8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 1/11

2012 45th Hawaii International Conference on System Sciences

Design, implementation, and evaluation of a Virtual Meeting Tool-basedinnovation for UML technology training in global organizations

Mervi Koivulahti-Ojala & Timo KäköläUniversity of Jyväskylä

40014 University of Jyväskylä, Finland{meelheko, timokk}@j yu.fi

Abstract End-user training is complicated to

implement in global corporationswhose activities are typically scattered across multiple sites in different countries and leverage information

systems in various ways.This is especially true in global

software development where the sitesmay leverage a development tool for totally different purposes. Web-based Virtual Mee- ting Tools (VMT) enable

synchronous communication globallythrough interactive audio, onlinechats, video, and the sharing of

presentations. They provide potentiallya cost effective way to train evencomplex topics to large numbers of

people in global settings. Fewindustrial experiences from the designand use of VMT-based training innovations have been repor- ted. This

paper draws upon a case study in a global corporation to describe thedesign, implementation, and evaluation of a training innovation,consisting of a set of courses delivered by means of a VMT and conferencecalls, to support the global deployment of a Unified Modeling Language(UML) modeling tool and to developUML modeling skills. Evaluation isbased on interviews to verify 1) theimpacts of the innovation on skills,knowledge and motivation, 2)

perceived learner satisfactionwith respect to the innovation.The innovation proved successful

in improving skills, knowledge,and motivation in the caseorganization and learners were

satisfied with it. Other organizations may benefit fromusing VMT to train people to use

similar complex information systems for supporting global software development.

8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 2/11

2012 45th Hawaii International Conference on System Sciences

978-0-7695-4525-7/12 $26.00 © 2012 IEEEDOI 10.1109/HICSS.2012.198DDDDDDDDDDDD

978-0-7695-4525-7/12 $26.00 © 2012 IEEEDOI 10.1109/HICSS.2012.198

3980

8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 3/11

14140024444

complex language, requiring ample, long-term trai-ning and learning efforts [8;17]. UML modelingrequires the use of versatile UML modeling tools thatoffer, for example, graphical editors to enablearchitects, developers, and engineers to model requi-rements, architectures, data structures, dynamic beha-viors, and other characteristics of systems [16]. UMLand the supporting modeling tools constitute a criti-cally important technology (hereafter “UML tech-nology”) for supporting global software develop-ment. This technology, due to its complexity andcomprehensiveness, is a challenging domain for training. It is thus an excellent domain of study todetermine whether VMT are adequate and scalable totrain hundreds or thousands of people to master complex topics in global corporations.

Indeed, the use of UML and UML modeling toolsdo not automatically lead to productivity improve-ments. For example, Dzidek et al. [9] found thatUML is beneficial when developers must extend non-trivial systems with which they are unfamiliar andthat better UML modeling tools and more experien-ced personnel could yield even larger returns on in-vestment. Productivity improvements from theadoption of the UML technology may not be reachedwithout the cost-effective training of end-users.

The case organization had to find a cost effectiveway to improve its employees’ UML technologyrelated skills, knowledge, and motivation globally. Itdecided to use Virtual Meeting Tools for UMLtechnology training. However, the extant literature

provided the organization with little guidance for designing and implementing such training.

Following the problem-centered approach of Peffers et al. [23], this research was initiated in thecase organization to fill the identified gap inknowledge. The research question is as follows:

• Can the UML technology training be organizedand delivered through Virtual Meeting Tools in waysthat learners are satisfied with the training and thetraining positively impacts the skills, knowledge andmotivation of the learners?

To answer the research question, the four-phaseddesign science research methodology presented byPeffers et al. [23] was deployed. First, Problem

Identification and Motivation revealed that the UML

technology related research did not provide anyinsights into the design and implementation of VMTinnovations for UML technology training. Second,Objectives for an Innovation were defined to resolvethe problem based on the experiences of the caseorganization. Third, the key components of theinnovation such as content, organization of training,training materials, and trainers’ skills and knowledgewere Designed and Developed . Fourth, learner

satisfaction and improvements in skills, knowledge,and motivation were Evaluated .

The main contribution of this research is thedesign, implementation, and evaluation of a VMT-

based innovation for UML technology training.Although the design of the innovation has beencreated based on the experiences in the caseorganization, we have made every effort to generalizeit and to identify potential prerequisites for theinnovation, so other organizations can maximallyleverage it in UML technology training.

The paper proceeds as follows. Section “VirtualMeeting Tools, Unified Modeling Language andUnified Modeling Language Tools” introduces basicconcepts related to VMT, UML modeling, andmodeling tools. Section “A Systematic LiteratureReview” presents research related to UML trainingand VMT adoption in training. Section “Descriptionof the case organization and the research method”describes the case organization; the objectives for theinnovation; the research method; and the innovation(i.e., key features of training such as contents,organization of training, training materials, andtrainers skills and knowledge). Section “Preliminaryevaluation of the Virtual Meeting Tool-basedinnovation,” details the results of evaluation. Section“Conclusions and Future Research” concludes the

paper, addresses the limitations of the conductedresearch, and provides an outlook to further research.

