ENABLING CSCW SYSTEMS TO AUTOMATICALLY BIND

EXTERNAL KNOWLEDGE BASES

Thomas Bopp and Jonas Schulte

Heinz Nixdorf Institute, University of Paderborn, F

urstenallee 11, 31102 Paderborn, Germany

Thorsten Hampel

Department of Knowledge and Business Engineering, University of Vienna, Rathausstr. 19/9, A-1010 Vienna, Austria

Keywords:

Digital Library, Integrated Search, Distributed Knowledge Spaces, CSCW, Automatic Binding.

Abstract:

The usage of CSCW systems for teaching, training, and research collaboration increases constantly, as it offers

a time- and place-independent, as well as a cost-effective platform. The user’s search should not be restricted

to local material; in fact, users beneﬁt from different search environments, as, for example digital libraries to

ﬁnd appropriate working material. Searching and further processing of documents imply a media breach since

the search cannot be invoked in current CSCW systems directly. This paper presents the ﬁrst prototype of a

CSCW system which enables users to search in external sources without media breach. To provide arbitrary

search environments no restrictions to data formats or search functionalities are allowed. Hence we have

enhanced search environments with self description capabilities in order to realize an automatic binding of

search environments in CSCW systems. By search environments we address any system offering searchable

knowledge bases, such as digital libraries or the CSCW system itself. Furthermore our concept supports local

search and searching in different external sources at the same time.

1 INTRODUCTION

CSCW (Computer Supported Cooperative Work) sys-

tems are an efﬁcient solution to establish cooperative

work independent from the local places of partici-

pants. Their ﬁelds of application are not limited to

universities, as it is also beneﬁcial to the industry in

regard to employee training. In many applications, as

for example structuring cooperative work processes

or ﬁeld of employee trainings, it is essential to have a

reliable database to work with. Usually, not all data

needed for arbitrary trainings and research/teaching

activities is already contained within the CSCW sys-

tem. Accordingly, users search for information in the

internet as well as in digital libraries. The credibil-

ity of external information is of crucial importance,

which is often not given by freely accessible data from

the web.

In contrast the credibility of information avail-

able from digital libraries is generally undoubted and

therefore digital libraries are an excellent data source

for reliable documents, which furthermore can be ac-

cessed independently from time and place. Enabling

publication of and access to information in a digital

format has already been realized by digital libraries

such as DSpace(Smith, 2002), Fedora (Payette and

Lagoze, 1998), and DuEPublico (Gollan et al., 1999).

There is a signiﬁcant trend of providing access to in-

formation in a digital way which enables an easy fur-

ther processing as well as time and place independent

utilization.

The goal of our approach is to develop mecha-

nisms which allow for a ﬂexible integration of dif-

ferent knowledge sources into any collaborative sys-

tems. In this context, the new generation of web ser-

vices are of central importance in regard to our mistel-

approach. In contemporary digital libraries the search

is often invoked by users via a web-interface. To uti-

lize material from digital libraries in CSCW systems,

there are currently two steps to be effected: First of

all, the user has to search via the web-interface of

the digital library. Afterwards, the appropriate results

are added into the content of the CSCW system. In

most cases, this requires to buffer the ﬁles from the

result page of the digital library and then to upload

them to the CSCW system. It is obvious that this

323

Bopp T., Schulte J. and Hampel T. (2007).

ENABLING CSCW SYSTEMS TO AUTOMATICALLY BIND EXTERNAL KNOWLEDGE BASES.

In Proceedings of the Ninth International Conference on Enterprise Information Systems - DISI, pages 323-329

DOI: 10.5220/0002382603230329

 SciTePress

process is not only inefﬁcient and not user-friendly,

but also slows down the establishment of CSCW sys-

tems and digital libraries. However, performing these

steps is the only possible way at the moment to use

data from digital libraries in CSCW systems. Our ob-

jective is to avoid this media breach between search-

ing in external sources and using founded material in

CSCW systems in order to boost the establishment

of both platforms as well as their utilization. A so-

lution without media breach would require the user

being able to invoke search functionalities of digital

libraries within the CSCW system. To enable other

systems to use their search functionalities, some digi-

tal libraries already provide appropriate interfaces by

deploying web services (Petinot et al., 2004; Gollan

et al., 1999). However, no contemporary CSCW sys-

tem beneﬁts from this possibility.

In this paper we address the usage and search of

external information in CSCW systems without me-

dia breach. We not only describe the integration of

external search functionalities in a CSCW system, in

fact it is obvious that an automatic binding of digital

libraries/repositories in CSCW systems is deﬁnitely

more beneﬁcial than a manual integration. To en-

able a complete automatic binding process, several

requirements for the search environments have been

identiﬁed. As an example implementation we have

integrated the DuEPublico binding in the CSCW sys-

tem open sTeam (Bopp et al., 2006b). Here, a de-

tailed view of the realized concept is presented in or-

der to point out the advantages of an integrated search

functionality. Our approach is not limited to the men-

tioned demonstrational systems. Instead, it is a open

web service implementation which can be added eas-

ily to any collaborative system and digital library.

