general documentation requirements and more advanced requirements for content management.

1. The clients should access any media, i.e. the content is independent from clients

2. Web services should be offered for all kinds of devices, including mobile devices

3. Content should be searched for, found and retrieved with many different methods

4. Content modules should have working and public versions 5. Audit trails are needed for administrative purposes 6. The workflow and lifecycle are considered important in all organisational units 7. An extended enterprise wide content management platform with standard

access 8. Role based security architecture is important for an extended enterprise (e.g.

LDAP, Active directory, Group management master system) 9. Reusability is one big advantage 10. Modular content management features are used in the memory management

process

Modularity could go deeper into the object‐oriented level, including inheritance, encapsulation and polymorphism if extended enterprise cooperation process requires them. Sometimes



245

the language versions are easier to manage with modular content. However, these must be measured in terms of costs and benefits.

On the process level, extended enterprise content management should move from containers towards reusable content that is independent from systems offering the core information services for explicit organisational knowledge. On a simple process level, the organisation could develop a directory structure whereby a physical reference exists at the place where the required knowledge can be retrieved. This improves the speed of search and retrieval considerably. Also, the type of knowledge container formatting is important for quick retrieval and interpretation. Some standardisation of documents, messages, speech and so forth, may help, if only to retrieve the knowledge and information more easily. The development of an organisation‐wide infrastructure may help in the efficient dissemination and exchange of knowledge, in order to reach the objectives of increased organisational integration, coordination, and distribution of memory contents to the appropriate actors. However, this is only half of the picture. It should be possible to create, use and deliver content with different tools and systems. Content management should go from evolution to design, i.e. extended organisation‐wide planning. Even though the content will be independent from the systems, the number of selected systems should be reduced in the long run, following the laws of economy. However, only one enterprise wide system, especially in the case of collaboration, is not the optimum either.

Current knowledge or content management systems are successful in impacting on the process, but do not appear to have an impact on communication and innovation. They make processes more effective and efficient. Knowledge management should not forget the subjective communication of people. Innovation here does not mean renovating, but real divergent inventions, as well as creative research based on discipline.

The organisational culture plays a critical role in acceptance and adoption of systems, the modification of their use and the overall success in facilitating and conducting the process. The sharing culture should support the use of systems, but content is a critical factor. It should be classified from process to communication, and sharing of content between those two extremes is different. There is a difference in sharing also between content data, information, knowledge, epistemic knowledge or knowformation. The quality of the content and the most relevant appropriate information is critical for the processes. Innovative, dynamic communications require a different approach. The culture of sharing means the inclusion of a presentation layer of the systems to support searching and retrieving public documents containing epistemic knowledge, instead of futile appearances at meetings.

From the workers’ viewpoint the most valuable skill is the creation of new knowledge. Information sharing diminishes its value, and the fewer workers can achieve the same results, the better for organisation’s profitability. From the workers’ point of view the inflation of the value of information does not matter if he/she can generate new, valuable information. If that is done in collaboration, it generates and captures weak but rapid communication; however, the long term memory and the ability to apply its content are much more significant.



246

CONCLUSIONS

Conclusions are summarised here, but also presented after each case study, in the abstract, after various sections, and more generally in the whole Part III. The research question, how can knowledge be managed with information systems, was answered on two levels. The answer to the primary level is that information management systems can not replace management of knowledge. However, they can be used as an organisational memory containing epistemic knowledge and process‐oriented information, and ICT can be used as a supporting media for knowledge through communication. This is not based on processes, but at most working procedures when constructing knowformation between information and knowledge.

The second level aim of the research was to determine in which way information systems can be used for management of knowledge in an (extended) organisation. This started with a categorisation of knowledge – that is the question of what kind of knowledge is meaningful for memorisation in what kind of systems. Also, the background of knowledge management was presented from its history to state‐of‐the‐art. The species of information are different, and should be categorised when planning systems. As the system users do not explicitly classify content, a new concept of knowformation was introduced. Organisational memory was presented in parallel with knowledge management theory and a new memory management model of an extended enterprise was created.

Categorisation or classification is also a key for learning and memory. As organisation consists of individuals, organisational memory can be compared to an individual memory in that sense. There are certain differences, but one benefit on an organisational level will be achieved from archiving. Information redundancy and overload could be avoided with archiving and deletion – but before that short term and long term memories may be separated.

Organisational memory content and media could be separated as well. In all research cases organisations had the same knowledge‐intensive approach that raised systems as the most important bin of organisational memory. However, those were not dedicated organisational memory systems. When systems are built, context and content are learnt and memorised.

The first case systems studied were data mining, data warehouses, data marts, and related knowledge management systems. Case I started from data management, but all systems will integrate. Visualisation and languages, together with other standards, will make portals and web interfaces more unified.

Normally data management is a core on all levels of data, information and knowledge (DIK) management systems. One risk revealed in case I is that processes are too fixed to the systems. Process is an independent entity in the memory management model. Data from operative systems are often criteria for decisions, which designers should be aware of. Meta‐data, ownership, and archiving are relevant for changing architecture of an extended enterprise. Structured data (and even compound documents) are results of knowledge maturity, which was figured with knowformation in individual and organisational triangles. The maturity of containers and knowformation related concepts was presented in a separate table.

Secondly, document management was studied as a generic system for management of knowledge. Case II analysed systems, and also information level from managers, users and administrators viewpoint. Compound documents work only for modular information, which is a minor part of documentation. The formality of the process did not correlate with the maturity of the content. The system requirements began to divide into communication‐oriented and process‐oriented. Also this case indicated that content should be independent from systems.

Maturity is concerning systems as well. Systems are an independent entity in memory management model, but not only designed – they also evolve. Lack of content is one problem when getting new systems into use. Even with legacy systems, version discrepancy is a symptom of knowformation between communication and process. Communication is needed, but e‐mail as a storage is a wrong solution. Communication could be helped with easy access. However, modern ‘knowledge sharing’ is not the answer. Valuable information requires appropriate levels of access control. Mobile clients make the access ubiquitous, which was one the most required system features. Content should be available with any thin client, but the storage platform could be unified. Finally, architecture and containers are not enough; meaningful content plays the main role.



247

So the content of the memory management model seems to be the reason for systems, at least when the human knowledge is the final aim. Case III moved from information to knowledge management in an extended enterprise. Data and information can be normalised on system and organisational level. However, knowformation is not even recommended to be normalised. Documentation alone can not ‘transfer knowledge’ – interaction is needed. Collaboration with an information system is different behaviour than communication through an ICT system, which was indicated already by information species (presentation vs. communication). Trying to make ‘informal knowledge’ explicit by communication has been perceived in many occurrences. Knowledge is too widely interpreted, and – as an implicit layer between information and knowledge has been shown to be of crucial importance – the knowformation concept will help to withdraw ambiguous expressions. Technical media does not convert messages into ‘knowledge’. Even in constructivist sense knowledge must include a justified true belief as such, or it means something vague, as in contemporary language.

Knowformation formulates without management. Knowledge management, on the other hand, requires organisational commitment; not only separate processes, such as documentation. In an extended enterprise, the management is different: maybe extended organisation should be seen as a community of practice. In all cases the interviewees wanted a centralised tool or portal through which information and knowledge could be easily accessed for finding relevant documents or people. However, it looks like finding a ‘right’ person does not actually mean to find the right information, or – not necessarily even the right person. In human memory the context is memorised – sometimes even better than the actual data.

Expertise could be searched for on meta‐level: systems can offer personal background summary that is normally delivered during introductions. The collaboration of individual experts created new knowledge that could only partially be transferred to information services. In the memory management model, content should be independent from systems, also what comes to all kinds of searching. The search functionality appeared to be crucial in all case studies. Rational keywords, Boolean searches etc. will be appropriate in knowledge work. In addition, the retrieval of content (recall of a memory) should be independent on the system. Independence from the constraints of time and space has been even more emphasised with the practicality of mobile devices in case III.

Systems between mature information (data, documentation) and ‘dynamic knowledge’ (including communication) should be able to manage emergent processes of knowformation. In the data, information and knowledge categorisation, one observation is explained by maturity: importance of data seemed to be growing when moving from organisational unit towards extended enterprise. This might depend on communication policy, organisational history, and collaboration culture. The maturity could also be basis for selecting the communication channel. DIK layers were again in focus: alternative sets of data, information and knowledge were figured as the result, including knowformation and entropy. Regarding memory management model entities: systems themselves do not learn, while content and process could reflect learning.

The case studies and the research afterwards verified that memory is the right metaphor for the model. The system entity, however, is not a single system, but proper systems integrated to offer a role‐based, web‐enabled access and on‐line visibility. Their content can be normalised, but knowformation can not. It is mainly an amending interpretation in organisational short term memory. It is supported by communication that is not ‘knowledge sharing’. Knowledge has context, but presentation even with references is different than communication. Information can be modular in documents. Information sharing between collaborators seemed to be purer than internal ‘knowledge sharing’. Extended enterprises should interact for learning, and use standards for collaboration. Classification is required in processes, which should have an end. New kinds of search possibilities might benefit from sociotechnical architecture, and communication technology will offer a ubiquitous access. Because of these findings, a new model was required.

As the final summary of results, a memory management model of an extended enterprise was defined. Related to that, an organisational strategy how to locate different kinds of knowledge to suitable categories was discussed. The knowformation concept was formulating in continuous studies.

Many and varied methods were used for both gathering and analysing research data. The different methods were applicable to different layers of the research objectives. Literature surveys for



248

comparing the case study research were the iterative basis for the whole research. Thus knowledge management literature and other relevant articles were studied constantly using qualitative methods. However, the data was gathered from interviews, ethno‐methodological work observation and participation, literature and articles, but – maintaining an objective grounded theory case study – not from system specifications or marketing materials, or any other system providers’ presentations without critical analysis.

Field observations through work participation, and expert evaluations concerned system‐specific subparts of the research. The quantitative method was used for small samples, when statistical analysis was appropriate. The researcher evaluated the applications, but also referred to existing evaluations. The experts were categorised mainly as Managers, Users and Administrators. On knowledge work, external experts of an extended enterprise were also interviewed, together with other researchers and system specialists. Structured interviews and free form discussions were analysed.

Interviewees, i.e. internal and external experts and other employees, whose knowledge work was managed, filled in the questionnaires and gave interviews. The workers' viewpoints were enlightening concerning the actual problems with knowledge management. Tacit knowledge was a feature of these workers that could be observed somewhat as the researcher was working in the organisation. The explicit‐tacit dichotomy was also reconsidered with the help of knowformation concept.

The passing of time helped to achieve objectivity. The work was carried out in three different layers: 1) KM theory vs. observations, 2) KM systems vs. observations and 3) their combination in the extended organisational memory management model, and strategy. The participatory but timely distant case study research method was complemented with constant follow up of current literature and scientific articles, and iterative observations.

The main result, memory management model (M3.exe), include four high‐level entities. On one end, the process‐based data is managed, while on the other, non‐transactional communication, knowledge creation by research, and organisational networking is supported only. The content and systems make up the other two independent entities. The border of these entities lies between information and knowledge. The systems between information and knowledge require a new collective kind of epimnemonic layer called knowformation, mainly because the epistemic criteria is not required either from the content of information systems nor the current knowledge management literature. System content and applied knowledge concept are not evaluated or categorised scientifically in practice, and the memories affect to construction.

Knowformation is an organisational layer between information and knowledge, mainly based on content of the extended collaborative systems. Mobile ICT supports knowformation, because communication and, thus, ‘knowledge’ is dynamic before being transformed into more mature information. Even though the new technologies for communication are becoming increasingly important, for actual knowledge management the significance is twofold. Communication is often emotional, and therefore ‘knowledge sharing’ has become popular. Any kind of get‐together in virtual meetings is considered as knowledge sharing, even though the actual value of information could be zero or negative. This might be disadvantageous for enterprises. Communication technology between information and knowledge could be used to deliver valuable information to create new epistemic knowledge, although the emotional belief simply consolidates the image of plausible knowformation.

Memory in human beings and machines works on different basis. The critical distinction between human and artificial memory could be seen in the analogy between knowledge and information. Human memory is generally strongly affective, and the mental model requires construction of memories. It should be seen as an advantage that systems do not possess feelings themselves – even though they can deliver emotional messages in human to human communication. Organisational memory includes both aspects, as well as knowledge management theory in general. As a repetition, the common denominator for information management, memory and learning is classification. It is different than the already mentioned species of information, because it is individually constructed, and relates to the human memorising, or learning of concepts. However, classification is also basic to the planning of systems, i.e. modelling the world. The individually constructed classification (how we remember things) and objective species of information converge in knowformation.



249

As there are plenty of paradigms indicating what knowledge management should be, it might be better to specify the concept depending on context. In the future, there could be extended enterprise architecture, which would include generic parts of knowledge management anchored to strategy, employee interactions, group dynamics and technological infrastructure. The values, culture, systems and structures surround this more concrete knowledge management paradigm. Most likely KM will be located within the enterprise and group levels, but it should take account the collaboration between enterprises and other stakeholders. The process versus communication axis should be separated, as the time factor is crucial for the lifecycle of knowformation.

The research has shown that while over ten years ago the focus was on the analysis of what an enterprise knows, now the focus has moved towards communication, learning and innovations. Both approaches are needed. However, as we move further away from knowledge, the new knowformation becomes more relevant. The value of research and innovations is just as high as the value of information created, no matter whether it is due to collaboration or individual work. But the information must be communicated and taught if it is to be transformed into actions.

Mobility has made information ubiquitous. This trend will evolve even more in the future. Systems between information and knowledge are developing in terms of volume of data and information, as well as in increased speed and ease of changing content. The communication status is becoming more and more important, as new knowformation is continuously created and constructed, and process‐based information is changing more rapidly and therefore also distorted and deleted or archived.

This research has shown that only process‐based information can be managed, while the knowledge constructed in communication is only supported by ICT. The knowformation layer covers the missing parts and could be basis for an enterprise to develop its culture and strategy for enhanced communication and learning. Some aspects of classified knowledge, like formal organisational structure, communication technology, and data and information systems can be managed, but the wider concept of knowledge in the organisational culture itself can not be managed. The critical, organisation‐specific distinction is: what can be included in the knowformation of an extended enterprise?

The question of management is also time‐scaled: knowledge creation can not be managed, simply because knowledge to be created is not yet known, or mature enough to be classified. Subjective information can be somewhat managed – even though it means different things to different people. The only well‐managed entity is objective data. However, most content is interpreted, no matter if is derived from data, information or knowledge management systems. Ultimately the traditional data, information and knowledge layers could be reconsidered. In an alternative classification presented, information is the widest category.

The systems support objective knowledge by functioning as organisational memory for storing explicit information and helping tacit knowledge to become explicit through implicit knowformation, but this support of subjective knowledge requires communication technology. The sense‐making aspect of communication will help to transform knowledge into more mature data for processing.