2. Virtual Meeting Tools and UnifiedModeling Language Tools

This section explains the concepts of VMT andUML Tool.

2.1. Virtual Meeting Tools

Virtual Meeting Tools enable real-time interac-tions through features such as chat tools and audio,video, and user interface screen sharing. They usecommon browser plug-ins and connect through alocal or remote hosting service [10]. Most VMTs are

platform independent, allowing users on PCs, Macs,and Linux machines to share identical features [10].

At the appointed times, participants log on to join thesessions.

VMT has been used most extensively in educa-tion [10], for example, to arrange remote lectures.But other types of organizations are increasinglyusing VMT for collaboration and training purposes.For example, individuals can use VMT to collaboratein geographically distributed projects. Learning has

become more flexible as VMT has provided more

8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 4/11

4003441421111

opportunities for learning at any place. There is oftena sense of community even if the collaborators arethousand miles away from each other. Without thetime and expense of travel, experts can attend classesfrom any location and respond to the questions of other participants in real time.

Training through VMT has limitations. Mostimportantly, the trainers have reduced control over the virtual class-rooms compared to on-site training.As a result, the trainers have to be highly experiencedin using VMT to interact effectively with their audiences while missing many visual and other cues.

2.2. Unified Modeling Language tools

Some UML tools can generate software fromUML models and UML models from the software.Some also have a built-in knowledge of UML rules,so they can automatically validate the correctness of UML models. Table 1 presents typical high-levelfeatures for the class of UML modeling tools.

3. A Systematic Literature Review

Literature was reviewed to verify to which extentexisting studies cover VMT usage for UML andUML tool training in industrial settings. To improvethe rigor of the study, a systematic literature reviewwas conducted following the principles of Kitchen-ham et al. [15]. VMT related literature is fragmentedand keywords such as e-Learning, online learning,web-based learning, computer-based training, Inter-net-based training, and web-based training are used.We thus decided to use a broader term “training”.UML tool related literature is also fragmented (e.g.,using keywords such as “UML tool” or “CASEtool”), so we decided to use a broader term “UML”.The following criteria and process were used:

1. The first criterion was to find UML trainingrelated articles by searching words “UML” and“training” in title, abstract, and keywords. Decisionwas based on the title and the abstract of the article.

2. The second criterion was to categorize researchaccording to industrial experiences, that is, whether the research reported industrial experiences or not.The content was visited when it was impossible to

determine based on the abstract and the title whether the article reported industrial experiences.

3. The third criterion was to categorize researchaccording to e-Learning, that is, whether the researchreported experiences related to e-Learning or not.The content was visited when it was impossible todetermine based on the abstract and the title whether the article reported e-Learning related experiences.

Table 1. Main features of UML modeling tools(adapted from [16]).

Feature Purpose of the feature is to helpModeling &Diagramming

Create, remove, and edit modelelements and diagrams; view themodels from different perspectives.

HierarchyManagement

Create, update, and delete hierarchiesin which model elements are assigned.

Collaborationand Versionmanagement

Multiple concurrent users to managedifferent versions of assets and toresolve conflicts; integrate the UMLtool to version control and/or changemanagement systems as necessary.

Publishing Compose and publish views of the se-lected models or model elements; pro-vide data in different formats (e.g,JPG); create reports and documents

based on the selected model(elements).

Tracebility Create, remove, update, and tracerelationships between models or model elements.

SimulationandValidation

Simulate dynamic behaviors of models or interface or integrate thetool to simulation tools; validate UMLmodel correctness and completeness.

Model andCode Synch-ronization

Generate code based on models; crea-te models based on code (reverse engi-neering); integrate UML tools to sour-ce code systems, Eclipse, or Model-driven architecture tools such asAndroMDA.

User Management

Manage access and connectivity to theorganization’s directory services (e.g.,Active Directory).

In the first phase of search, IEEE Explore, ACMPortal, and Elsevier’s Science Direct were searched.These databases covered both IS journals andconferences. The number of found articles andrelated references are described in Table 2. Sevenarticles related to UML training were found. Three of them reported UML modeling and/or UML modelingtool training in industrial settings. In addition, Virvouand Tourtoglou [28,29] present two potential systemsto support UML learning but no industrial

experiences were reported. Both Anda et al. [1] andAndersson et al. [2] did not mention usage of any e-Learning tools for UML modeling related training.Bunse et al. [6] reported industrial experiences fromthe design, organization, and execution of a training

program blending e-Learning and face-to-facetraining to teach 42 employees UML in anautomotive branch of a large German corporation.The program started with an online learning phase in

8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 5/11

4004441431111

the form of web-based training, in which the learnersworked self-directed with the courseware to enhancetheir knowledge and skills in applying the UML. The

phase was a prerequisite for the face-to-face trainingsof the second phase. After the face-to-face training, aseveral weeks long coaching phase concluded thetraining program. The coach consulted the learnersabout applying UML in their day-to-day-work. NoUML modeling tool related training was included.

