ARIADNE V4 Filip Neven, Stefaan Ternier & Erik Duval Dept. Computerwetenschappen, Katholieke...

22
ARIADNE V4 Filip Neven, Stefaan Ternier & Erik Duval Dept. Computerwetenschappen, Katholieke Universiteit Leuven, Belgium E-mail: [email protected]

Transcript of ARIADNE V4 Filip Neven, Stefaan Ternier & Erik Duval Dept. Computerwetenschappen, Katholieke...

ARIADNE V4

Filip Neven, Stefaan Ternier & Erik Duval

Dept. Computerwetenschappen, Katholieke Universiteit Leuven, Belgium

E-mail: [email protected]

Ariadne V3 architecture• Access to database through Query Tool (TM5)

• Maintenance is difficult – Database structure is hardcoded at many places

– Duplication of functionality (TM5, replication)

• Installation is difficult– Different tools have to be installed seperately

KPS

WEBLE

Query Tool (TM5)

TM6

Replication

...Document Repository

A new V4 architecture: Why?• Programmers point of view

– Old 2-tier architecture is difficult to maintain– Exploit KPS as a component

• Users point of view– End users

• Want easy user inferface (integrated in the tool they use)

• Automatic metadata generation

– Application developers• Technical users (e.g. Blackboard)

• Provide easy solution for accessing the knowledge pool

• Blackboard, SILOGIC, ...

Ariadne V4 architecture

KPS Client

SOAP Interface

Query Tool

SQL

SOAP API

JAVA API

KPSDocument Repository

Ariadne V4 packages

KPS Client

SOAP Interface

Query Tool

KPS client development

package

Query tool package (TM5)

KPSDocument Repository

KPS package

Ariadne V4 packages

KPS

SOAP Interface

KPS Client

Query Tool

KPS Client

Blackboard

Your (non java)application

KPS Client

SILOGIC

Document Repository

Soap Interface• XML technology

– Easy processing of metadata

– Validation of metadata

– LOM export

• Hide complexity of database for query tool– Query Service

• Keyword Query

• Advanced Query

• Get Results

– Insert Service• Insert

• Delete

• Update

– Download Service

KPS Client Application Development Kit• Offers Java API that talks with SOAP interface

– Query, Insert, Update, ...

– methods to handle metadata

Towards a critical mass• Share and reuse -> critical mass

– Content

– Users

• Why don’t we reach this critical mass ?

Indexation process takes

too long

# new documents stable

Low number of query results

Motivation to index new material low

Towards a critical mass• Share and reuse -> critical mass

– Content

– Users

• Why don’t we reach this critical mass ?

Indexation process takes

too long

# new documents stable

Low number of query results

Motivation to index new material low

Towards a critical mass• Share and reuse -> critical mass

– Content

– Users

• Why don’t we reach this critical mass ?

Indexation process takes

too long

# new documents stable

Low number of query results

Motivation to index new material low

V4 Indexation- and Query Tool• Simplify the indexation procedure

– Profiles

– Automatic metadata generation

– Less = more• E.g. Hide concept management

V4 Indexation- and Query Tool• Integration in existing infrastructure

– Institutions have their own house-style• Allow to customize the look-and-feel

– Institutions have specific metadata requirements• Allow to define new metadata fields

– E.g. FEICO -> Product Line

V4 Indexation- and Query Tool

KPS

KPS Client

SOAP Interface

Query Tool

Query Tool

AriadneMetadata

InsertAdvanced

Query

XML Configuration files

V4 Indexation- and Query Tool

KPS

KPS Client

SOAP Interface

Query Tool

Query Tool

AriadneMetadata

InsertAdvanced

Query

XML Configuration files

Java layer

GUIXML

V4 Indexation- and Query Tool

KPS

KPS Client

SOAP Interface

Query Tool

Query Tool

AriadneMetadata

InsertAdvanced

Query

XML Configuration files

Java layer

GUIXML

XSLT ProcessorLook and feel(Stylesheet)

HTML

V4 Indexation- and Query Tool

<insertpage >

<section name =" general " caption ="General Info ">

<field name =" title " type ="input " caption ="Document Title " size ="60" multivalued =" true “

multilingual ="false " addcaption ="Add "/>

<field name ="publdate " type =" date " caption ="Publication Date "/>

<field name =" restrictions " type =" list " caption =" Restrictions " listsource ="restrictions "/>

<field name ="usagerights " type ="list " caption ="Usage Rights " listsource ="usagerights "/>

<field name ="author " type =" hierarchical " caption ="Author(s) " multivalued =" true"

addcaption =" Add ">

<field name =" lastname " type ="input " caption =" Last Name " size ="60"/>

<field name =" firstname " type =" input " caption =" First Name " size ="60"/>

</field >

<field name ="description " type =" input " caption ="Description " size ="60"/>

</section >

</insertpage >

V4 Indexation- and Query tool

Demo

Blackboard• Own document repository with a (poor & optional)

description of LO’s

• Building Block program– Offers an API to third parties to extend functionality of Blackboard

– Their way of doing “open source”

Blackboard

User Support

http://www.cs.kuleuven.ac.be/cwis/research/hmdb/bugzilla/index.cgi

Status & planning• Almost finished

– Web services layer • All functionality is implemented

• Testing phase

– Indexation- and Query tool• Most functionality is ready

• Todo– Advanced query

– Support of multiple languages

– ...