Systems help to store and retrieve information, as well as meta‐knowledge, such as who knows what, and where they can be located; ICT even enables connections with these people, but the communication technology above the systems is required to communicate, and to some extent create knowledge. The increasing complexity, dynamics, fragmentation and decentralisation of knowledge must be isolated under communication, while the systems also manage non‐structured content. Sense‐making knowledge creation activities are in the communication domain, while objective knowledge is memorised as content, managed with systems and processed. These combine into knowformation.

The complexity of knowledge in organisations has been mirrored in the complexity of designing and implementing enterprise‐wide approaches to knowledge management, utilising appropriate knowledge management systems but also taking adequate account of human and organisational factors. From a practitioner’s perspective, only a small percentage of organisational knowledge is written. Most of the ‘knowledge’ is covered by informal, undocumented practices and artefacts. All contacts within an organisation can lead to ‘sharing knowledge’ (i.e. communication) even though most are not intended for this.



250

New knowledge is created on top of the old. Memory requires categorisation. Organisational conceptual research and the memorising process resemble language learning. An organisation learns by comparing new facts, ideas, or tools with existing categories. The systems help to formalise knowledge when analysis work transforms the communication into explicit information.

Leveraging knowledge is not easily done. Documenting procedures, linking people electronically or creating web sites are not enough to get people to think together, share insight and generate new knowledge. Information systems and communication technology, however, are crucial because to leverage knowledge organisations need to focus on the community that owns it and the people who use it, not only the information presentation itself. For human beings, the media is also a message. All systems with communication technology could be unified for the user. Knowformation is recursively generated in the memory management model, between information and knowledge.



251

SYSTEMS BETWEEN INFORMATION AND KNOWLEDGE: DISCUSSION

The post‐industrial era, or information age, or third wave or knowledge society has been here for a while. No matter whether we use term from Huber, Shapiro, Veria, Hope or Drucker, the unifying theme is knowledge. The paradigm shift has put knowledge in the centre of organisational performance.

The creation and use of knowledge is claimed to be one of the major building blocks of an organisation today. The new organisational structure and functioning require new support mechanisms, such as IT tools for storing and sharing information. The new knowledge‐related technologies are in the area of communication. New ways to implement KM systems are needed in order to collaborate efficiently, when both people and knowledge are dispersed in terms of time, physical distance, and organisational structures.

Knowledge is a core asset, but it could be managed traditionally only as the presentation of epistemic information. Managing knowledge is not an advanced form of information technology. The inclusive way to manage other information species is through organisational change and rotation of personnel. Communication somewhat takes care of this sharing. Data must be managed with appropriate systems, and information with processes and communication infrastructure. Infrastructure (memory and communications media) can be managed, as can the data and explicit parts of the information. But only the objective parts of communication‐based knowledge can be ‘managed’, mainly offering a supporting media. The information systems content and practical knowledge as described in KM literature are categorised on other basis than organisational memory really works. Thus knowformation as a content‐based concept between communication and processes is needed.

The significance of systems between information and knowledge can be appreciated only within a given context, social locus and time. As the relationships between actions and knowledge are based on sharing of information and reconstruction of knowledge, communication and presentation technologies are of crucial importance. Reduction in time spent looking for expertise, and the knowledge flows in communities are becoming as critical as the actual data and information that are stored in the organisational memory.

The challenge is that information systems and most knowledge management are mainly based on other than constructivist philosophy. Tightly rationalistic knowledge management literature is epistemologically impoverished, seemingly oblivious to the thousands of years of vigorous and inconclusive debate about the nature of human knowledge. The real importance of the knowledge, learning and communication debate may require researchers to return to their philosophical roots and raise new questions about the nature of the knowledge they are trying to produce. Knowformation concept is used here when we discuss the meaning of content in systems between information and knowledge.

Rationalistic management is based on the cognitive paradigm. This positivistic IT‐engineering approach ignores the value of the construction of knowledge in human interactions, which lead to trust, learning, and knowledge socialisation, while systems increasingly support the sharing and storing of information. Rational causality is not valid in social entities, and management in this context does not accord the reality. Social construction of knowledge depends on the social context, as does all human knowledge processing. There are no great benefits if we try to make technology imitate human thinking, and vice versa. The crucial point is to utilise the advantages of both paradigms.

On the other hand, tacit knowledge has become popular along with rationalism. The tacit vs. explicit knowledge dichotomy is not enough. Polanyi’s practical knowledge – skills based on tacit knowledge – have got explicit roots. There are also skills based on habits and traditions, i.e. beliefs and thus doxastic information. Tacit‐explicit could be replaced with conceptual vs. factual – as has been done with communication vs. presentation dichotomy. Practical information is more like pragmatic communication, where the criteria of epistemic knowledge have no meaning. The communication vs. process ratio scale can be used with knowformation.

Knowledge management paradigms (for example, in referred Anglo‐American and Japanese literature) often deal with ‘knowledge’ concepts that can never be managed, or content that is not knowledge, as observed in the beginning of this research. Systems manage information, but the categorisation into information or knowledge species is not required by default. If we want to manage



252

knowledge, it should be defined as a subset of information. The new knowformation concept, that may be the missing concept of KM theories, recursively circulates between information and knowledge, including priming and other memories in conversion from organisational and individual. In the end, physical media is needed for the memorisation, and media also supports communication.

It is dangerous to blur the concepts of information and knowledge, but the border between these is indistinct in the literature and, on the other hand, mixed on purpose as memory is used as metaphor. Epistemic knowledge as a subset of information presentation can be stored. In practice, applied information must be in context and knowledge is grounded in experiences and other memories. Together they form the fuzzy knowformation on the mnemonic layer of organisational consciousness. Rationalistic management and engineering is based on the results: we use our knowledge in order to achieve particular ends. A corollary of this is that achieving these ends justifies the means. System based on business benefits convert pseudo‐rationalism to a dogma. The way that human knowledge is constructed differs from this, although some parts can be objectively rationalised. Data and document bases provide the means to retrieve and store data and information, while mobile devices and web‐based applications provide the means to present and communicate. The first practical step to adjust systems towards knowformation could be to combine the logical data models of operative systems and reporting. But creating and applying knowledge need the social processes emphasised by the philosophical distinction between information and knowledge species to build the model supporting knowformation. Because memorising needs categorisation, it can be based on unified model.

The syntactic, pragmatic and data‐derived information species are useful for conceptual clarification of ICT, and reflect the reality from the extended organisational point of view, when mixed together, like in the memory management model. Information species based on probability are wrongly considered more crucial, although human interpretation is always needed for knowledge. Everything is converted to numbers, both in information species based on probability, and in the concurrent systems. Semiotics could be attached to semantics, although the probability of information is not calculated. ICT based information represents something, and most often the calculated probabilities are misleading without interpretation. However, some classification should be done by system planners. The users classify when learning, but from an organisational point of view system content mix up with knowledge workers knowformation as the distributed cognition in a short term memory.

A categorisation of information is a presupposition of working systems between information and knowledge, even though knowledge workers do not necessarily require it. In an extended enterprise, that should be done based on the information species, but also based on organisations, different groups and finally individuals. As the most important memory media, information systems are crucial for storing purposes. In addition to explicit information storage, knowformation from communication can be captured and even categorised. However, most KM systems suffer from information overload lacking a utility value. The best practices and lessons learnt work only with special information species, in quite similar context. Also ICT based KM systems have often made the error of delivering irrelevant data recursively. The memory management model, however, offer a framework for building better systems and processes, also for an extended enterprise.20

The dynamic nature of new content is a big challenge for information systems, especially in an extended enterprise. Groupware that supports communication within and across temporal and spatial boundaries creates opportunities, but also limitations. Human knowledge and information needs are dynamic, and change when new information is found. Actually the knowformation concept describes this better, because semantic refer to formation of knowledge. In English semantics ‘knowledge’ in practice is attached to a person, it requires commitment, and is difficult to detach from the person. The wide concept of knowledge depends on the context.

It could be beneficial to make a distinction between system environments, where people are experts in an overall domain, and environments where people are novices or occasional users of the

20 Constructs and models similar to knowformation and memory management model of an extended enterprise were looked for constantly during the research (as recommended by Kitchenham et al 2007; Bereton et al. 2009), and afterwards, using various journals and collective sources (such as Emerald). Similarities and differences are relative: nor completely opposite, neither identical result was found.



253

technology at stake. The case studies concentrated mainly on the experts (in telecommunications and aerospace domains). For example, searching is different if ‘user knows what he or she does not know’ versus searching for unknown without any meta‐information. Functions including search and retrieval work only in well‐defined situations. Search engines depend on the context; otherwise the results are mostly irrelevant. Information overload is not an absolute phenomenon. Knowledge‐based expert systems and information filtering are examples of good engineering, but even there we need user profiles for different users and their roles. In addition, other organisational mnemonic functions such as knowledge acquisition, retention and maintenance should be viewed through information species and context. Applying function is needed to get content into action and archiving to avoid overload.

Discussion about information systems continues to follow the rationalistic paradigm. On the other end, human thinking, with practical and strategic skills and the ability to learn and self‐initiate, are far from the system models. The artificial dichotomy of theory and practice often disables credible interchange between practitioners and research. There are too many tribes interpreting what knowledge is: epistemology, network theory, business, communication, cognitive sciences…

It could be that new knowledge workers’ tribes should learn three fundamentals from aborigines of Australia. Firstly, they are as mobile as can be – they have no permanent residence. Secondly, in their worldview, information is not a separate item – it is everywhere. Last, but not least: they don’t need management of any kind.

Information systems for the management of knowledge are efficient and effective only when used right. Their content should not be perceived as a substitute for human knowledge, but as a memory. Knowledge and information should not be approached in similar manners, and knowformation is between them. Data and information can be processed with systems, while knowledge work with communication can only be supported. Ultimately, the extended enterprise could organise itself on the basis of information. What the importance of content is, that can be managed with systems, is a strategic question for each organisation.

REFERENCES

Abecker, Andreas; Bernardi, Ansgar; Hinkelmann, Knut; Kühn, Otto & Sintek, Michael – German Research Center for Artificial Intelligence (1998): ‘Toward a Technology for Organisational Memories’ in IEEE Intelligent Systems. May/June 1998, pp. 40 – 48

Ackerman, Mark & Stein, Eric W. (1996a): ‘Organisational Memory’. Tutorial of the Hawaii International Conference of System Sciences (HICSS 96)

Ackoff, Russell, L., (1989) ‘From Data to Wisdom’ in Journal of Applies Systems Analysis, Volume 16, pp. 3 – 9. Ackoff, Russell, L., (2003) ‘Creating a Competitive Strategic Advantage’ in Journal of Innovation Management,

Volume 9, Nr. 1. Agger, B. (1991) ‘Critical Theory, Post‐structuralims, Post‐modernism: Their Sociological Relevane’ in Annual

Review of Sociology, Vol. 17, pp. 105 – 131. Alavi, M. and Leidner, D.E. (2001): ‘Knowledge management and knowledge management systems: Conceptual

foundations and research issues’ in MIS Quartely, Vol. 25 No. 1, pp.. 107 ‐ 135. Alur, Nagraj & Davis, Judith R. (1997): ‘How to Improve RDBMSes – Seven long‐term requirements for managing

complex data’ in Byte, April 1997: 115 – 120. Alversson, M. (1991): ‘Organisational symbolism and ideology’ in Journal of Management Studies, Vol. 28, No. 3,

pp. 207 – 225. Alversson, M. and Wilmott, H. (1991): Make Sense of Management: A Critical Analysis. Sage, London. Anderson, John R. (1983): The Architecture of Cognition. Cambridge, Mass. Harvard University Press. ISBN:0‐6740‐

4425‐8 Andren, John – Business Technology Writer, Gartner Group (1997): ‘Electronic Document Management: Selecting

Successful Strategies, Systems and Products’. Executive Symmary Report 5414 for IT Executives. Based on analysis and written product by Gartner Groups's Research, Advisory and Strategic Planning (RAS) services. Published by InfoEdge, Inc, Two Stamford Landing, Stamford, Connecticut, USA.

Arrow, Kenneth (1962): ‘The Economic Implications of Learning by Doing,’ Review of Economic Studies 29, No. 3, 153–173. June 1962.

Attewell, Paul & Rule, James (1984): ‘Computing and Organisations: What We Know and What We Don't Know’ in Communications of the ACM, December 1984 Vol. 27, Nr. 12, pp. 1184 – 1192.

Bacon, Francis: (1605): The Advancement of Learning. Printed for Henrie Tomes (Renaissance Edition), London http://darkwing.uoregon.edu/~rbear/adv1.htm

Bair, James: (21.11.1997): Tietoviikko p. 21 Banker, Rajiv D. / Kauffman, Robert J. (2004): ‘The Evolution of Research on Information Systems: A Fiftieth‐Year

Survey of the Literature in Management Science’ in Management Science, Vol. 50, No. 3, March 2004, pp. 281 – 298. ISSN 0025‐1909

Bannan, Joan (1997): Intranet Document Management. Reading, MA. Addison‐Wesley Developers Press. ISBN 0‐201‐87379‐6

Barabas, C. (1990): Technical writing in a Corporate Culture. Ablex, Norwood, NJ. Barnes, Michael & Kelly, David (1997): ‘Play by the Rules’ in Byte, June 1997: 99 – 102. Barnes, Stuart (ed.) (2002): Knowledge Management Systems – Theory and Practice. Thomson Learning, Oxford.

ISBN 1‐86152‐616‐4 Barney, J. (1991): ‘Firm resources and sustained competitive advantage’ in Journal of management, 17, pp. 99 –

120 Bartunek, J. (1988): ‘The dynamics of personal and organisational reframing’ in R. Quinn & K. Cameron (eds)

Paradox and transformation pp. 137 – 162. Cambridge, MA, Ballinger ISBN 0‐88730‐156‐8 Beckman, Thomas J. (1999): ‘The Current State of Knowledge Management’ in Knowledge Management

Handbook, Liebowitz, Jay (ed.), CRC Press LLC. Beijerse, Roelof Puit (1999): ‘Questions in knowledge management: defining and conceptualising a phenomenon’ in

Journal of Knowledge Management, 1999, No 2, pp. 94‐109. Berger, Peter & Luckmann, Thomas (1966): The Social Construction of Reality – A Treatise in the Sociology of

Knowledge. Penguin Books, USA Bernstein, Cynthia L – Andersen Consulting (1997): ‘Global Sharing of Consulting Knowledge’ in AAAI Spring

Symposium, Knowledge Management Bhatt, Ganesh D. (2002): ‘Management Strategies for Individual Knowledge and Organisational Knowledge’ in

Journal of Knowledge Management, 2002, Vol. 6, Nr. 1. Bieber, M., et al. (2002): ‘Towards knowledge‐sharing and learning in virtual professional communities’ in the

Proceedings of the 36th Annual Hawaii International Conference on System Sciences. Hawaii, 7. – 10. Jan 2002.