Whenever industrial experiences from UML andUML modeling tool training were reported, theimportance of organizing UML and UML modelingtool training in a cost effective way was clear ([1],[2]). Anda et al. [1] investigated the adoption of UML modeling principles and tools in a projectwhere a global company applied a UML-baseddevelopment method in a large, international projectwith 230 system developers, testers and managers.Adoption was supported by face-to-face training andmentoring. Maximum benefits from UML-baseddevelopment were not achieved because training (1)was not adapted to the needs of the project and (2)was considered too expensive to provide to projectmembers who were not directly involved with UML-

based development. Andersson et al. [2] researchedthe adoption of UML/SysML modeling principlesand tools in an aerospace systems engineering projectat Saab Aerosystems. Introducing UML/SysML witha methodology and a supporting toolset in theorganization required a clear strategy including just-in-time, face-to-face training and mentor support.

Table 2.Results of literature review.

Search criteria Number of articlesfound

Refe-rences

UML training Seven [1,2,6,22,27,28,29]

UML and/or UML tooltraining in industrial settings

Three [1,2,6]

Experiences of e-Learningadoption for UML trainingin industrial settings

One [6]

Relevant articles may have been published but notfound in this literature review. Nevertheless, we

conclude that the application of e-Learning in generaland VMT in particular for UML modeling and/or UML modeling tool training has not received muchattention in the literature.

4. Description of the case organizationand the research methodology

4.1. Case OrganizationA VMT-supported innovation for UML

technology training was implemented in the global

high-technology corporation. To support productdevelopment, a new UML modeling tool was beingrolled out globally. A web-based VMT was provided

by the IT department. It was not used for voicesharing but instead employees could use phone linesor a VoIP application to perform conference callswhile using the VMT.

4.1.1. Introduction of the UML modeling tool inthe case organization

Most of the intended UML tool end-users werefrom the R&D organization. Other organizations suchas partners using the same IT infrastructure were

involved as well. Deployment was supported by ateam consisting of personnel from the global ITdepartment, the department responsible for processand information systems development and supportfor R&D, and the subcontractors working for thesedepartments. The system was intended to graduallyreplace some existing systems and the number of end-users was thus growing.

4.1.2. The need for the Virtual Meeting Tool-based training innovation

The need for a new way of training was noticed based on two surveys conducted in 2009. The middlemanagement responsible for the tool rollout andsupport decided to conduct the surveys to evaluatehow satisfied the end-users were with the tool and thequality of service. The results of the two user satisfaction surveys are presented in detail by [12].The team analyzed the results of the surveys andconcluded that instructions, user guides, and training

practices had to be improved. It initiated severalimprovement activities accordingly. The challengewas that end-users were working in distributed siteswhile at the same time there was pressure to extendthe use of VMT to cut down travelling costs.

The team had previously used VMT ininformation sharing. Face-to-face trainings wereorganized in co-operation with the UML tool vendor which provided globally UML technology trainingand consultancy as well as technical support for their

products. However, the vendor had no experience of VMT-based training. The team decided to design and

pilot an innovation for UML technology training toimprove the overall effectiveness of this training.Since then, the content, material, and organization of

8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 6/11

4005

Name of thesession

Description of the session

Introductionto UML

Main diagrams of UML, history andevolution of UML language

Introductionto tool

Main features of tool, how to getstarted with hands-on example,support resources such as Intranet,guides, and IT support

Classdiagrams

Class diagram in UML and demonstra-tions showing how to create classdiagrams using the tool

Sequencediagrams

Sequence diagram in UML and de-monstrations showing how to createsequence diagrams using the tool

Compositestructurediagrams

Composite structure diagram in UMLand demonstrations showing how tocreate such diagrams using the tool

Statemachinediagrams

State machine diagram in UML anddemonstrations showing how to createstate machine diagrams using the tool

Use Casediagrams

Use case diagram in UML anddemonstrations showing how to create

use case diagrams using the toolIntroductionto collabo-ration

Features to support the collaborativemaintenance of UML models usingtool (presentation and demonstration)

How to publishmodels

Features to support the sharing of mo-dels in different formats or throughIntranet (presentation anddemonstration)))

1441111

trainings have been iteratively improved based onfree-form feedback from end-users. The middlemanagement initiated more formal evaluation during2010 in the form of a survey but the response ratewas unsatisfactory. Interviews were then determinedto be the best method to verify that the innovationwas viable in 2011 after two years of deployment.