Our architecture combines local and external

search resources offering extended search capabili-

ties to the user. In this regard, we have identiﬁed a

number of scenarios (see section 1.1): A user might

search for a search term in all possible resources or

explicitly choose an external resource. Regarding the

concrete example of the object oriented open sTeam

CSCW system, external and local search results are

usually imported into knowledge spaces and can be

arranged using the room and spatial metaphors in or-

der to make appropriate search results available for

other CSCW users.

This paper is organized as follows: related work

is presented in section 2, while the description of our

developments is divided according to requirements in

section 3. Next, a detailed view of the system inter-

action is given in section 4. Before concluding our

results the adjusted concept for searching in parallel

in different search environments is explained in sec-

tion 5.

1.1 Motivating Scenarios

In this section three different scenarios are presented

which point out the various search requirements for

a user-friendly CSCW system. Flexible search func-

tionalities as the ones described in this paper, are

quite new to common CSCW systems. Our analysis

presents a concept to realize these three scenarios in

CSCW systems mentioned below. A basic framework

and ﬁrst prototypes developed on its basis will be in-

troduced, which support searching according to the

different scenarios and offer to invoke external search

functionality of one digital library.

Scenario 1: Local Search

A user searches for local objects, as, for exam-

ple a speciﬁc document, group, user, or knowledge

space. If the user searches for a certain group to re-

quest membership, he already knows the local avail-

ability of the group and searching in local resources

is sufﬁcient.

Scenario 2: Literature Research

This scenario describes the research for litera-

ture in digital libraries. A user searches for gen-

eral keywords or more speciﬁc meta data such as

authors, conferences, or book titles. Digital li-

braries/repositories offer a large number of articles

and provide rich meta data for the entries. Usually,

not only a single repository is searched by the user,

but also literature in the established digital libraries

of ACM, IEEE, and CiteSeer are scanned for results.

Therefore, within a CSCW system the user would in

this case choose several external resources to which

the search request should be sent.

Scenario 3: General Searching

A user might have no precise idea of where to ﬁnd

something. Therefore, he searches without further

speciﬁcations all available resources, which possibly

might lead to a large number of search results.

2 RELATED WORK

Digital libraries such as DSpace (Smith, 2002), Fe-

dora (Payette and Lagoze, 1998), and DuEPublico

(Gollan et al., 1999) provide digital access to their

inventories, as well as web services to invoke their

search functionalities remotely. To enable searching,

it is necessary to identify and collect metadata from

documents.

The Open Archives Initiative provides the OAI-

PMH (Lagoze and de Sompel, 2001) standard for

ICEIS 2007 - International Conference on Enterprise Information Systems

324

harvesting metadata of documents. This approach is

meant to be a low-barrier solution with roots in en-

hancing access to e-print archives. The need for a

common metadata format has been identiﬁed before

and resulted in the development of the Dublin Core

Metadata Element Set (Weibel, 1998). This set be-

came an ISO standard and most digital libraries like

DSpace, Fedora, and DuEPublico, provide the data

using Dublin Core (DC) elements. The prototype pre-

sented in this paper bases on the awareness that arbi-

trary search environments do not provide an uniform

metadata format. Accordingly, an automatic bind-

ing requires a search environment’s ability to describe

their functionalities and formats on their own.

Since the DC elements are designated to be used

for metadata description, it lacks description terms for

service functionalities and for additional information

required to perform searches. Therefore, the ZeeRex

format (ZeeRex, 2004) can be used to describe ser-

vice functionalities. In the process of an automatic

binding the self description method of a search envi-

ronment returns a ZeeRex document, which is eval-

uated from the CSCW system in order to enable the

user to invoke the search functionalities without me-

dia breach. There is no contemporary CSCW system

which offers the search in external sources without

media breach. The developed prototype extends the

open sTeam (Bopp et al., 2006b) CSCW system.

Another approach to standardize the access of

search functionalities is the standard search protocol

Search/Retrieve Web Service (SRW, 2005), which in-

cludes the search, scan, and explain operations. How-

ever, this standard is not (fully) supported by arbitrary

search environments and rather does not deﬁne the ex-

change formats of the metadata. That is why even if

SRW is supported, an automatic binding requires the

search environment’s capability of self description.