255

Bielawski, Larry & Boyle, Jim (1997): Electronic Document Management Systems – A User Centered Approach for Creating, Distributing, and Managing Online Publications. Prentice Hall New Jersey 07458. ISBN 0‐13‐591520‐1

Binney, Derek (2001): ‘The Knowledge Management Spectrum – Understanding the KM Landscape’ in Journal of Knowledge Management, Volume 5 Number 1 Page 33 – 42. Emerald 2001

Blanchard, A.L. & Markus, M.L. (2000): ‘Sense of Virtual Community – Maintaining the Experience of Belonging’ in Proceedings of the 35th Hawaii International Conference on System Sciences 2002. IEEE Computer Society.

Blau, Peter M. (1955): Dynamics of Bureaucracy. Chicago, University of Chicago Press Blechar, M., Varma, S., Plummer, D & Duggan, J. (30.7.1997): Gartner Group Systems Software Architectures (SSA)

Research Note SSA: E‐260‐255 Bohn, R. E. (1994): ‘Measuring and Managing Technological Knowledge’ in Sloan Management Review, vol. 36, no.

1 pp. 61 – 73 Boloix, Germinal & Robillard, Pierre N.: ‘A Software System Evaluation Framework’ in IEEE Computer 28(12) 1995,

pp. 17 – 26. Boyatzis, R. (1998): Transforming Qualitative Information Thematic Analysis and Code Development. Sage

Publications, Thousand Oaks, CA. Boy, Guy (2002): ‘KM as a Scientific Discipline’ in WISE – State‐of‐the‐art, Chapter 5 Brereton, P., Kitchenham, B., Budgen, D., Turnen, M. Khalil, M. (2007): ‘Lessons from applying the systematic

literature review process within the software engineering domain’ in The Journal of Systems and Software, Nr. 80, pp. 571 – 583

Brey, Philip (1997): ‘Philosophy of Technology meets Social Constructivism’ in Techné: Journal of the Society for Philosophy and Technology. Vol. 2, Nr. 3 – 4.

Broadhurst, Roger & Hendley, Tony (1998): Document Management Directory – A Comprehensive Directory of Document Management and Imaging Products and Services. Compiled by Roger Broadhurst, Guide written by Tony Hendley. Published on behalf of Cimtech Limited by University of Hertfodshire Press. 9th edition, 1998. ISBN 0‐900458‐91‐7

Brown, John Seely & Duguid, Paul (2000): The Social Life of Information. Harvard Business School Press, Boston Massachusetts. ISBN 0‐87584‐762‐5

Brown, John Seely & Duguid, Paul (2000: 2): Balancing Act: How to Capture Knowledge without Killing It? Harvard Business Review Article 1.5.2000

Buckland, Michael K. (1991): Information and Information Systems. Praeger, New York. ISBN 0‐275‐93851‐4 Bunge, Mario (1967): Scientific Research I – The Search for System. Springer‐Verlag, Berlin Van Buren, M. (1999): A yardstick for knowledge management’ in Training and Developmen Journal. Vol. 53, No.

5, pp. 71 – 78. Burn, J.M. & Ash, C. (2000): ‘Managing Knowledge for Strategic Advantage in the Virtual Organisation’ in:

Malhotra, Y. (ed.) (2000) Knowledge Management and Virtual Organisations. Idea Group Publishing, pp. 144 – 161.

Burstein F.; Linger, H.; Stein, E. & Jennex M. (1998): ‘Researching Organisational Memory Systems’. Presentation at the Thirty‐First Annaual Hawaii International Conference on System Sciences, Task Force on Organisational Memory, January 1998

Burton, B (25.4.1997) ‘DBMS and Middleware Markets Converge’ in Gartner Group Systems Software Architectures Research Note SSA: M‐401‐211

Burton, B., Percy, A., Radcliffe, J. & Varma, S. (16.4.1997) ‘Informix Loses Its Way: DBMS Vendor Takes Severe Revenue Hit’ in Gartner Group Inside IGC: 041697‐01

Cameron, K. & Quinn, NR. (1988): ‘Organisational paradox and transformation’ in R. Quinn & K. Cameron (eds) Paradox and transformation pp. 1 – 18. Cambridge, MA, Ballinger ISBN 0887301568

Carayannis, E. (1999): ‘Knowledge transfer and technlogical hyper‐learning in five industries’ in Technovation, Vol. 19, No. 3, pp. 141 – 161.

Carnelley, P., Woods, E & Vaughan D. et al. (2001): Ovum Forecasts: Global Software Markets, Ovum. CEN Workshop Agreement (2004): European Guide to good Practice in Knowledge Management – Part 5: KM

Terminology. Brussels. Ref. No.: CWA 14924‐5:2004 E Cerebyte (2005): http://www.cerebyte.com Chandler, Alfred (1977): The Visible Hand: The Managerial Revolution in American Business, Cambridge, Harvard

University Press. Cho Hee‐Jae (1999): ‘Organisational Learning and Memory’. <http://trochim.human.cornell.edu/

gallery/cho/cho.htm#organisational memory>, 25. January 1999 19:29:27 GMT Clark, P. and Staunton, N. (1989): Innovation in Technology and Organisation. Routlege, London. Cleveland, Harlan (1982): ‘Information as Resource’ in The Futurist, December pp. 34 – 39. CMMISM (2002): Capability Maturity Model Integration for Systems Engineering, Software Engineering,

Integrated Product and Process Development, and Supplier Sourcing. Carnegie Mellon University.

256

McClintock, C., Brannon, D. & Maynard‐Moody, S. (1979): ‘Applying the logic of sample surveys to qualitative case studies. The case cluster method’ in Administrative Science Quarterly, pp. 612 – 629.

Conklin, Jeffrey E. (1996): Designing organisational memory: preserving intellectual assets in a knowledge economy. Glebe Creek, MD: CogNexus Institute. Available at: <http://cognexus.org/dom.pdf> and <http://www.gdss.com/DOM.htm> 21. May 1999 20:34:04 GMT

Constant II, Edward W. (1999): ‘The Social Locus of Technological Practice: Community, System, or Organisation?’ in The Social Construction of Technological Systems, Massachusetts Institute of Technology, 7th printing, ISBN 0‐262‐02262‐1.

Curley, Kathy ‐ Lotus Institute ‐ Director of Research (21.9.1998): ‘The Business Drivers of Knowledge Management’ in Lotusphere Europe, Berlin 21. ‐ 25.9.1998

Datamation (1997): Product of the Year 1997 – Server Databases. Datamation, February 1997: 65 – 66 Davenport, Thomas (1998): ‘Some Principles of Knowledge Management’.

<http://www.bus.utexas.edu/kman/kmprin.htm> 28.10.1998 20:48:48 GMT Davenport, T.H., & Prusak, L. (1998): Working Knowledge: How Organisations Manage What They Know.

Cambridge MA: Harvard Business School Press. Davenport, Thomas H. with Prusak, Laurence (1997): Information Ecology – Mastering the Information and

Knowledge Environment. Oxford University Press, New York, ISBN 0‐19‐511168‐0 Davenport, Thomas (01.02.1998): Some Principles of Knowledge Management.

http://www.mccombs.utexas.edu/kman/kmprin.htm 01.02.1998 12:19:48 GMT Dam K.W. (2001): The Rules of the Global Game, University of Chicago Press. McDermott, Richard (1999): ‘Why information technology inspired but can not deliver knowledge management’ in

California Management Review, Nr. 41, pp. 103 – 117. Derrida, J. (1976): Speech and Phenomena, North Western University Press, Evanston, IL. Derrida, J. (1978): Writing and Difference, Routlege and Kegan Paul, London. Dixon, Nancy M. (2000): Common Knowledge – How Companies Thrive by Sharing What They Know. Harvard

Business School Press, ISBN 0‐87584‐904‐0 Documentum (1999): ‘Documentum – Technical Fundamentals’. Presentation the 20th of April, 1999, Tieto

Corporation. Dyer, J.H. and K. Nobeoka, ‘Creating and managing a high‐performance knowledge‐sharing network: The Toyota

case’ in Strategic Management Journal, 2000. 21(3): p. 345 – 367. Edwards, Deming, W. (1986): Out of Crisis, Cambridge University Press Edvinsson L. and Malone, M.S. (1997): Intellectual Capital: Realizing Your Company’s True Value by finding its

Hidden Brainpower. Harper Press, New York, NY. Ehrlich, Kate & Cash, Debra: (1994): ‘Turning information into knowledge: information finding as collaborative

activity’ in Proceedings of the First Annual Conference on the Theory and Practice of Digital Libraries, pp. 119 – 25. http://www.csdl.tamu.edu/DL94/paper/lotus.html.

Douglas, Mary (1987): How Institutions Think. Syracuse University Press. ISBN 0‐8156‐0206‐5 Dreyfus, H.L. & S.E. (1986): Mind over Machine: The Power of Human Intutition and Expertise in the Era of the

Computer. Blackwell, Oxford. Drucker, Peter F. (1969): The age of discontinuity: guidelines to our changing society. New York, NY: Harper and

Row. Drucker, Peter F. (1988): ‘The coming of the new organisation’ in Harvard Business Review, January‐February, pp.

45 – 53 Drucker, Peter F. (1992): ‘The new society of organisations’ in Harvard Business Review, 95‐105. Drucker, Peter F. (1993): Post‐Capitalist Society, Butterworh/Heinemann, Oxford. Drucker, Peter F. (1995): Managing in a Time of Great Change, Truman Talley. ISBN 0‐525‐94053‐7 Drucker, Peter F. (2000): ‘Managing Knowledge Means Managing Oneself’ in Leader to Leader. 16 (Spring 2000):

8‐10. Earl, M. (1998): What on Earh is CKO? – Survey IBM, London Business School, Londen. Ebers, M. (ed.) (1997): The Formation of Inter‐Organisational Networks. Oxford University Press. Eco, Umberto (1986): Travels in Hyperreality. Harcourt Brace, Jovanovich, San Diego 1986. ISBN 0‐15‐191079‐0 Edvinsson, Leif & Malone, Michael S. (1997): Intellectual Capital. London: Judy Piatkus (Publishers) Limited. Eisenhardt, Kathleen M. (1989): ‘Building theories from case study research’ in Academy of Management Review,

Vol. 14, Nr. 4, pp. 532 – 550. Engeström, Yrjö (1992): Interactive expertise: Studies in Distributed Working Intelligence. Helsinki, Univeristy of

Helsinki, Department of Education, Research Bulletin 83, 105 p. EKMF (2003): European Knowledge Management Forum Epstein, K. M. & Henderon, J. C. (1987): ‘Data Envelopment Analysis for Managerial Control and Diagnosis’ in

Decision Sciences, vol. 20, no. 1, pp. 90 – 117

257

McFadden, Fred R. & Hoffer, Jeffrey A. (1991): Database Management. 3rd ed. Redwood City, CA: Benjamin‐Cummings. ISBN:0‐8053‐6040‐9

Von Foerster, Heinz (1970): ‘Thoughts and notes on cognition’ in Garvin, Paul L. (ed.): Cognition: A Multiple View. Spartan Books, New York, pp. 25 – 48. ISBN: 9‐87671‐703‐2

Foley, Mary Jo (1996): ‘Hot New Database Technologies’ in Datamation, September 1996: 45 – 49. Foucault, M. (1980): Power/Knowledge, Pantheon, New York. Galbraith, J. R (1996): Designing Organizations: An Executive Briefing on Strategy, Structure, and Process. San

Francisco: Jossey‐Bass Garvin, D.A. (1993): ‘Building a learning organisation’, Harvard Business Review, January – February, pp. 45 – 53. GEM (2003) Observations from WISE Group Elicitation Method – discussion results and various interviews. Gerrissen Jack & Shcuwer, Robert – Netherlands Open University (1998): ‘Convergence of Personal and

Organisational Knowledge Management in VirtualBusiness‐based Learning’. Seventh International Worksop on Hypertext Functionality: Organisational Memory Systems and HTF. University of Helsinki 13.12.1998

Gersick, C. (1988) ‘Time and transition in work team. Toward a new model of group development’ in Academy of Management Journal, Nr. 31. pp. 9 – 41.

Gibbons, M., Lomoges, C., Nowotny, H. Schwartzman, S. Scopp, P. and Trow, M. (2000): The New Production of Knowledge. Sage, London.

Gilad T. & Gilad B. (1986): ‘Business intelligence: the quiet revolution’ in Sloan Management Review, vol. 27, no. 4, pp. 53 – 61

Glaser, Barney G. & Strauss Anselm L. (1971): The Discovery of Grounded Theory – Strategies for qualitative research. Aldine & Atherton. Chicago & New York. USA. ISBN 0‐202‐30028‐5

Goh, A. (2004) ‘Enhancing Organisational Performance Through Knowledge Innovation: A Proposed Strategic Management Framework’ in Journal of Knowledge Management Practice, October 2004.

Knowledge Garden, The Leadership Alliance Inc. http://www.tlainc.com/articl73.htm Grant, R.M. (1991) ‘The Resource‐based theory of competitive advantage: implications for strategy formulation’ in

California Management Review, Nr. 33, pp. 114 – 135 Grant, R.M. (1996) ‘Towards a knowledge‐based theory of the firm’ in Strategic Management Journal, Vol 17, No1,

pp. 109 – 122 Greimas, Algirdas Julien (1980): Strukturaalista semantiikkaa. Sémantique structurale. Gaudeamus, Tampere.

ISBN 951‐662‐259‐3 Habermas, J. (1972): Knowledge and Human Interests. Heinemann, London Habermas, J. (1984): The Theory of Communicative Action, Vol. 1, Beacon, Boston, MA. Hackney, Raymond – Brunel Business School professor (2008): ‘Trends in Global Knowledge Management’.

Keynote speech in Dokumenttien hallinta ja tietojohtaminen conference, Espoo 13.11.2008 Haig, Brian D. (1995): Grounded Theory as Scientific Method. Philosophy of Education. University of Canterbury. Halper, Mark (1997): ‘Universal Databases Are Still on the Bleeding Edge’ in Datamation, 15.4.1997: 48 – 50 Hamel, Gary & Prahalad, C. K. (1994): Competing for the Future: Breakthrough Strategies for Seizing Control of

Your Industry and Creating the Markets of Tomorrow. Boston: Harvard Business School Press ISBN 0‐8758‐4416‐2

Hammer, M. and Champy, J. (1993): Re‐engineering the Corporation, Harper Business. Hannabuss, Stuart (2001): A wider view of knowledge in Library Management, Vol. 22, Nr. 8/9. pp. 357 – 363. MCB

University Press. ISSN 0143‐51224. Harper R. (1999): ‘Information that counts: a sociological view of information navigation’ in Munro, Alan J., Höök,

Kristina & Benyon, David (eds.): Social Navigation of Information Space. Springer, London, pp. 81 – 89. ISBN 1‐85233‐090‐2

Hars, Alexander – University of Souther California (1998): ‘Position Statement’. Seventh International Worksop on Hypertext Functionality: Organisational Memory Systems and HTF. University of Helsinki 13.12.1998

Hatch, M.J. (1997): Organisation Theory: Modern, Symbolic and Postmodern Perspectives. Oxford University Press.