4.2. Research Methodology

4.2.1. Design ScienceThe first author of this paper was a member of the

team responsible for UML tool deployment. Designscience research was deemed as the most effectivemethodology for designing the innovation for UMLtechnology training. Design science is a discipline of information systems research which has recently gotample attention among information systemsresearchers. Design science research is relevant to

practitioners as it aims at solving practical andtheoretical problems by creating and evaluating ITartifacts intended to solve identified organizational

problems [19;11;23]. The artifacts are the finalresults of the design process. March and Smith [19]define artifacts as constructs, models, methods andinstantiations. There are several extensions to their list of artifacts. Rossi and Sein [25] include thefollowing artifacts: conceptual designs (e.g.,definition of a relational model), methods (e.g.,design patterns), models and systems (e.g.,

prototypes and commercial applications), and better theories (e.g., relational algebra). Järvinen [13]includes informational and human resources as

potential artifacts, too.The designed innovation for UML technology

training is partly an IT artifact but it also includes hu-man (e.g., trainers, end-users, and their skills,motivations, and stocks of knowledge) and informa-tional resources (e.g., contents of training materials).The case organization experimented with many diffe-rent ways of supporting UML tool end-users. Noother combination of IT artifacts and informationaland human resources was found cost effective by themanagement or appealing by end-users. This paper focuses on the innovation that reflects the onlyeffective combination of the IT artifact (i.e., the VMT

tool) and informational and human resources. Theinnovation is an artifact resulting from the systematicapplication of the design science methodology.

4.2.2. Design of the Virtual Meeting Tool-basedInnovation

The design of the VMT-based training innovationwas a result of two years of development work

between 2009 and 2011. An initial set of training ses-

sions was created and executed in 2009 in co-opera-tion with the UML tool vendor. The latest set of sessions is introduced in Table 3. A set of sessionswas organized typically once every two months. Eachset of sessions was delivered during two weeks, soend-users were able to learn the basics withinreasonable time. Each session was designed to last

between one and two hours, including the timereserved for questions and answers. After eachsession, feedback was asked via e-mail from

participants.Trainers were not experienced in applying VMT

tools for UML technology training when the trainingwas started. They were specialists in both the UMLtechnology and traditional face-to-face training. TheUML tool vendor had to make a substantial effort toinstall and learn to use the VMT tool and theconference call system the case organization hadchosen. The vendor then organized training sessionsin its physical premises and delivered them via VMT,decreasing the traveling costs of trainers.

Table 3.Names and descriptions of thesessions.

8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 7/11

40064414511

Each set of on-line training sessions wasadvertised through email and Intranet pages.Employees registered in the sessions got personalinvitations to the calendar system used in the caseorganization. However, it was possible to join thesessions without registration because the conference

phone number and the link shared during each VMTsession remained the same. This flexibility was wellreceived by employees but the VMT technology didnot provide the team with possibilities to keep track of the employees joining the sessions. The number of trained employees is thus an estimation based on theinvitations sent. 107 employees were invited for thetraining sessions in 2009 and 150 in 2010. Employeescould join the sessions in their offices, in meetingrooms they had reserved, or in other premises, for example, when traveling. They used their mobile

phones, conference phones, or PC software (such asVOIP) for conference calls. Both muted and non-muted lines were reserved for calling purposes. Whencalling to muted lines, they could not make anycomments or questions verbally. However, it was

possible to send questions or comments through chatto the trainers, who checked the questions andcomments and answered them as necessary duringthe sessions.

Training materials were originally developed for the purposes of face-to-face training. Each concept(e.g., a UML diagram or feature) was introduced firstand then the use of the UML modeling tool wasdemonstrated in the same context. Later on thematerials were further developed to better meet thetraining needs when there is no face-to-face contact.

For example, questions were added that trainerscould ask to activate learners remotely. Questionscharted the ways of using the UML technology (e.g.,“Do you use Class Diagram (Yes/No)? Do you findSequence Diagram useful in your work?”) and testedthe learners (e.g., “Which one of the following state-ments is correct?”). It was also possible for the trai-ners to share information during the sessions aboutthe test results and the opinions of learners. Trainingmaterial was available in Intranet for end-users tostudy before, during, or after the training. All thematerials followed the same agreed upon way of

presentation (e.g., all menu options were presented in

italics ).

4.2.3. Methodology for validating the innovationThe qualitative data was collected through seven

interviews after two sets of sessions were organizedduring June 2011 and September 2011. To keep theinterviews informal, semi-structured questions wereused. The interviews were conducted over the phone.They were transcribed to a standard format following

the semi-structured questions and related themes andsent to the interviewees for review. The transcriptswere cross-checked by the research team to capturemisunderstandings and potentially missinginformation. Surveys could not be used for thisresearch despite the substantial number of learners

because the response rates for surveys are very low inthe case organization. Formalized ways of testingimprovements in skills and stocks of knowledge

before and after the training sessions (pre-testing and post-testing) were impossible to deploy as the end-users were located all over the world and there wereno resources available to collect all the necessarydata from them.

To improve the rigor of interviews, the followingstudies were applied when planning the questions:• Koivulahti-Ojala and Käkölä [16] for cate-

gorizing the ways of using UML modeling tools,• Kang and Santhanam [14] and Kraiger et al. [18]

for identifying potential areas for improvementsin skills, stock of knowledge, and motivation,

• Azadeh and Songhori [4] for identifying potential areas of learner satisfaction.

Koivulahti-Ojala and Käkölä [16] proposed that theways of using UML models for communication can

be categorized as follows: human to human, humanto machine, machine to human, and machine tomachine. Kang and Santhanam [14] identified threeknowledge domains that training programs shouldcover: Application knowledge covering commandsand tools embedded in IS applications; business con-text knowledge covering the use of IS applications toeffectively perform business tasks; and collaborative

task knowledge covering the task interdependencies between various actors and how the IS applicationcoordinates and mediates these interdependencies(Table 4). End-users of UML modeling tools need tomaster all the knowledge domains. The businesscontext needs to be mastered because UML modelingtools are general purpose tools applicable to several

business processes. Collaborative task knowledge isvital too because UML modeling tools (possiblyintegrated with other tools) mediate collaborativeactivities in distributed software development. Table4 illustrates UML modeling tool related knowledgeneeds with examples. This study focuses on

application and collaborative training as the trainingsessions supported them. Although Kang andSanthman [14] did not consider motivational aspectsin their study, training can positively affectindividuals’ motivations [18]. Interviews thus chartedalso motivation issues from application andcollaborative task perspective.

Azadeh et al. [4] proposed seven factors thatshould be taken into account when evaluating end-

8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 8/11

400744146

Table 4.The model for training users of UML tools (adapted from [14]).Domain of knowledge Definition Example in UML tool contextApplication Knowledge1) Command based2) Tool-procedural3) Tool-conceptual

1) Commands/keystrokes needed toexecute an operation

2) Knowledge required to combinemultiple commands andcomplete a generic task

3) Knowledge to understand the bigger picture of what to do witha tool

1) Commands/keystrokes in order tocreate a UML element

2) Combine multiple commands tocomplete a UML diagram

3) Which types of diagrams should beused together and when, and howthe tool facilitates this?

Business context knowledge1) Business-procedural2) Business-motivational

1) How to apply the above levels of knowledge to execute a specific

business task?2) What the tool can do for my job?3) What is the role of the tool in the

organization?

1) Which diagrams to apply and whento support a particular business

process (e.g., requirementsmanagement)?

2) Which business processes of theorganization are supported by theUML tool and why?

Collaborative task knowledge1) Task interdependencies2) Collaborative problem

solving approach

1) Interdependencies between tasksand their effects upon using aUML tool

2) Collaborative problem solvingeffort between users

1) How tasks completed through theUML tool affect and are affected byother users of the tool (and/or relatedtools)?

2) Knowledge sharing between users tosolve problems

user training programs from learners’ perspective:relevance of the course to the learner's job,satisfaction with course content and presentation,quality of instruction, effectiveness of the trainer, andoverall satisfaction with the training. Interviewscovered all the factors comprehensively.

5. Preliminary Evaluation of the Virtual

Meeting Tool-based Innovation

Interviewees’ previous knowledge of UMLtechnology varied a lot. Three interviewees hadseveral years of experience of using UML technologyand had used this particular UML tool for more thanone year. Two had applied UML technology but hadused this particular UML tool little or not at all. Twohad very little knowledge of UML technology. All of the interviewees shared models with other employees

but only one used built-in collaborationfunctionalities. They did not use models for communication between humans and machines (e.g.,

code generation or reverse engineering) on a regular basis but some knew such possibilities exist or hadeven tried using them. Interviewees representeddifferent continents (Asia and Europe) and

programmer and architect roles. Most interviewees joined five or more sessions. Those joining less thanfive sessions were more experienced and wanted tolearn specific topics.

5.1. Skills, knowledge and motivation aftertraining

All interviewees were able to name new UMLdiagrams (or semantics related to a particular UMLdiagram) or functionalities they had learned duringon-line training, indicating that their tool-proceduraland tool-conceptual skills had improved. However,

the results varied with respect to learning commandlevel skills. An interviewee with limited previousUML technology knowledge mentioned: “If theapplication is new, you cannot learn everything atone glance.” Learners with limited knowledge maythus be overloaded and unable to follow detailedcommand level instructions. One interviewee hadfound a solution to support his learning of commandlevel skills. He had completed notes during trainingso he could later find the right menus more easily.Another interviewee proposed that training sessionsshould be recorded so the instructions can bereviewed whenever necessary. We can conclude that

learners were able to find their ways to learncommand level skills over time with the help of on-line training.