Hayek, F.A. (1945): ‘The uses of knowledge in society’ in American Economic Review. Vol 35, pp. 1 – 18. Hayes, N. (2001): ‘Boundless and bounded interactions in the knowledge work process: the role of group ware

technologies’ in Information and Organisation. Vol 11, No. 2, pp. 79 – 101. Hedlund, G. (1994): ‘A model of knowledge management and the N‐form corporation’ in Strategic Management

Journal Nr. 15: p. 73‐90. Hedlund, G. and I. Nonaka (1993): ‘Models of knowledge management in the west and Japan’ in P. Lorange, et al.

(Editors): Implementing Strategic Processes, Change, Learning, and Cooperation. Basil Blackwell: London. pp. 117 – 144.

Henry, Nicholas L. (1974): ‘Knowledge Management – a New Concern for Public Administration’ in Public Administrations Review. 34(1974): 3, pp. 189 – 196.

Heyl, B.S. (2001): ‘Ethnographic Interviewing’ in: P. Atkinson, A. Coffey, S. Delamont, J. Lofland & L. Lofland (eds): Handbook of Ethnography. Sage 2001, pp. 369‐383.

258

Heron, J. (1996): Co‐operative Inquiry: Research into the Human Condition. Sage, London. Highsmith, Jim (2001): ‘Documentation is not understanding’ in e‐Project Advisory Service, Cutter Consortium

15.3.2001 Hilwa, A. (9.7.1997): ‘Is the RDBMS a Commodity?’ in Gartner Group InSide IGG ‐070997‐03. Hintikka, Jaakko (1982): Kieli ja mieli – Katsauksia kielifilosofiaan ja merkityksen teoriaan. Otava, Helsinki. ISBN

951‐1‐06757‐5 Hjørland, Birger (1997): Information seeking and subject representation. An activity‐theoretical approach to

information science. New Directions in Information Management, Greenwood Press, Westport, CT. ISBN 0‐3132‐9893‐9

Huber, G. P. (1995): Organizational Learning: The Contributing Processes and the Literature in Cohen & Sproull (eds.): Organizational Learning. CA: Sage.

Hudson, William J. (1994): Intellectual Capital – How to Build It, Enhance It, Use It. John Wiley & Sons, Inc. IEC 82045‐1:2001 standard Jelinek, Mariann & Litterer, Joseph A. (1994): ‘Toward a cognitive theory of organisations’ in Stubbart, Chuck;

Meindl, Jamer R & Porac, Joseph F. (eds.): Advances in Managerial Cognition and Organisational Information Processing. Volume 5. Greenwich, Connecticut, Jai Press Inc. pp. 3 – 42

Johnston, R. and Blumentritt, R. (1998): ‘Knowledge moves to centre stage’ in Science Communication. Vol. 20, No 1, pp. 99 – 105.

Kahneman, D. & Tversky, A (1973): ‘On the psychology of prediction’ in Psychological Review, Nr. 80, pp. 237 – 251.

Kakabadse, N., Kakabadse, A. & Kouzmin, A. (2003): ‘Reviewing the Knowledge management Literature: Towards a Taxonomy’ in Journal of Knowledge Management, Vol 7, Nr. 4, pp. 75 – 91.

Kanter, J. (1999): ‘Knowledge management Practically Speaking’ in Information Systems Management, Fall, pp. 7 – 15.

Kellogg, Wendy A. and Richards, John T., ‘The human factors of information on the internet’, in Nielsen, Jakob (Editor): Advances in Human‐Computer Interaction, Volume 5, Ablex, Norwood, NJ (1995), pp. 1 – 36.

Kelly, Rainer (2001): Managing Organisational Data and Information. Auburn University, August 2001 Kemp, J., Pudiatz, M., Perez, P., & Ortega, A.M. (2002). KM Framework (2nd Release). European KM Forum (IST

2000‐26393): D1.1. Final report. Kerr, Roger M. (1991): Knowledge‐based Manufacturing Management: Applications of Artificial Intelligence to the

Effective Management of Manufacturing Companies. Sydney, Australia: Addison Wesley ISBN 0‐2014‐1622‐0 Kiianmaa, Antero (1996): Moderni Totemismi – Tutkimus työelämästä, solidarisuudesta ja sosiaalisista

verkostoista keskiluokkaistuvassa Suomessa. Keuruuprint Oy. ISBN 951‐97463‐2‐3 Kiesler, Sara & Sproull, Lee (1992): Connections: New Ways of Working in the Networked Organisation.

Cambridge: MIT Press, ISBN:0‐262‐19306‐X Kimberly, J. (1988): A review of Walter R. Nord and Sharon Tucker. ‘Implementing routine and radical innovations’

in Administrative Science Quarterly, Nr. 33, pp. 314 – 316 Kitchenham, B., Brereton, P., Budgen, D., Turnen, M. Bailey, J. Linkman, S. (2009): ‘Systematic literature reviews in

software engineering – A systematic literature review’ in Information and Software Technology, Nr. 51, pp. 7 – 15

Kitson, Terry; Baltup, Mike – Severn Trent (23.9.1998): ‘Delivering Business Benefits through Knowledge Management’. Lotusphere Europe, Berlin 21. – 25.9.1998.

Klemettinen Mika & Ahonen‐Myka Helena (2000): Course on Data Mining, University of Helsinki, Autumn 2000. Klueber, R., Alt, R. & Österle, H. (2000). ‘Implementing Virtual Organizing in Business Networks: A Method of Inter‐

Business Networking’ in: Malhotra, Y. (ed.) (2000). Knowledge Management and Virtual Organisations. Idea Group Publishing, pp. 43‐62.

Knorr, Karin Cetina (1999): Epistemic Cultures – How the Sciences Make Knowledge. Harvard University Press, London, England. ISBN 0‐674‐25894‐0

Konradt, Udo (1995): ‘Strategies of Failure Diagnosis in Computer‐Controlled Manufacturing Systems’ in International Journal of Human Computer Studies, Nr. 43, pp. 503 – 521.

Kontzer, T. (2001): ‘Management legend: trust never goes out of style’ in Call Center Magazine. <http://www.callcentermagazine.com/article/IWK20010604S0011>

Kouzmin, A. (1980): ‘Control in organisational analysis: the lost politics’ in Dunkerley, D. and Salaman, G. (eds) 1979 International Yearbook of Organisational Studies. Routlege and Kegan Paul, London, pp. 56 – 89.

Krippendorff, Klaus (1975a): ‘Some Principles of Information Storage and Retrieval in Society’ in General Systems 2 (1975), pp. 15 – 35

Krippendorff, Klaus (1975b): ‘Information Theory’ in Hanneman and McEwen (eds.): Communication and Behavior. pp. 351 – 389. Reading, MA: Addison‐Wesley

Von Krohg, Georg; Roos, Johan & Slocum, Ken (1994): ‘An Essay on Corporate Epistemology’ in Strategic Management Journal 15, Special Issue, pp. 53 – 71

259

Krulwich, B. & Burkey C. (1997): ‘The Information Finder Agent: Learning Search Query String Through Heuristic

Phrase Extraction’ in IEEE Expert, Sept.‐Oct. 1997, pp. 22 – 27. ISSN:0885‐9000. Kuhn, T. (1970): The structure of scientific revolutions. (2nd ed.) Chicago, University of Chicago Press. Chicago, IL. Kuutti, Kari – University of Oulu (1998): ‘Supporting Perspective Making and Perpective Taking: A Framework for

Contextual Information’. Seventh International Worksop on Hypertext Functionality: Organisational Memory Systems and HTF. University of Helsinki 13.12.1998

Kvale, S. (1996). InterViews. An Introduction to Qualitative Research Interviewing. Sage Publications, Thousand Oaks, CA.

Kühn, Otto & Abecker, Andreas (1997): ‘Corporate Memories for Knowledge Management in Industrial Practice: Prospects and Challenges’ in Journal of Universal Computer Science, Special Issue on Information Technology for Knowledge Management. Vol. 3, No. 8 <http://www.iicm.edu/jucs_3_8/ corporate_memories_for_knowledge/ paper.html> pp. 923 – 954

Käkölä, Timo – University of Jyväskylä (11.3.1999): ‘Groupware Architecture for Requirements Processes in New Product Development’. Requirements management Forum, Helsinki

Lamping, J.; Rao, R. & Pirolli, P. (1995): A Focus + Context Technique Based on Hyperbolic Geometry for Visualizing Large Hierarchies. Proceedings of the SIGCHI conference on Human factors in computing systems. ACM Press, New York.

Lang, Josephine Chinying (2001): ‘Managerial concerns in knowledge management’ in Journal of Knowledge Management, Volume 5 Nr. 1 pp. 43 – 59 ISSN 1367‐3270.

Lank, E. (1997): ‘Leveraging invisible assets: the human factor’ in Long‐Range Planning, 30, pp. 406 – 412 Latour, Bruno (1996): Aramis or the Love of Technology. Harvard University Press, Cambridge ISBN 0‐674‐043‐235 Lawton, George (2001): ‘Knowledge management: ready for prime time?’ in IEEE Computer, Vol. 34, Nr 2, pp. 12 –

14. Lee, J.H., Kim, Y.G. and Yu, S.H. (2001): ‘Stage model for knowledge management’ in Proceedings of the Hawaii

International Conference on System Sciences, Maui, Hawaii. Leonard, Dorothy & Sensiper, Sylvia (2002): ‘The role of tacit knowledge in group innovation’ (first published in

California Management Review, Vol. 40, Nr. 3, 1998, pp. 112 – 132) in Choo, Chun Wei & Bontis, Nick: The Strategic Management of Intellectual Capital and Organisational Knowledge. Oxford University Press, New York, ISBN: 0‐19‐513866‐X

Lewin, A. Y. & J. W. Minton (1986): ‘Determining organizational effectiveness: another look and an agenda for research’ in Management Science 32(5): pp. 514 ‐ 538

Levinthal, D.A. & March, J.G. (1993): ‘The myopia of learning’ in Strategic Management Journal, vol. 14 pp. 95 – 112 Winter 1993

Linde, Charlotte (2001): ‘Narrative and social tacit knowledge’ in Journal of Knowledge Management, Vol. 5, Nr. 2, pp. 160 – 170.

Lindstaedt Stefanie – University of Colorado at Boulder & Zimmermann, Bea – NYNEX Science & Technologies (1999): ‘GIMMe – Group Interactive Memory Manager’. <http://www.cs.colorado.edu/ ~stefanie/GM_Info.html > 15. March 1999 00:30:07 GMT

Lipnack, Jessica & Stamps, Jeffrey (1997): Virtual Teams. John Wiley & Sons. ISBN 0‐471‐16553‐0 Lovio Raimo (1993): Evolution of Firm Communities in New Industries ‐The Case of the Finnish Electronics

Industry. Acta Universitatis Oeconomicae Helsingiensis, Helsinki. Lucas, H. C. (1996): The T‐Form Organization: Using Technology to Design Organizations for the 21st Century. San

Francisco: Jossey‐Bass Lueg, C. (2001): ‘Information, knowledge and networked minds’. Journal of Knowledge Management, Vol. 5 nr. 2,

pp. 151 – 159. MCB University Press. ISSN 1367 – 3270. Lyons, John (1979): Semantics, 1‐2. Cambridge University Press, Cambridge. ISBN 0‐521‐29165‐8 Van Maanen, J. (1988): Tales of the field on writing ethnography. Chicago University of Chicago Press. Malhotra, Yogesh (2000). ‘Knowledge Management and new organisational forms’ in: Malhotra, Y. (ed.) (2000).

Knowledge Management and Virtual Organisations. Idea Group Publishing. Hershey, PA. Malhotra, Yogesh (2002). ‘Why KM systems fail? Enablers and Constraints of Knowledge Management in Human

Enterprises’ in: Handbook of Knowledge Management (ed. C.W. Holsapple) Springer‐Verlag, Heidelberg, in the ‘International Handbook on Information Systems’ Series.

Maltz, D. & Ehrlich, K.(1995): ‘Pointing the way: active colaborative filtering’ in Katz, Irvin R., Mack, R., Marks, L. Rosson, M.B. and Nielsen J (eds.): Proceedings of the Annual ACM SIGCHI Conference on Human Factors in Computing Systems (CHI'95), ACM Press, New York pp. 202 – 209. ISBN 0‐89791‐694‐8.

Maltzahn, Carlos (1995): ‘Community Help: Discovering Tools and Locating Experts in a Dynamic Environment’. <http://info.sigchi.acm.org/sigchi/chi95/Electronic/documnts/shortppr/cm_bdy.htm>

Mandviwalla, Munir; Eulgem Stefan; Mould, Carol & Rao, Srinivasan V.: (1998): ‘Organisational Memory System Design’. Hawaii International Conference on System Sciences (HICSS 98)

March, J. and Simon, H. (1958): Organisations. Wiley, New York.

260

March, S. & Smith, G. (1995): ‘Desing and Natural Science Research in Information Technology’. Decision Support Systems, Vol. 15, Nro. 4, pp. 251 – 266

Marshall, Alfred (1965): Principles of Economics. Macmillan, London; first publishing year 1890 Martin, Paul – Victoria University of Wellington (1998): ‘Scholarly Web Sites as Organisational Memory System’.

Seventh International Worksop on Hypertext Functionality: Organisational Memory Systems and HTF. University of Helsinki 13.12.1998

Maturana, Humberto R. (1980): ‘Biology of Cognition’ in Maturana, H. & Varela, F. (eds.) Autopoiesis and Cognition: The Realization of the Living. Vol. 42, D. Reidel Publishing, Dordrecht, Holland. ISBN 90‐277‐1016‐3

Maurer, H. (1998): ‘Web‐based knowledge management’ in IEEE Computer, Nr. 31, pp. 122 – 123. Mazarr, M.J. (2001): Global Trends 2005, Palgrave. Megill, Kenneth a: & Schantz, Herb (1999): Document Management – New Technologies for the Information

Services Manager. Published by Bowker‐Saur, © Reed Business Information Ltd. West Sussec, UK. ISBN 1.85739‐246‐9

Meyer, M.H. and Zak, M.H. (1996) ‘The design of information products’ in Sloan Management Review. Vol. 37, No. 3, Spring, pp. 43 – 59.

Miles, M.B., (1979) ‘Qualitative data as an attractive nuisance: the problem of analysis’ in Administrative Science Quarterly, Nr. 24, pp. 590 – 601.