Interviewees were not able to name any conceptsor practices (e.g., collaborative maintenance of models) related to collaborative task knowledge after training. Only one of the interviewees usedcollaborative modeling and it can be expected thatinterviewees focused on those sessions they

8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 9/11

4008

Domain for learner satisfaction

Example answers

Content “Content was a compact packet.”“I think content was good and some

good examples were presented. ““Potentially it could go into moredetails and more advanced things.”

Demonstra-tions

“Demonstrations were clear and presented smoothly.” “Good to havethe sessions on UML modeling and theuse of the tool after each other.”“When showing how to make menuselections, the trainer should pause andshow the selections slowly.”

Trainingmaterial

“Material was ok”. “I have read thosematerials I need for the three types of diagrams I use.” “Good enoughwhatever there was during those twosessions I attended.” “Material doesnot support finding information.Searching and linking capabilitieswould be improvements.”

Trainer “Trainer knew the material and theUML tool.” “Trainer was fluent inEnglish and knew what he was doing.”“He did know his topic and was clear

presenting it.” “Content was good butsometimes too difficult to follow due tofast speed.”

Voice “No problems.” “Mostly ok.”Means of

presentation“No problems”. “Surprisingly good.Only one small break.”

Questionsfor thetrainer

“It was good to have a chance to makequestions. Trainer answered them

promptly.”Questionsfor other interviewees

“Most learners only listened. As far as Iremember, one person in two sessionsasked something.”

The way theon-linetrainingsessionswere

organized

“A full day session is difficult to alloca-te nowadays. This [short session] wasgood for me... I would probably miss itif it were a longer face-to-face or on-line session. On-line sessions are diffi-

cult to follow if they last several hours.Face-to-face trainings need full dayallocations and negotiations with themanager.” “Length of sessions was ok”.

considered most relevant to their immediate needs.Thus it cannot be concluded that on-line training isunsuitable for learning collaborative knowledge.Instead, lack of such knowledge after trainingindicates that interviewees lacked motivation to learnsuch knowledge. Some interviewees stated that theyhad not started to use collaborative modeling andtherefore had now skipped the related session butwere interested to join such a session later.

Most interviewees agreed that they were moremotivated to use the UML tool after training. For oneuser, the usage of the tool was compulsory and heindicated that training neither increased nor decreased his motivation. Another user had a longexperience of UML technology and his expectationsfor the course were learning business-proceduralskills and knowledge rather than application levelskills. On the other hand, one experienced user indicated increased motivation due to the possibilityto refresh his UML technology knowledge. Our

preliminary conclusion is that those using the toolvoluntarily and joining sessions to learn applicationlevel skills were more motivated after the training.Interviewees did not express increased motivations tosolve UML technology related problems with other end-users after the training. They mentioned their own teams, Intranet, and Internet as the sources theywould use to solve the problems. Training thusimproved or maintained motivation at the applicationlevel but not at the collaborative level.

In sum, the innovation for UML technology trai-ning improved application related skills and know-ledge and increased or maintained the motivation to

apply UML technology. However, improvements incommand level skills and collaborative task knowledge and motivation were limited.

5.2. Learner Satisfaction

Interviewees were satisfied with content, trainingmaterial, voice, presentation sharing, and the waylearning was organized (Table 5). As the content andtraining materials had been specifically tailored for on-line training of application and collaborative task knowledge during the previous two years, it is

possible that the interviewees did not see any major

improvement proposals necessary. The proposals for new content came mainly from the users having mostextensive previous knowledge. They indicated needsfor training either business context related knowledgeor very detailed additional knowledge. But additionaldetails might neither be interesting nor useful for novices. Accordingly, the scope of using the VMTinnovation must be extended for training businesscontext knowledge in future.

Table 5.Examples of answers for learner satisfaction.

One interviewee proposed that it should be possibleto find information in the material more easily. Thematerial was tailored for training purposes and didnot support the searching of particular pieces of

8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 10/11

4009

information. As learners were not able to fully learncommand level skills during training sessions, thematerial should support the searching of relevantcontent after training.

The organization of the training sessions got some positive remarks. UML was always introduced firstand the use of the tool was focused on after that. Thiscombination of tool-conceptual and tool-proceduraltraining was seen beneficial. In addition, the lengthsof the sessions were suitable both from practical andlearning perspectives (see the last row of Table 4).