Miles, M.B. & Huberman, A.M. (1984): Qualitative data analysis. Beverly Hills, CA Sage Publications. Millen, D.R. (2000). ‘Rapid Ethnography: Time Deepening Strategies for HCI Field Research’ in Proceedings of

DIS’00, pp. 280 – 286. Miller, L. & Nilakanta, S. (1997): ‘Tools for Organizational Decision Support: The Design and Development of an

Organizational Memory System’ in Nunamaker, J.F., Jr. & Sprague, R.H., Jr. (eds.): Proceedings of the Thirtieth Hawaii International Conference on System Sciences, Vol. 3, pp. 360 – 368,

Mintzberg, Henry & Gosling, Jonathan R. (2002). ‘Reality Porgramming for MBAs’ in Strategy + Business, 1st Quarter. Booz & Company.

Moilanen (9.9.1997): Discussions with Petri Moilanen, Informix Country Manager Monash Technology Glossary (2001) <http://www.monash.com/glossary.html>, August 2001 Morrison, Joline & Weiser, Mark (1998): ‘A Research Framework for Empirical Studies in Organisational Memory’.

Proceedings of the 29th Hawaii International Conference on System Sciences (HICSS‐29) <http://www.computer.org/conferen/proceed/hics29/vol3/ABSTRACT.HTM> 17. February 1998 19:01:15 GMT

Muller, Robert J: ‘Database Management Review – Oracle8: Worth the Wait?’ in BYTE September 1997: 111 – 113.

Murray, P. and Myers, A. (1997): ‘The facts about knowledge’ in Information Strategy. Vol. 2 No. 7, September, pp. 29 – 33.

Murray, P. (2000): ‘Designing business benefits from knowledge management’, in Despers, C. and Chauvel, D. (eds): The Present and the Promise of Knowledge Management. Butterworth‐Heinemann, Boston, MA, pp. 171 – 194.

Myers, M.D. and Newman, M. (2007): ‘The qualitative interview in IS research: Examining the craft’ in Information and Organization. Nro. 17, pp. 2 – 26

Naisbitt, John (1994): Global Paradox, Nicholas Brealey Publishing. ISBN 1‐85788‐051‐X Nance, Barry: ‘IBM Builds a Better DBMS’ in BYTE August 1997: 133‐134 Nardi, Bonnie A. & O´Day, Vicki L. (1999): Information Ecologies, MIT Press, Cambridge. ISBN 0262140667 Niiniluoto, Ilkka (1980): Johdatus tieteenfilosofiaan. Käsitten ja teorianmuodostus. Otava, Keuruu. ISBN 951‐1‐

05435‐X. Niiniluoto, Ilkka (1986): Tiedon asema informaatioyhteiskunnassa in Kanava, Vol. 14, Nr. 2, pp. 71 – 78. Niiniluoto, Ilkka (1993): The Aim and Structure of Applied Research in Erkenntnis, 38: pp. 1 – 21. Kluwer Academic

Publishers, Netherlands. Nisbett, B. & Roll, L. (1980): Human interference – Strategies and shortcomings of social judgement. Englewood

Cliffs, NJ, Prentice Hall. Nonaka, I., Reinmöller, P., & Toyama, R. (2001). ‘Integrated information technology systems for knowledge

creation’ in: M. Dierkes, A. Berthoin‐Antal, J. Child, & I. Nonaka (eds.), Handbook of Organisational Learning and Knowledge Management, Vol. Paperback Edition. Padstow, UK: Oxford University Press, pp. 827 – 848.

Nonaka, I. & Takeuchi, H. (1995): The Knowledge‐Creating Company. How Japanese Companies Create the Dynamics of Innovation. Oxford University Press. ISBN 0‐19‐509269‐4.

Nonaka, I. & Konno, N. (1998): ‘The Concept of ‘Ba’: building a foundation for knowledge creation’ in California Management Review. Vol. 40 No. 3, pp. 40 – 54.

Nonaka, Ikujiro; Toyama Ryoko; Konno Noboru (2001): ‘SECI, Ba and Leadership: a Unified Model of Dynamic Knowledge Creation’ in: Managing Industrial knowledge, creation, transfer and utilisation. Sage Publications.

Nurminen Markku (1986): Kolme näkökulmaa tietotekniikkaan. WSOY, Helsinki. ISBN 951‐0‐13535‐6 ObjectSpace (22.9.1997): ObjectSpace – Next‐Generation Distributed Computing. <http://www.objectspace.com/

22.9.1997>

261

ObjectStore (22.9.1997): ObjectStore – Database for the Web. <http://www.odi.com/ 22.9.1997> O´Dell, Carla; Grayson, C.; Jackson J.R. (1998): If Only We Knew What We Know – The Transfer of Internal

Knowledge and Best Practice; Chapter 9: ‘Culture, The Unseen Hand’, The Free Press, New York, ISBN 0‐684‐84474‐5.

O’Leary, Ann: ‘Managing Mission‐Critical Text’ in Byte, September 1997: 43 – 44 O'Leary Daniel E. (1998): ‘Enterprise Knowledge Management’ in Computer, Vol. 31, No. 3, March 1998, pp. 54 –

61. Oinas‐Kukkonen, Harri – University of Oulu (1998): ‘Organisational Knowledge Creation in Web Information

Systems’. Seventh International Worksop on Hypertext Functionality: Organisational Memory Systems and HTF. University of Helsinki 13.12.1998

Orlikowski, W. J. (1992): ‘Learning from Notes: Organisational Issues in Groupware Implementation’ in J. Turner & R. Kraut (eds.): Proceedings of CSCW’92, Toronto, Canada, November, 1992, pp. 362 – 369

Orlikowski, W. J. (2002): ‘Knowing in practice: enacting a collective capability in distributed organizing’ in Organisation Science. Vol. 13, No3, pp. 249 – 273.

Orr, J. E. (1996). Talking about machines: An ethnography of a modern job. Cornell University Press. Parsons, T. (1959): ‘General Theory in Sociology’ in Merlon, R.; Brooom, I. & Cottrell, I. S. Jr. (eds.): Sociology

Today: Problems and Prospects. Basic Books, New York, pp. 3 – 37 Pemberton, Steven (2004): Lecture in Helsinki XML conference, 14.9.2004. Pearce, W. Barnett (1989): Communication and the Human Condition, Southern Illinois University Press. Percy, A. (27.6.1997): Oracle Unveils Version 8. Gartner Group Systems Software Architecture Research Note SSA:

E‐550‐236 Persson, Per‐Arne – Linköping University (1998): ‘Linking the Experts – a Case for a Hypertext Application’. Seventh

International Worksop on Hypertext Functionality: Organisational Memory Systems and HTF. University of Helsinki 13.12.1998

Pettigrew, A. (1988): ‘Longitudinal field research on change – Theory and practice’. Paper presented at the National Science Foundation Conference on Longitudinal Research Method in Organisations, Austin.

Pfeffer, J. (1982): Organisations and organisation theory. Marshfield, MA, Pitman. Piatetsky‐Shapiro, G. & Frawley, W. (1991): Knowledge Discovery in Databases. AAAI Press, Menlo Park, California Pirttilä, Anneli (1997): ‘Competitor Information and Competitive Knowledge Management in a Large, Industrial

Organisation’. Doctoral Thesis, Lappeenranta University of Technology. ISBN 951‐764‐175‐3 Polanyi, Michael(1958): Personal Knowledge: Towards a Post‐Critical Philosophy. Chicago, University of Chicago

Press. Reprinted 1974. ISBN 0‐7100‐7691‐6 Prusak, L. (2001): ‘Where did knowledge management come from’ in IBM Systems Journal, Volume 40, Number 4. Purvis, R.L., Sambamurthy, V. & Zmud, R.V. (2001): ‘The assimilation of knowledge platforms in organisations: an

empirical investigation’, Organisation Science, Nr. 12, pp. 117 – 135. Quinn, James B. (1992): Intelligent Enterprise: A Knowledge and Service Based Paradigm for Industry. New York,

Free Press Quinn, R. E. & Rohrbaugh, J. (1983): ‘A spatial model of effectiveness criteria: Towards a competing values

approach to organizational analysis’ in Management Science, vol. 29 no. 3, pp. 363 – 377 Radcliffe, J: Oracle8: ‘Raising the Bar for OLTP Scalability’ in Gartner Group Strategic Data Management Research

Note SDM: P‐450‐241 Reber, Arthur (1995): Implicit Learning and Tacit Knowledge: An Essay on the Cognitive Unconscious. New York,

Oxford University Press. ISBN 0‐19‐510658‐X Ruggles, R. (1998): ‘The state of the notion: knowledge management in practice’ in California Management

Review, Nr. 17, pp. 45 – 62. Santos, Jack (2008): A CIO’s Perspective on Information Overload. Burton Group presentation 25.4.2008. Sauser, Brian (2007): From Systems Biology to the Biology of Systems. Stevens Institute of Technology. Schein, E. (1999): The Corporate Culture Survival Guide – Sense and Nonsense about Culture Change. Jossey‐Bass

Publishers. Scherberger, K. (27.6.1997): ‘SQL Remote Shows Cracks in Mobile Environments’ in Gartner Group Strategic Data

Management Research Note SDM T‐520‐254 Schroeck Michael J. (2001): ‘Insights from the Front Line: ODS: The Cornerstone of Information Integration’.in DM

Review, January 2001 Schwabe, Gerhard (1993): ‘Providing for Organisational Memory in Computer‐Supported Meetings’.

Arbeitspapiere, Lehrstuhl für Wirtschaftsinformatik, Universität Hohenheim, Stuttgart. Schwartz, D., Divitini, M. & Brasethvik, T. (2000): Internet‐Based Organisational Memory and Knowledge

Management. Idea Group Publishing. Schön, D. (1983): The Reflective Practitioner – How Professionals Think in Action. Basic Books, New York.

262

Selvin, Albert M. – Bell Atlantic Corporation (1998): ‘Supporting Granular Reuse of Knowledge Elements in an Organisational Memory System’. Seventh International Worksop on Hypertext Functionality: Organisational Memory Systems and HTF. University of Helsinki 13.12.1998

Semich, William (1996): ‘Database technologies race to meet business demands’ in Datamation, September 1996: 5

Shannon, Parker (2001): Enterprise Data Architecture <http://www.dbstuff.com/architecture.htm>, August 2001 Shegda, Karen (1999 and 2000): ‘Integrated Document Management Software: Perspective’ in Datapro – Gartner

Advisory 2.10.1999 and .29.2.2000 Shum, Simon Buckingham – Knowledge Media Institute, UK (1998): ‘Organisational Memory and Interpretation:

Roles for Hypertext Functionality’. Seventh International Worksop on Hypertext Functionality: Organisational Memory Systems and HTF. University of Helsinki 13.12.1998

Skarra, Poul – NetSys Technology Group – Management Consulting (March 18, 1999): ‘Knowledge as a corporate asset, managed and developed to create a competitive advantage’. Executive IT Forum 1999. Wanha Satama, Helsinki, March 17 – 18, 1999

Skolimowski, H. (1994): The Participatory Mind: A New Theory of Knowledge and of the Universe. Arkana books, London.

Sluimer, Hugo – Vice President, European Operations, Verity Inc. (10.6.1999): ‘Knowledge Management – Expansion of Information and How to Control It’. Sybase Seminar, Vantaa

Snowdon, D. (1998): ‘I only know what I know when I need to know it – embracing the actived management of tacit knowledge’ in Knowledge Management. Vol 3, No 1, March, p. 1.

SOLID (1997): Negotiations, e‐mail correspondence and telephone conversations with Sales director and President of Solid Information Technology Ltd. 16.9.1997

Spender, J.‐C. – New York Institute of Technology (1998): ‘Pluralist Epistemology and the Knowledge‐Based Theory of the Firm.’ <http://iris.nyit.edu/~spender/text/Plurorg326.htm> 31. March 1998 14:07:32 GMT

Stein, Eric W. & Zwass, Vladimir (1995): ‘Actualizing Organisational Memory with Information Systems’ in Information Systems Research, Vol. 6, no. 2, pp. 85 – 117

Strange, K. (27.6.1997): ‘Oracle8 and Data Warehousing: Are Big Databases Enough?’ in Gartner Group System Software Architectures Research Note SSA: P‐300‐241

Strauss, Anselm (1987): Qualitative analysis for social scientists, Cambridge, England, Cambridge University Press. Strauss, Anselm & Corbin, Juliet (1998): Basics of qualitative research. Techniques and procedures for developing

grounded theory. Second Edition. Sage Publications. California. USA. ISBN 0‐8039‐5939‐7 Strange, K. & Radcliffe, J. (30.7.1997) ‘Oracle for Data Warehousing: The Experience So Far’ in Gartner Group

Strategic Data Management Research Note SDM: P‐600‐262 Sveiby, Karl‐Erik (1990): Valta ja johtaminen asiantuntijaorganisaatiossa. Jyväskylä, Weilin+Göös. 336 p. ISBN 951‐

35‐3984‐4 Sveiby, Karl‐Erik (1997): The new Organisational Wealth. Berrethh‐Koehler, San Francisco, CA. Sveiby, Karl‐Erik; Miller, Riel (2000): ‘A dialogue about knowledge work in the future’. Seminar at Helsinki

University of Technology, 1.9.2000. Swan, J. & Newell (2000): ‘Linking knowledge management and innovation’ in Hansen, H.R., Bichler, M. and

Mahrer, H. (eds), Proceedings of the 8th European Conference on Information Systems, Vienna University of Economics and Business Administration, pp. 591 – 598

Swan, J. (2001): ‘Knowledge management in action: integrating knowledge across communities’ in Proceedings of the Hawaii International Conference on System Sciences, Maui, Hawaii.

Sydänmaanlakka, Pertti (2003): Intelligent Leadership and Leadership Competencies. Developing a Leadership Framework for Intelligent Organizations. Helsinki University of Technology.

Szulanski, Gabriel (1997): ‘Intra‐Firm Transfer of Best Practices’ in Campbell, Andrew & Sommers Luchs, Kathleen (eds.): Core Competency‐Based Strategy. International Thomson Business Press, pp. 208 – 235.

Tampoe, M. (1994): ‘Exploiting the core competencies of your organization’ in Long Range Planning, vol. 27, no. 4, pp. 66 – 77

Tapscott, Don & Caston, Art (1992): Paradigm Shift – The New Promise of Information Technology. McGraw‐Hill, Inc

Tissen, R., Andriessen, D. and Depres, F.L. (2000): The Knowledge Dividend. Financial Times and Prentice Hall, London.

Trimmer, D. (1993): Downsizing. Addison‐Wesley. Tsui, Eric (2000): Exploring the KM toolbox. Knowledge Management, Vol.. 4, Nr. 2, pp. 11 – 14. URL

<http://www.bl.com.au/km/articles/exploring.htm> Turban, Efraim (1995): Decision Support and Expert Systems; Management Support Systems. Englewood Cliffs,

NJ, Prentice Hall ISBN: 0‐0242‐1701‐8 Tuomi, Ilkka (2000): Corporate Knowledge. Theory and practice of intelligent organizations. Metaxis. ISBN 951‐

98280‐0‐1 Twenty First Century Schools: http://www.21stcenturyschools.com/Vygotsky.htm

263

Tyson, Kirk W. M. (1986): Business Intelligence: Putting It All Together. Lombard, IL: Leading Edge. ISBN: 094‐11010‐0‐2

Varma, S. (23.9.1997): ‘Uncertain Universality: Coming Soon Through a Vendor Near You!’ in Gartner Group Strategic Data Management (SDM) Strategic Analysis Report

Virtanen, Turo (1989): ‘Informaation lajit ja tietohallinto – Informaation tutkimuksen ja suomalaisen tietohallintokeskustelun anti tietohallinnon tutkimuskohteen määrittelyssä’, Hallinnon tutkimus 3, pp. 180 – 212.

Wagner, E. (2000): ‘Communities of practice: the structure of knowledge stewarding’ in Despres, C. and Chauve, D. (eds): The Present and the Promise of Knowledge Management. Butterworth‐Heinemann, Boston, MA, pp. 205‐ 224.

Walsh, James P. (1995): ‘Managerial and Organisational Cognition: Notes from a Trip Down Memory Lane ‘ in Organisation Science 6, 3 pp. 280 – 321.

Walsh, James P., & Ungson, G. R. (1997). ‘Organisational Memory’. Academy of Management Review, Vol. 16, No. 1, 1991 pp. 57 – 91. Reprinted in Prusak, Laurence (ed.): Knowledge in Organisations. Butterworth‐Heinemann, 1997, ISBN 0‐7506‐9718‐0

Wareham, J. and Gerris, H. (1999): De‐Contextualising Competence: Can Business Best Practice be Bundled and Sold? European Management Journal Vol 17, No 1, pp. 39 ‐ 49

Watson‐Manheim, M.B. & Belanger, F. (2000): ‘Exploring Communication‐Based Work Processes in Virtual Work Environments’. Proceedings of the 35th Hawaii International Conference on System Sciences 2002. IEEE Computer Society.

Watt, Stuart: (2000): ‘Agents and Objects; Agents, Knowledge and Social Work – How people interact with complex systems.’ Knowledge Media Institute, http://www.comp.rgu.ac.uk/staff/sw/presentations/kmiagents.ppt.

Watterson, Karen (1996): ‘Database replication explained’ in Datamation, September 1996: pp. 62 – 66. Web‐enabled Information Services for Engineering (2002): Review of KM Tools and Products Weick, Karl. E. (1979): The Social Psychology of Organizing. 2nd ed. Reading, MA: Addison‐Wesley. ISBN: 0‐0755‐

4808‐9 Weldon, Jay‐Louise (1997): ‘RDBMSes Get a Make‐Over’ in Byte, April 1997 pp. 109 – 114. Weldon, Jay‐Louise (1997B): ‘A Career in Data Modeling’ in Byte, June 1997 pp. 103 – 106. Wenger, Etienne (1998): Communities of Practice – Learning, Meaning, and Identity. Cambridge University Press.

ISBN 0‐521‐43017‐8. Verkasalo, Matti (1997): On the Efficient Distribution of Expert Knowledge in a Business Environment.

Department of Electrical Engineering, University of Oulu Whitehead, Steven D. – GTE Laboratories Incorporated (1994): ‘Auto‐FAQ: an experiment in cyberspace

leveraging’. <http://www.ncsa.uiuc.edu/SDG/IT94/Proceedings/Agents/whitehead/whitehead.html>, 18. September 1994 19:41:02 GMT

Wickramasinghe, Nilmini (2003): ‘Do we practise what we preach? – Are knowledge management systems in practice truly reflective of knowledge management systems in theory? ‘ in Business Process Management Journal, Vol. 9, Part 3, pp. 295 – 316

Wiig, Karl M. (1993): Knowledge Management Foundations – Thinking about thinking. How People and Organisations Create, Represent and Use Knowledge. Schema Press, ISBN 0‐9638925‐0‐9

Wiig, Karl M. (1994): Knowledge Management Methods – Practical Approaches to Managing Knowledge. Schema Press, ISBN 0‐9638925‐3‐3

Wiig, Karl M. (1994): Knowledge Management. The Central Focus for Intelligent‐Acting Organisations. Schema Press, ISBN 0‐9638925‐1‐7

Wiig, Karl M. (1997): ‘Knowledge management – where did it come from and where will it go?’ in Expert Systems with Aplications. Vol 13, Nro. 1, pp. 1 – 14.

Wiig, Karl M. (1998): Introducing Knowledge Management into the Enterprise. Chapter manuscript for Handbook of Knowledge Management. Liebowitz, Jay & Wilcox, Lyle (editors).

Wiig, Karl M. (14.5.1998): Knowledge management and knowledge productivity in contemporary society and organisations. The Council for the Promotion of Information Society through Education, Research and Culture, Hanasaari Congress Center, Espoo, Finland.

Wiio, Osmo A. (1981): Information and Communications: A Conceptual Analysis. University of Helsinki 1F/6/1981. ISBN 951‐45‐2189‐7

Wijnhoven, Fons (1998): ‘Designing Organisational Memories: Concept and Method’ in Journal of Organisational Computing and Electronic Commerce, Vol. 8, No. 1, pp. 29 – 55

Wilson, F. Desmond, J. and Roberts, H. (1994): ‘Success and failure of MRPII implementation’, in British Journal of Management. Vol 5, No3, pp. 221 – 240.

Wilson, T.D. (2002) ‘The nonsense of 'knowledge management’ in Information Research, 8(1), paper no. 144 <http://InformationR.net/ir/8‐1/paper144.html>

Winograd, T. & Flores, F. (1986): Understanding Computers and Cognition: A New Foundation for Design. Ablex Publishing, Norwood, NJ. ISBN: 0‐89391‐050‐3

264

Winter, S. G. (1993) ‘On Coase, Competence, and the Corporation,’ The Nature of the Firm, Oxford University

Press, Oxford, UK WISE (2003): Web‐enabled Information Services for Engineering – Facts sheet and presentations WISE D1.1 (2003): State‐of‐the‐Art Surevey: Managing Engineering Knowledge WISE D1.3 (2003): Review of Knowledge Management Tools Yin, R.K., (1981): ‘The Case Study Crisis: Some Answers’ in Administrative Science Quarterly, No 26, pp. 58 – 65. Yin, R.K., (1994): Case Study Research: Design and Methods, 2nd ed., Sage, Thousand Oaks, CA. Yolles, Maurice (2000): ‘Organisations, Complexity, and Viable Knowledge Management’ in Kybernetes, Vol. 29,

Issue 9/10. Zack, M.H. (1999): Knowing and Strategy, Butterworth‐Heinemann. Zack, M.H. (1999/2): ‘Managing codified knowledge’ in Sloan Management Review, Vol. 40, No 4, pp. 45 – 58. Zeleny, Milan (1985): ‘Management Support Systems: Towards Integrated Knowledge Management’ in Human

Systems Management, 7(1987)1, pp. 59 – 70. Zmud, Robert W. (1978): An Empirical Investigation of the Dimensionality of the Concept of Information. Decision

Sciences, 9: 4, pp. 187 – 195 Zuboff, S. (1988): In the Age of the New Machine. Basic Books, New York.

265

APPENDIX 1: ACRONYMS AND ABBREVIATIONS

AI Artificial Intelligence API Application Programming InterfaceAPROS Asia‐Pacific Researchers in Organisation StudiesBFILE Binary File (outside the database)BLOB Binary LOB (unstructured, in database)BU Business Unit CA Computer Associates CLOB Characters LOB (single byte characters, e.g. ASCII)CMMI Capability Maturity Model IntegrationCOLD Computer Output to Laser DiskCO Case Organisation CoP Community of Practice CSCW Computer Supported Co‐operative WorkDAML DARPA Agent Markup LanguageDARPA Defense Advanced Research Projects AgencyDB Database DBMS Database Management SystemDCE Distributed Computing EnvironmentDCOM Distributed Common Object ModelDIKW Data, Information, Knowledge and WisdomDM This acronym has multiple meanings (Data Management, Data Mining, Decision Making, Document

Management…) but is used mainly in Document Management context in this research. DMA The Document Management Alliance (DMA) Task Force was formed in 1995 under the Dauspices of

the Association for Information and Image Management (AIIM). Its objective was a standard that assures any PC can request and receive documents from any repository regardless of platform or application across an enterprise. The DMA 1.0 Specification, approved in December 1997, contains a framework for intranet/Internet document management. A comprehensive client and server standard, DMA 1.0 enables multiple clients to talk to multiple document management systems. Cf. web protocols and standards like WebDAV.

DMS Document Management SystemDSS Decision Support System DTD Document Type Definition ECMS Enterprise Content Management SystemEDMS Enterprise Document Management SystemEIS Executive Information SystemFAQ Frequently Asked QuestionsFINSE Finnish Systems Engineering CouncilGDSS Group Decision Support SystemGUI Graphical User Interface HTF Hyper Text Functionality HTML Hyper Text Markup LanguageIBM International Business MachinesIBIS Issue Based Information SystemICT Information and Communication TechnologyIDOM Integrated Document and Output Management IM Information Management IPR Invention process IS Information System IT Information Technology JDBC Java Database Connection KM Knowledge Management LgP Legal Process LOB Large Object (BFILE, BLOB, CLOB or NCLOB)

LDM Lotus Document Management

LP Learning Process

266

M3.EXE Memory Management Model of an Extended Enterprise

Meta‐data Information about the object, e.g. document. Name of Author, creation date, version number, format, etc. are typical elements of meta‐data.

MIS Management Information SystemsNASA National Aeronautics and Space AdministrationNCLOB N‐Chararacter LOB (many multiple‐byte chars, fixed)OCR Optical Character RecognitionODBC Open Database ConnectivityODMA The Open Document Management API (ODMA) is a client‐side standard that enables desktop

applications to talk to a single document management system. ODMA API is platform‐independent. ODS Operational Data Store OLAP On‐line Analytical ProcessingOLE Object Linking and Embedding. Microsoft's Object Linking and Embedding (OLE) is a set of services

that enable applications to interact and interoperate. OLE has become a de facto standard for linking objects.

OMIS Organisational Memory Information SystemOMS Organisational Memory SystemOMT Object Modelling TechniqueOODBMS Object‐Oriented Database Management SystemsORDBMS Object‐Relational Database Management SystemsOS Operating System PC Personal Computer PDF Portable document formatPDM Product Data ManagementPr Project PTI Palkka‐ ja työsuhdeilmoitus — Salary and Employment Notification R & D Research and DevelopmentRDF Resource Description FrameworkRDBMS Relational Database Management SystemROLAP Relational Online Analytical ProcessingSAP R/3 SAP R/3 is the former name of the main enterprise resource planning software produced by SAP AG.SDK Software Development Kit SECI Socialisation, Externalisation, Combination, Internalisation SMP Symmetric MultiProcessingSOAP Originally defined as Simple Object Access Protocol SQL Structural/Sequential Query Language TP Transaction Processing TREC Text REtrieval Conference UDDI Universal Description, Discovery and IntegrationVLDB Very Large DataBase WAIS Wide Area Information ServersWebDAV Web‐based Distributed Authoring and Versioning. An interface standard that defines the syntax

used by an authoring tool when interacting with a Web server. WOLAP Web‐based Online analytical processing WSDL Web Services Description Language XML A simplified subset of SGML, eXtensible Markup Language (XML) is a universal format for defining

and managing complex data on the Web. It separates content from style, thus enabling distribution to multiple media or platforms. Intended primarily to meet the needs of Web content providers, XML provides industry‐specific markup, vendor‐neutral data exchange, and media‐independent publishing.

267

APPENDIX 2: SYSTEMS INVESTIGATED

Category Platforms, products, solutions, applications, systems, services or infrastructure

Companies

RDBMS, Business Intelligence

InterBase, (Embedded) Empress, Raima Database Manager++, Velocis Database Server, ADABAS D, SQL Server, Unify DataServer, DB2 Universal database, Oracle products, Informix OnLine Dynamic Server, SOLID Server, SQL Server, SQL Anywhere, Red Brick, DB2, RDB, Ingres/OpenIngres, Teradata, Progress, SQLBase

Borland, Empress Software Inc. Raima Corp., Simba Technologies Software AG, Sybase Inc., Unify Corp., IBM, Oracle, Informix, Solid, Microsoft, Watcom, Ingres, Various other

Desktop Databases

4th Dimension, Access, Paradox, Approach, FileMaker Pro, Visual FoxPro, dBase, Clipper

Microsoft, Borland, Lotus, Clarion/Claris corp.

ODBMS GemStone, ObjectStore, Raima Object Manager, Falcon, Versant ODMBS, ObjectSpace Voyager

Gemstone Systems, Object Design Inc., Raima Corp., Unidata Inc., Versant Object Technology, ObjectSpace Inc.