No interviewees mentioned other types of diagramsthat should also be covered in training sessions butsome detailed proposals for other topics werementioned (e.g., how to move elements in ahierarchical model).

Interviewees were mostly satisfied withdemonstrations and the trainer but they agreed that

presentation speed was sometimes too fast. This isunderstandable as the trainer could not see learners’reactions and adapt the speed as necessary. On-linetraining thus requires paying special attention to

presentation pace.Interviewees were familiar with VMT and confe-

rence calling. They were satisfied with voice and pre-sentation sharing but stated that sometimes PC appli-cations used for presentation sharing or conferencecalling were not working properly. However,interviewees knew from their earlier experiences thatsuch incidents happen from time to time. This mayexplain why the incidents did not decrease their

perceived satisfaction. When the studied sets of sessions were organized, all trainers had previous

knowledge of applying VMT and voice sharing for training. In organizations where trainers or learnerslack similar VMT skills and knowledge, learner satisfaction may be lower than in this organization.

5.3. Generalizable findings from theevaluation

The evaluated innovation for UML technologytraining is based on experiences from oneorganization during the period of two years.However, it is possible to make some generalrecommendations because both the innovation and

the related informational and human resources have been specified. VMT can be applied for complextechnology training successfully (in terms of learner satisfaction, sustained motivation to use thetechnology, and improved tool-conceptual and tool-

procedural skills and knowledge) in organizationswhere end-users are familiar with VMT and there aretrainers experienced in conducting customized on-line training using the innovation. Organizations,

searching for a viable solution for training largenumbers of globally distributed employees to usecomplex software technologies, should thus carefullyanalyze both employees’ and trainers’ abilities to useVMT and conference calls.

Subjective opinions of interviewees do notnecessarily correlate with real improvements in skillsand knowledge or learner satisfaction. However,other data sources within the organization support theinterview results. First, a user satisfaction surveycompleted in the organization indicated that after theUML technology training sessions were initiated,user satisfaction was increased (see details in [12]).Second, the case organization tried out other ways of supporting end-users’ efforts to learn UMLtechnology but they were unsuccessful in terms of

popularity amongst the end-users.It should also be noted that in the case organiza-

tion, both business context knowledge creation andcollaborative task knowledge creation were alsosupported by other means. Business contextknowledge creation was supported by UMLtechnology experts who joined deployment projectswhere teams or projects took the UML tool into use.Experts suggested suitable diagrams, structured themodels, and provided tailored training for team-,

project-, and department-specific purposes.Collaborative task knowledge creation was enhanced

by finding and training contact persons for each team, project, and department, and encouraging the sharingof experiences in user forums.

6. Conclusions and Future Research

This research described an innovation for UMLtechnology training that results from a few years of iterative development of the case organization,content, material, and trainers’ skills and knowledge.It was found that VMT can be applied for training

people to use complex technologies successfully (interms of learner satisfaction and motivation andknowledge to use the technology) in organizationswhere end-users routinely use VMT and there aretrainers experienced in on-line training. Informationsystems professionals benefit from the proposedinnovation for UML training when planning,

implementing, and evaluating UML training sessionsorganized through VMT. Information systemsmanagement can take advantage of the results whenmaking decisions about VMT usage in complextechnology training.

The single case study methodology may not pro-vide a sound basis for generalization. Future researchin other organizations is necessary to probe theapplicability of VMT in training people to use

8/2/2019 uml based

http://slidepdf.com/reader/full/uml-based 11/11

4010

especially nontrivial information systems. The UMLtechnology is considered to be complex and difficultto learn. This study indicates that it is possible tosupport the learning of complex technologies throughVMT by structuring the complex content in anappropriate way from the end-users’ perspectives.

7. References

[1] Anda, B., Hansen, K., Gullesen, I. and Thorsen, H.(2006). Experiences from introducing UML-baseddevelopment in a large safety-critical project. EmpiricalSoftware Engineering, 11(4), 555-581.[2] Andersson, H., Herzog, E., Johansson, G. andJohansson, O. (2010). Experience from introducing unifiedmodeling language/systems modeling language at SaabAerosystems. Systems Engineering, 13(4), 369-380.[3] Attewell, P. (1992). Technology diffusion andorganizational learning: The case of business computing.Organization Science, 3(1), 1-19.[4] Azadeh, A. and Songhori, M. (2006). End-user training