SQL Dev. Tools SQL_Connect, Platinum Fast Load for Sybase and Microsoft SQL Server, Platinum SQL‐Archive for Sybase, SQL‐Retriever

Objective Interface Systems Inc., Platinum Technology Inc., Santa Cruz Operation

OLAP Tools PaBLO, Essbase Andyne Computing, Arbor Software Data Mining AAdvantage, Stor/QM, VisiFlow, IntelliSTOR

QM, Nucleus, Atom, Molecule, KnowledgeSTUDIO & KnowledgeSEEKER, KnowledgeExcelerator & KnowledgeAccess, Predictor, DecisionMaker, Clementine, The Predictive Modeling Markup Language, Decision Maker, DBMiner, Data in Depth, Enterprise Miner, Fast Fuzzy Cluster, Model 1, IBM alphaWorks, IBM Intelligent Miner, IBM Knowledge Management Solutions, Neural Nets, InterLeap, Alice d'ISoft, Data Mining Software, KnowledgeMiner, Data Mining Software and Predictive Modeling Software, DataEngine, Mars, MineSet, PolyAnalyst, Thinkmap, Thinx Data Visualization Software, Two Crows Data Mining Technology Report, Viscovery SOMine, ADVIZOR, PV‐WAVE, JWAVE, IMSL, JNL, VRCharts, XpertRule, Zinnote, Data Mining Workbenches,

Acxiom Corporation, Alterian Ltd., Angoss Software Corp., Attrasoft, Clementine, The Data Mining Group, DogHouse Enterprises, Inc., Dimension 5, 3DV8 Inc., Enprotelligence, SAS Institute, Group 1 Software, IBM, Innovative Solutions, InterLeap Inc., ISoft, Knowledge Based Systems, Magnify, Inc., Management Intelligenter Technologien GmbH, Salford Systems, Megaputer Intelligence, Ltd., Silicon Graphics, The National Center for Data Mining – University of Illinois at Chicago (UIC), The Pisa KDD Laboratory, PolyVista, Inc., Rulequest Research, Smart Research BV., Spotfire, Eudaptics Software GMBH, Visual Insights, Visual Numerics, WizSoft Software,

Data Warehouse Informatica for Operational Data Storage, Statistical Data Warehouse, Crystal Decisions, MS Data Transformation Service, Progress, Artemis, SAP / R3, SAP Business warehouse, Oracle applications – Data Mart Suite, Designer Develop, Application Server;

Informatica, Business Objects, Microsoft, Progress, Artemis, SAP, Oracle

Data Marts Ingres, Adaytum, MS Excel and Access for data mart and reporting, Essbase

Ingres, Cognos, Microsoft, Hyperion

Information Retrieval

The Center for Intelligent Information Retrieval, Center for Networked Information Discovery and Retrieval, Correlate, Cross‐Languages Information Retrieval Resources, Diogene, Informatica, Information Retrieval, Information Retrieval and Information Extraction on the Web, Inmagic, Knowledge Navigation Suite, MultiCentrix, Omnidex Database Search Engine, Text REtrieval Conference, The unCommon Hub, Willow, CGKAT, Excalibur RetrievalWare, LinguistX Categoriser

University of Massachusetts, Correlate, Datagold, Delphes Technologies International, The Glasgow Information Retrieval Group, GuruNet, University of Glasgow, Painted Word Inc., Project Aristotle, University of Washington

268

Knowledge Discovery

XCise, ClearForest, DataSet, Enfish, IBM Intelligent Miner for Text, iCrossReaderTM, LexiQuest, Map.net, MapStan.net, MindServer, Research Outlet and Integration, SemioMap, Stratify (formerly Purple Yogi), System of Meaning Integrities structural Creation, Clusterer, Categoriser, Extractor, Skills Cartridges, Online Miner, Email Miner, TextAnalyst, TextSmart, TheBrain, Thinkmap, Trivium, The Vantage Point, Virtual Self, WEBSOM, Hyperknowledge

Antarcti.ca Systems Inc., Applied Semantics, Brosis Innovations Inc., Fractal Edge Limited, IBM, Inxight Software, Inc., Knowledge Base, LexiQuest Inc., Maps of the Web, Quiver, Inc., Readware Technology, RecomMind, Semio Corporation, Temis, Torch Concepts, Helsinki University of Technology, Yellowbricks

Collaboration

IBM Websphere (db2, Oracle), MS Office for Teamtools, Documentum, Oracle, Vignette, Brainstorming, Innovation Toolbox, ParaMind, AntWorld, Caucus, CollabFab, CoIcon, YEdit, Correlate Products, Dataware II Knowledge Management Suite, DocuTouch, GroupPort, Omniprise, Orbis Intelliware, JBlurb, Work2Gether, KnowledgeLEAD, Livelinked, Livelink 8, Papirus, SiteScape, ThoughtWeb, VisionCompass, Visual Issue Mapping System, OpenText Corporation KM suite,

IBM, Oracle, MS, EMC, Vignette, Brainstorming software, Idea Fisher, Commugen Inc., Communispace corporation, Correlate Technologies, Dataware Technologies, Inc., Groupthink Inc., Ikimbo, Intelliware Systems, KMTechnologies, Cadenza, OmniSpace Technologies, OpenText,

Knowledge Mapping & Ontology

Axon Idea Processor, Banxia Decision Explorer, IHMC Concept Map Tools, Inspiration, WebCokace, OntoBroker, OntoSeek, Shoe

Axon Research, Banxia Software, IHMC

Organisational Memory

Quest Map, Lotus Notes based apps, GroupWare, Answer Garden, (ASSIST, AnswerWeb), Knowledge Cache, Stakeholder behaviour histories, Goal Attainment subsystem, The Coordinator, Meeting memory, Pattern Maintenance subsystem, PRSM, FRUM, TAXOPS, CYRUS,TOPC, CABARET, Collabra Share, Café Construction Kit, Cybrarium, Web Storybase, Constellation, Collective Memory Palace, Corporate Memory (IBIS, gIBIS, CM/1, QuestMap); SIBYL; Living Design Memory, Raison d'être, ICE, AutoFAQ, FAQ Finder, Community Help, GIMMe, Team Memory, ABC, Display Systems, Scholarly Web Sites, System Architect.

Group Decision Support Systems, Price Waterhouse, McKinsey, Cypress Semiconductor, Federal Express, Xerox,

On‐line Training System

ABC Academy, Achievement Technologies, alto learning, AskMe Enterprise, A∙TrainES, a2zclass, VELOCITY Platform, Course Builder, Examiner, CLI e‐Learning Solution, ClassLeader, OnTrack, DOTS, EDU System, Edutorium, E‐education.com, eLearningKit, ePath Learning, EPPIK, E‐Presentations Pro, e‐train, e2train, ExamPlanner, EZTracker, Generative Resources, Hosted Test, ILIAS, IntraLearn, Intranet U, ISOtrain, Knowledge, KnowlegeBridge, Pinnacle Learning Manager, The Learning Manager, Aurora, Learn‐up, LUVIT AB (Publ), mGen Enterprise, NetBoomerang, Netucate Online, Online Expert, Pedagogue Solutions, RISC, ScholarWorks, Schoolspace, Serf, Simtrex, Skills2Go, SkillSpace, SocratEase, SOF,

AcuLearn, knowledge=power, Atlantic Associates, Nicheware, Awake Interactive Learning Systems, Boniva Software, Campus Crossing, @campus, Inc., Cisco Learning Institute, Courses by Wire, Coursewhere, Dennis Tester OnLine Learning Systems, DKSystems Online, Docent, Inc., WebRaven, Dover Software, Argus, Pro‐ductivity Systems, FlexTraining, Gyrus Systems, Hippo Software Inc., INTECH Interactive Technologies, Intellicue, Interact! Multimedia Corp., Interactive Media Corporation, InterWise Millennium, Softekpr, Knowlagent, The Knowledge Business, Knowledge Impact, KnowledgeWindow.com, Websoft Systems, Learnframe, Inc., WIN, The Learning Springs, LearningSoft, LearnLinc Corporation,

269

Solstra, SpeedTrainer, Teamscape, Study‐It®, Test‐It®, THEORIX, tmsSEED, Traineasy, Trainframe, Training Register, TrainingForce.com, Tuesday Interactive, Universal Educator, WebTrain, WIDS, LearningSpace

LearnWright LLC, Lectra, Link‐Systems International Inc., mGen, Inc., NetDimensions, ObjectJ Inc., LearnKey, Online Learning Ltd, Online‐Testing.net, rapidLD, Saba Software, Inc., Electronic Education Foundation, Ltd., British Telecom and Futuremedia PLC, SyberWorks, Inc., Syntrio Inc., Testmaker.com, VCampus Corporation, VeriLogix, Inc., Vitalect, Inc., Vuepoint.com, Wanadu

Document (content) management

Antarcti.ca , XCise Pro, Clear Case (sw & dm), PC Docs, ClearForest, DataSet, Documentum, Hyperwave Information Server, IBM Intelligent Miner for Text, iCrossReaderTM, LexiQuest, Map.net, MapStan.net, Optegra, Readware, RecomMindServer, Research Outlet and Integration, SemioMap, Stratify, System of Meaning Integrities structural Creation, Clusterer, Categoriser, Extractor, Skills Cartridges, Online Miner, Email Miner, TextAnalyst, TextSmart, TheBrain, Thinkmap™ Platform, SEE‐K, The Vantage Point, Virtual Self, WEBSOM

Antarcti.ca Systems Inc., Applied Semantics, Brosis Innovations Inc., EMC, Enfish Enterprise, Fractal Edge Limited, Inxight Software, Inc., Knowledge Base, LexiQuest, Inc., Quiver, Inc., Readware Technology, RecomMind, twURL, Semio Corporation, Stratify, Temis, Thinkmap.com, Torch Concepts, Trivium, Yellowbrix

Project & Resources

Progress with Project View, Track View, Capacity View, Need View; Artemis, MS Project, Hermes

Progress, Microsoft, Artemis, Hermes

Reporting Actuate Reporting System, Spectrum Report Generator, Visual CyberQuery, FlexQL, r‐tree, Safari ReportWriter,UDMS, IQ/Objects, KCSI Control Report Inquiry and Data Entry Utilities, RM/Companion, Platinum InfoReports, TPL Report, Report Writer, 4SS‐Report, Safari Report Writer; Crystal Decisions with Crystal Reports, Crystal Info, Microsoft Excel and Access, SAS based reports, Hyperion Enterprise, SAP SEM

Actuate Software, Cabletron Systems Inc., Cyberscience Corp., Data Access Corp., FairCom Corp., Interactive Software Systems Inc., IQ Software Corp, King Computer Services Inc., Liant Software Corp., Platinum Technology Inc., QQQ Software, Raima Corp., SuperNova Inc., Unidata Inc., Microsoft, SAP Business Warehouses

Search engines Copernicus, Fulcrum, <isindex>, WAIS (Z39.50 protocol), TREC, Verity, PLS, Yahoo!, AltaVista, Harvest, PointCast, Google, various portals

Fulcrum, Search Software America, Copernicus‐software

Not categorised KM related applications

Action Technologies, Process Handbook,GrapeVINE, Orenge, Meta4Mind Set, Meta4 KnowNet, K‐STATION, Lotus Domino/Notes, Lotus DiscoveryServer, Lotus K‐station, TeamRoom, QuickPlace, InterCommunity, IBM KnowledgeX, IBM Enterprise Information Portal, Microsoft exchange, Outlook, Verity Software K2, Enterprise, LiveLink, Fulcrum Knowledge Server/PCDocs/CyberDocs, Knowledge Management Suite, Active Knowledge, CommonKAds‐Centaur (CoKace), Wordnet, CriticalReach, QLoops

Meta4, IBM, Microsoft, Hummingbird, Cerebyte Inc., QCircuit

270

APPENDIX 3: QUESTIONNAIRES

Case I Questions about Data Warehouses (DW), Data Mining and Data Marts (DMs), Data Archiving (DA) and Knowledge Management (KM) applications and data: 1 Applications 1.1 What applications there are used for DW, DMs, DA and KM in the Case Corporation? Why just those applications have been selected? 1.2 What applications are on the market and not used in the Case Corporation? Are there any advantages in those alternative applications? 2 Data 2.1 What kind of data there is in DW, DMs, DA and KM systems? Where does that data come from? What is the quality of the data? 2.2 For what purpose the data are used? Are there any data models? To whom or to what other system the data are delivered? 3 Future 3.1 What kinds of plans there are for the future of DW, DMs, DA and KM applications and data? Are there any development projects? 3.2 What problems must be solved for the future? Are there any kinds of improvement proposals or ideas for the future? Case II questionnaires Phase 1 questionnaire Personal background and the case organisation in general

What are your current job and task functions in the case organisation? What are your current projects and duties in your department or laboratory? Experience with documentation How long have you been doing work with documentation? (As an Administrator, User, Manager) Processes What are your work processes and business rules that guide you in your work in the case organisation? Documentation entities on the case organisation level What kinds of documentation entities there are in the case organisation?

Processes, information search and distribution

Processes of documentation What are the processes of documentation like in your work in the case organisation? Needs for Delivery What are the needs for delivering documentation? How to Search How are you searching for information? Reading How many documents you are reading per day?

271

Writing, modification, updating What amount of documents you are writing or updating per day?

Needs and problems

What kinds of documents What kinds of documentation you are dealing with? Needs for own work What are your needs for documentation in your current job? The case organisation needs What are needs for documentation for the whole case organisation? Problems with own work What are problems concerning documentation in your current job? Case organisation wide problems What are problems concerning documentation for the whole case organisation?

Current tools and the future

What kind of information What kind of information is included in your documents? Tools Which products you are using for documentation (e.g. Word, PowerPoint, Excel, Outlook, but also other than Office products: Lotus Notes, html‐editors, project management tools etc.) Documentation management Are you using any systems or tools for documentation management (e.g. Documentum, Lotus Domino Document Manager, or PC‐Docs) How to enhance Documentation Management How the documentation could be managed better in the focal company? Comments Any other comments?

Phase 2 questionnaire All the following questions from category 2 to 7 are formed with first explaining the

feature and the asking ‘how important is feature for your work in the future’, for example:

Cross‐repository searching

One aspect that separates products targeted to workgroups from those meant for the enterprise is the capability to search across multiple servers, i.e. multiple repositories. Cross‐repository searching has become a core backbone service for enterprise document management software solutions.

How important is cross‐repository searching for your work in case organisation in the future? Not important… ……very important0 1 2 3 4 5 6 7

How well Documentum / Lotus Domino Document Manager is supporting cross‐repository searching?

Not well… ……very well0 1 2 3 4 5 6 7

272

1 Personal background information – name and selected process or project 2 System Background Information 2.1all Documentation management in general 2.2all Document management system is integrated 2.3all Compound document management system 2.4all Objects as attributes 3 Architecture 3.1a UNIX and Windows clients 3.1m WWW servers and thin clients 3.1u All types of unstructured information 4 Search and Retrieval 4.1u Search from attributes and contents simultaneously 4.1m Cross‐repository searching 4.1a Compound search ability 4.2u Search results representation 4.2m Filtering 5 Library Services 5.1u View checked‐out documents 5.1a Manual version numbers 5.1m Working and public versions 6 Control 6.1u Notifications 6.1a Audit Trails 6.1m Security levels 6.2a OS security features 7all Workflow and Lifecycle 7.1all Workflow 7.2all Lifecycle 8u Customized applications 8a Standards 8m Business Approach 8.1m Price of the software 8.2m Vendors of document management systems 9all Other requirements

273

Case III The discussion on state of the art was raised with the following questions:

1. Does your organisation have any formal policy concerning knowledge management? KM strategy? What does it say?

2. Does your organisation have formalised KM practice/processes/responsibilities? Can you describe the KM Tools in use?

3. Does informal KM practice exist? 4. What are the possibilities and obstacles of KM perceived to be?

The rest of case III interviews were more free form, and all the questions that have been discussed in three separate interview sessions and in frequent occasions after these, are combined here. 1) Background information • What is your job title? • What is your educational background? • Please describe your role in the organisation and in your current project? • Please specify your organisational location (management, design, customer etc.)? • Does your organisation have any formal policy concerning knowledge management? KM strategy? What does

it say? • How long have you been working in other companies? • How long have you been working in this company? • How long have you been doing the current tasks? • Are you involved in other projects at the same time when working in this project? • How much time do you spend on the projects? (Give percentage for each.) • How closely are you involved with collaborated projects? • How long experience of distributed product development projects do you have? • The role of the team in the projects (duties, sub‐processes, associates) 2) Distributed project work • How many partner companies do you collaborate with in your current project? • Where are these companies located? (countries and cities?) • How much collocated work is required in a distributed project in your opinion?

o Which issues determine how much collocated work you arrange in a project? o How long are the periods? o Who travels? o Where are the collocated periods arranged?

• Focal Company: Support for the partner: What kind of support practices you try to provide to the partner? • What kind of support are you offered for using systems? Is expert support available? • Do you have a resident engineer/liaison person in distributed projects? Why?

o How have you organised his/her work? What is his/her work description? o What have been the benefits of this arrangement? o How is answering to suppliers’ questions /problems arranged? o How do you make sure that the supplier acquires enough product‐related knowledge and

information? • Does the Focal Company offer product and technology training related to your work? If yes, how do you

evaluate this training? How often do you (want to) participate to these trainings? • Focal Company: Project monitoring in networked projects: How do you monitor networked projects? • What kind of information is collected? (e.g. amount of resources spent?)

o By whom? • How do you use the collected information? • Would you need to know more? What kind of information? • Do you have any special monitoring on your partners? • Do you have any tools to follow the progress in the project? (e.g. time control)

o What kind of tools? Who uses? Who inputs the information? • What is most problematic in partner monitoring in globally distributed projects? Why? • Do you inform project members about project progress? How? • How is the work divided into the network between you and your partners?

o Why is this division chosen? o Is the development work divided into separate modules? How?

274

o Has the division caused any problems? What kind of problems? • Have you used some process model in this project? What process model do you use? (a free form description

about the project phases is an alternative) o Which (product creation process) model? o Partner: Do you use the processes of the Focal company or your own processes? o (Have you had in this project any special decision points / milestones before the project

can move to the next phase?) o How often do you have milestones in this project? o What have been the most critical milestones in distributed projects according to your

experience? • How many teams do you have in this project? • What is the number of people working 1) in general in collaboration and 2) in one team? 3) and in the future

(increasing/decreasing)? • Why have you chosen this number of teams • Why have you chosen this number of persons? • Do you have teams that include members from different companies?

o Why? o Does that bring any special challenges?

• Do you have teams that include persons representing different professions or functions in organisations? o What is the benefit? o Does this pose any special challenges?

• Has there been a project kick‐off meeting? • Agenda? • In which phase of the project was it arranged? • Who did participate? Who was missing? • How long? • Was it useful? Disadvantages? Benefits? Why? • What kind of networked product development projects / collaboration do you have? • What kind of network structures do you usually have?

o How many companies? o What kind of companies? o What kind of roles do the companies in the network have?

• Is everything related to information exchange and communication agreed in the contracts with the partner? 3) The current work practices related to knowledge management and sharing o What kind of information do you use or need concerning human resources management of the programs, and

what kind of information is available? o Does your organisation have formalised KM practice/processes/responsibilities? Can you describe the KM

Tools in use? 3.1 Information management o Document management: How do you take care of document management in a networked project? o Do you have any common document management system between the partner companies? o How do different companies access the documents? (How are the documents transferred between

companies?) o How do you inform the project members (external/internal) about the changes in the documents or

processes? o Do you have a shared information storage? o Who has access? o Who is responsible? o Do you have established procedures for information exchange? 3.2 Re‐use of information and knowledge o What kind of unofficial work related groups do you belong to? (CA, enablers, research projects, work groups,

SIGs, news groups, study groups) o In your current project, do you re‐use information/knowledge from the previous projects? o If yes:

Why? What kind of information/knowledge do you re‐use? How does the re‐use happen?

275

o If not: Why? Would you need some information from previous projects?

3.3 Communication and meetings What kind of management reporting is there between you and your subordinates, managers and colleagues?

Where from do you get the information to your reports? What kind of reporting for designers is there in your organisation? How are the progress and content of work

reported in your team? What kind of communication is there between you and your partners during the project execution? (e.g.

formal, informal, e‐mail, face‐to‐face, document exchange…) What kind of formal meetings do you use in this project between companies? (e.g. weekly meetings,

videoconferences, milestone reviews, other reviews, inspections, netmeetings, conference calls) Does the distributed/collaborated project require different meeting arrangements compared to normal

project? How often do you meet face‐to‐face? Do you think that is often enough? Who are participating? Which issues are discussed in face‐to‐face meetings? What kind of issues should be

discussed in face‐to‐face meetings and not in remote ones? • Besides meetings, which other issues need to be communicated between companies?

o Who communicates? o How often? o How do you choose which media you use for the purpose in question (e.g.

questions/answers, change requests, informal communication)? • How successful has the communication been between companies in this project?

o Regarding the amount of communication o Regarding the quality of communication o Is relevant information (info that you need for your work) readily available and easily

found? o Are the important persons available?

• What are the most difficult problems in communication during a distributed project? Why? o How could communication be improved in your opinion?

• Contacts across company borders: Do you feel you have enough contacts & communication with those members in other companies, which are somehow important for your own working tasks?

o How have the contacts been created (kick‐off / earlier projects)? o If you feel you don’t have enough contacts, can you name some barriers for these

contacts? o (Can you imagine that cooperation with someone in the project could have helped you in

managing the project?) 3.4 Needs for knowledge management and sharing • What are the possibilities and obstacles of KM perceived to be? • Do you need information and/or knowledge from others? What sort of? How often/how much? • What kind of information and knowledge you need from the colleagues? • Do you need to inform the colleagues? In which situations? How often? • How, from where and from whom do you receive the most important information and/or knowledge for your

tasks and your work (in this project)? 3.5 Knowledge creation • Where is the knowledge created that is most relevant to you?

o In your own team / company o In the partner company

• In your opinion, what does it mean to create knowledge? • How do you create knowledge in your work?

o By yourself (e.g. …) o In your own team (e.g. …) o In collaboration with the partner (e.g. …..)

276

4) The effect of the work culture on knowledge sharing • Ways of sharing knowledge in your company: • How do you share knowledge? (e.g. face‐to‐face, information systems?) • The values and practices: do they support knowledge sharing?

o Incentives? • Do you have any common practices with your partner?

o To what extent? o Give an example?

• Differences in working practices between companies (or between organisational units): • Have there been any problems/challenges in working due to the different working habits between

companies? o What kind of problems; examples? o Do you think it is important to develop common working practices between cooperating

companies? o In what kinds of projects are common practices needed most, in your opinion?

• Problems: What were the major problems related to work practices during the project? o What were the causes of these problems? o How problematic did you see o Geographical distances? o Different work practices? o Crossing company borders? o Cultural differences?

5) ‘The knowledge border’ between the organisations • Team relationships: Teams and team borders

o Where do you feel that you belong to (e.g. this company/this project / project team inside your company / project team across company borders / technical expertise group)?

o Which team(s) have the most power to control the project in your opinion? Why? (What was the path to this situation?)

• Inter‐company cooperation o Do you prefer working in intra‐ or inter‐company projects? Why? o Do you get enough feedback across company borders? o What kind of feedback do you need? o Do you feel that your team’s (company’s) work was dependent on the work in other

companies? How? • Knowledge sharing between organisations:

o Do people share knowledge fluently across company borders? o Is knowledge sharing more difficult across the organisational border? o Why? How can the situation be improved? o How is your work dependent on the information or knowledge of the partner?

6) The role of information systems in knowledge sharing • Do you use information systems to exchange information and knowledge? • What kinds of information management systems and information sources do you use (including organisation

maintained, defined channels and informal exchange of tacit knowledge)? • Experiences on the usefulness of the information systems (positive/negative)? • Evaluate e.g. the following information sources

o Internal databases, portals, servers, intranet, virtual communities, newsgroups o External information services, standards, www o Meeting practices o Personal interaction

• How important are the different sources in different phases of your work? What is the content? How to find the right information source? The quality of various sources (topicality, completeness)? What kinds of problems there are with different sources?

• Is it possible to manage work without the information systems? • Do you have shared systems with the partners?

277

7) The barriers to knowledge sharing • The experiences on the barriers to knowledge sharing:

o What kind of problems have you faced? o How have the problems been solved?

• In your opinion, what are the obstacles for knowledge sharing • Between people? (e.g. distance, language, education…) • Between companies? (e.g. project personnel, lack of agreed practices…) • Give some examples? Characterise?

o Do you have solutions to these problems? 8) The facilitators of knowledge sharing • The experiences on the facilitators of knowledge sharing

o What kinds of helpful advice have you received from the colleagues? o How have you received this advice? o What procedures are helpful in getting advice?

• In your opinion, what are the factors that facilitate knowledge sharing o Between people? o Between companies? o Give some examples? Characterise? o Are these things already happening?

9) How can the prototype system facilitate knowledge sharing • Would you consider the following prototype system functions beneficial?

o Create a workspace for a project o Tracing information o Find relevant person (process/task related) o Find relevant documents (process/task related) o Find out who used the document and in which phase of process?

10) Experiences • What are the exceptionally positive experiences from knowledge management and information systems? • What kinds of problems there have been in finding or sharing knowledge?

o What kind of required information you have never reached? o How have you solved the situation when information is not available?

• How do you process the information you have created? o What information results does your work produce? o Where is this information stored and by which channels?

• Do you have time to share knowledge?

278

APPENDIX 4: LIST OF TABLES AND FIGURES

Table 1. The SECI Process ........................................................................................................................... 30 Table 2. Managerial interpretation of knowledge ..................................................................................... 31 Table 3. Examples of knowledge management instruments ..................................................................... 32 Table 4. Narrowing road of organisational memory .................................................................................. 34 Table 5. Process of building a model from case study research ................................................................ 42 Table 6. Comparison of cases ..................................................................................................................... 47 Table 7. Types and Gaps of Knowledge ...................................................................................................... 64 Table 8. Types of Knowledge in Organisational Memory ........................................................................... 66 Table 9. Retention Structures ..................................................................................................................... 74 Table 10. Knowledge maturity and effectiveness of organisational memory media ................................ 80 Table 11. Levels of analysis, criteria and results for organisational memory means ................................. 80 Table 12. Number of experts responding from different organisations .................................................... 91 Table 13. Workflow and lifecycle ‐ importance by role ........................................................................... 129 Table 14. Invention process requirements ............................................................................................... 138 Table 15. Learning process requirements ................................................................................................ 140 Table 16. Projects requirements .............................................................................................................. 142 Table 17. Administrative process requirements ...................................................................................... 143 Table 18. Legal process requirements ...................................................................................................... 145 Table 19. Case organization (CO) requirements ...................................................................................... 147 Table 20. Importance of system features ................................................................................................ 190 Table 21. Importance of system features in summary, compared between processes and roles .......... 191 Table 22. A sample from technical evaluation reports ............................................................................ 199 Table 23. Variables of the cases ............................................................................................................... 202 Table 24. Retention media and memory contents................................................................................... 206 Table 25. Species of data, information and knowledge – reflecting management systems ................... 211 Table 26. Knowformation and meta‐information .................................................................................... 214 Table 27. An alternative to tacit vs. explicit dichotomy ........................................................................... 228

279

Figure 1. Introducing a memory management model of an extended enterprise (M3.EXE) ....................... 7 Figure 2. Search based on meta‐data in external documents ................................................................... 24 Figure 3. Data, information and knowledge pyramid ................................................................................ 27 Figure 4. Collaborators ............................................................................................................................... 29 Figure 5. The knowledge spiral ................................................................................................................... 31 Figure 6. Information species based on probability ................................................................................... 53 Figure 7. Enhanced semiotic triangle ......................................................................................................... 55 Figure 8. Information species based on communication ........................................................................... 57 Figure 9. An example of classes, objects and relations .............................................................................. 57 Figure 10. Information species based on presentation.............................................................................. 59 Figure 11. Information species based on processing ................................................................................. 60 Figure 12. Example from common knowledge management strategy enablers ....................................... 62 Figure 13. Relations between memory types, operational activities and problem perceptions ............... 73 Figure 14. Processes of organisational memory ........................................................................................ 76 Figure 15. Concepts related to organisational memory ............................................................................. 82 Figure 16. First indications from extended enterprise ............................................................................... 89 Figure 17. Enterprise data architecture ..................................................................................................... 93 Figure 18. Content originates from Asset Management, but is diffusing .................................................. 96 Figure 19. Document management systems in general ........................................................................... 117 Figure 20. System integration .................................................................................................................. 119 Figure 21. Compound system ................................................................................................................... 120 Figure 22. Workflow ................................................................................................................................. 122 Figure 23. Simplified document lifecycle ................................................................................................. 122 Figure 24. Lifecycle ................................................................................................................................... 124 Figure 25. Very important DM system features ....................................................................................... 125 Figure 26. Search results representation ................................................................................................. 127 Figure 27. Workflow and lifecycle scattergram........................................................................................ 129 Figure 28. Document lifecycle activities ................................................................................................... 131 Figure 29. Invention process requirements ............................................................................................. 138 Figure 30. Learning process requirements ............................................................................................... 139 Figure 31. Process requirements .............................................................................................................. 141 Figure 32. Administrative process requirements ..................................................................................... 143 Figure 33. Legal process requirements .................................................................................................... 144 Figure 34. Case organisation requirements ............................................................................................. 146 Figure 35. Work packages of case III ........................................................................................................ 150 Figure 36. Extended Organisation include community members groups ................................................ 156 Figure 37. Topics related to reference persons ....................................................................................... 157 Figure 38. Initial plan of case III to integrate data, information and knowledge ..................................... 170 Figure 39. Case III system architecture .................................................................................................... 171 Figure 40. Starting page from case III prototype ...................................................................................... 173 Figure 41. Tracing as context logging in WISE .......................................................................................... 176 Figure 42. Search results require context ................................................................................................ 178 Figure 43. Knowledge object – an example from ‘who knows what’ in case III ...................................... 198 Figure 44. Mnemonic functions for memory management model of an extended enterprise ............... 207 Figure 45. Reverse Knowledge Axis .......................................................................................................... 212 Figure 46. Knowformation in individual and organisational triangles ..................................................... 213 Figure 47. Knowledge objects .................................................................................................................. 215 Figure 48. Alternative sets of data, information and knowledge ............................................................ 215 Figure 49. Relevant knowledge objects in personal context ................................................................... 220 Figure 50. Case III advanced search options ............................................................................................ 223 Figure 51. Sketch for an expanded SECI model ....................................................................................... 227 Figure 52. Memory management model of an extended enterprise (M3.exe) ....................................... 230 Figure 53. Knowformation and the systems between information and knowledge ............................... 232 Figure 54. Maturity of containers and knowformation related concepts ................................................ 233 Figure 55. Lifecycle of knowformation in M3.exe .................................................................................... 236

Systems between information knowledge

Documents

Transcript of Systems between information knowledge