programs planning model based on InformationTechnology and Information Systems (IT/IS) impact onindividual work. In Proceedings of the IEEE IndustrialTechnology, 2107 -2112.[5] Bostrom, R., Olfman, L. and Sein, M. (1990). Theimportance of learning style in end-user training. MISQuarterly, 14(1), 101-119.[6] Bunse, C., Grutzner, I., Peper, C., Steinbach-Nordmann,S. and Vollmers, C. (2006). Coaching professional softwaredevelopers-an experience report. In Proceedings of the 19thConference on Software Engineering Education andTraining, IEEE, 123-130.[7] Compeau, D., Olfman, L., Sein, M. and Webster, J.(1995). End-user training and learning. Communications of the ACM, 38(7), 24-26.[8] Dori, D. (2002). Why significant UML change isunlikely. Communications of the ACM, 45(11), 82-85.[9] Dzidek, W.J., Arisholm, E. and Briand, L.C. (2008). ARealistic Empirical Evaluation of the Costs and Benefits of UML in Software Maintenance. IEEE Transactions onSoftware Engineering, 34(3), 407-432.[10] ELI (2006). 7 things you should know about virtualmeetings, Educause Learning Initiatives.http :// co nn ec t. ed u ca u s e . e du/ L ibr a r y / E L I /7 T hin g sY ouShouldKnowAbout/39388[11] Hevner, A., March, S., Park, J. and Ram, S. (2004).Design science in information systems research. MISQuarterly 28(1), 75-105.[12] Islam, N.A.K.M., Koivulahti-Ojala, M., and Käkölä,

T. (2010). A lightweight, industrially-validated instrumentto measure user satisfaction and service quality experienced by the users of a UML modeling tool. In Proceedings of theAMCIS 2010.[13] Järvinen P. (2004). On research methods. Opinpajankirja, Tampere.[14] Kang, D. and Santhanam, R. (2003). A longitudinalfield study of training practices in a collaborativeapplication environment. Journal of ManagementInformation Systems 20(3), 257-281.

[15] Kitchenham, B., Pearl Brereton, O., Budgen, D.,Turner, M., Bailey, J. and Linkman, S. (2009). Systematicliterature reviews in software engineering-A systematicliterature review. Information and Software Technology,51(1), 7-15.[16] Koivulahti-Ojala, M. and Käkölä, T. (2010).Framework for Evaluating the Version Management

Capabilities of a Class of UML Modeling Tools from theViewpoint of Multi-Site, Multi-Partner Product LineOrganizations. In Proceedings of the 43rd HawaiiInternational Conference on Systems Sciences (HICSS-43).IEEE. 1-10.[17] Kobryn, C. (2002). Will UML 2.0 be agile or awkward? Communications of the ACM, 45(1), 107-110.[18] Kraiger, K., Ford, J. K. and Salas, E. (1993).Application of cognitive, skill-based, and affective theoriesof learning outcomes to new methods of trainingevaluation. Journal of Applied Psychology, 78(2), 311-328.[19] March, S. and Smith, G. (1995). Design and naturalscience research on information technology. DecisionSupport Systems 15(4), 251-266.[20] Nelson, R., Whitener, E. and Philcox, H. (1995). The

assessment of end-user training needs. Communications of the ACM 38(7), 27-39.[21] Object Management Group (2009). Unified ModelingLanguage: Superstructure. Formal Specification, version2.2, 2009.[22] Pavlov, V. L. and Yatsenko, A. (2005). UsingPantomime in Teaching OOA & OOD with UML. InProceedings of the 18th Conference on SoftwareEngineering Education & Training, IEEE, 77 -84.[23] Peffers, K., Tuunanen, T., Rothenberger, M. andChatterjee, S. (2007). A design science researchmethodology for information systems research. Journal of Management Information Systems 24(3), 45-77.[24] Piccoli, G, Ahmad, R., and Ives, B. (2001). Web-basedvirtual learning environments: A research framework and a

preliminary assessment of effectiveness in basic IT skillstraining. MIS Quarterly, 25(4), 401-426.[25] Rossi, M. and Sein, M. (2003). Design ResearchWorkshop: A Proactive Research Approach. Presentationdelivered at IRIS 26, August 9 – 12, 2003 .h tt p ://tiesrv. h kkk. f i/iris 26 /p rese n ta ti on /w o rks ho p _d e sig n Res.pdf last accessed January 16, 2004.[26] Sabherwal, R., Jeyaraj, A. and Chowa, C. (2006).Information system success: individual and organizationaldeterminants. Management Science 52(12), 1849-1864.[27] Turner, S., Perez-Quinones, M. and Edwards, S.(2005). minimUML: A minimalist approach to UMLdiagramming for early computer science education. Journalon Educational Resources in Computing (JERIC), 5(4), 1-28.[28] Virvou, M. and Tourtoglou, K. (2006). An AdaptiveTraining Environment for UML. In Proceedings of theSixth International Conference on Advanced LearningTechnologies, IEEE, 147 -149.[29] Virvou, M. and Tourtoglou, K. (2006). IntelligentHelp for Managing and Training UML SoftwareEngineering Teams. In Proceedings of the Seventh JointConference on Knowledge-Based Software Engineering.IOS Press, 11-20.