3 REQUIREMENTS FOR

AUTOMATIC BINDING

In this section the requirement analysis for automat-

ically integrating search functionalities of external

search environments into a CSCW system is pre-

sented. Since the requirements result from the process

of using external search functionalities, section 3.1

explains the workﬂow of the search invoking. The

resulting requirements for the search web service are

described in section 3.2.

Our recent conceptual approach in the ﬁeld of Com-

puter Supported Cooperative Work and Learning

(CSCW/L) has been to build cooperative knowledge

spaces. One resulting implementation from this ap-

proach is the client/server system called open sTeam

(Bopp et al., 2006b). The idea of this system is

to combine document management facilities with a

room metaphor in collaborative knowledge spaces.

This concept is especially advantageous since one

user can perform a search, set results in relation with

already existing documents, and make them available

for other users in the distributed knowledge space.

Our development in the CSCW system open

sTeam provides access to search functionalities of ar-

bitrary digital libraries and additionally, of any search

environment which deploys adequate web services.

The identiﬁed requirements are independent from the

CSCW system, which should interact with the search

environment. Hence, our results can be transferred to

extend other CSCW systems in a similar manner.

3.1 General Search Process

Distributed knowledge spaces are containers for in-

formation from various external data sources. Ac-

cordingly, the content stored in the CSCW system

bases on external information and the search func-

tionalities usable in open sTeam should not be limited

to one (external) search environment. In our solution

the user can select several search environments the

web services of which are invoked, which conforms to

the principle of meta search engines. Hence, the ﬁrst

step for the user is to select the search environments

to which his or her search request is sent. As a sec-

ond step, the user speciﬁes his or her search request

in an adequate search dialog in the web-interface of

open sTeam. Please note that the presented search

dialog depends on the selected search environments.

Search dialogs for several search environments can

differ since they may support different indexed search

ﬁelds like keywords, author, title, etc. After the user

has submitted the search dialog, open sTeam gener-

ates search requests conform with the selected en-

vironments. The web services are invoked and the

results are presented in the web-interface of open

sTeam. Since the results may be retrieved from dif-

ferent environments, an aggregation of the result hits

is required (Bopp et al., 2006a). A result page pro-

vides the opportunity to add the found documents di-

rectly to a distributed knowledge space for sharing

them with other participants. The system therefore

offers not only a standardized frontend to different

knowledge sources, but allows the ﬂexible integration

of new knowledge source in various ways.

ENABLING CSCW SYSTEMS TO AUTOMATICALLY BIND EXTERNAL KNOWLEDGE BASES

325

3.2 Functionalities of Search Web

Services

To provide highest ﬂexibility, it is essential to cou-

ple new search environments with the CSCW system

without additional effort. However, each search envi-

ronment can arbitrarily deﬁne the supported indexed

search ﬁelds and conﬁguration options. Hence, it is

necessary that web services from a search environ-

ment (web service provider) are able to explain them-

selves. Only if explanation methods are supported,

CSCW systems as web service consumers can con-

ﬁgure themselves automatically, build dynamically

search dialogs, and automatically generate adequate

search requests. The new approach of our develop-

ment is to divide the usage of a search web service

into the following three steps: First of all a connec-

tion with an optional authentication process has to

be established. Afterwards the CSCW system and

the search environment negotiated the interaction pro-

cess, which requires the search web service to explain

its functionalities. Finally, the CSCW system is able

to use the underlying web service. Figure 1 illustrates

this process.

Connection establishment

Usage of the service

Negotiation

process

Figure 1: Usage of Search Web Services.

In this section we focus on step 2 - the negotiation

process between the CSCW system and the search

environment. To provide highest usability of the de-

scription of the search web service’s functionalities a

simple and customizable exchange format is required.

Therefore we use the ZeeRex format (ZeeRex, 2004)

which extends the Z39.50 standard maintained by the

Library of Congress. ZeeRex is an attempt to de-

ﬁne an ”Explain”-like mechanism for search environ-

ments. If a digital library supports this explain mecha-

nism, a CSCW system can invoke the explain method

to retrieve an XML document that describes in details

the search environment which is required for an au-

tomatic binding. According to the ZeeRex DTD the

XML document for explaining a search environment

has the following structure:

</explain>

As described above, a ZeeRex document con-

tains, at most ﬁve sections surrounded by an

explain

tag. The

serverInfo

section contains basic infor-

mation required for connecting to the database such

as protocol, host, port, and database name. We

focus on the

indexInfo

and the

schemaInfo

sec-

tions. The

schemaInfo

section refers to available

XML schemata useable for returning results. Know-

ing the format of the results is necessary to transform

them for the ﬁnal presentation in open sTeam. In the

indexInfo

section(s) all usable ways in which the

server may be interrogated are recorded. These are

described by using

<index>

elements. The

<index>

element is a representation of a single type of search.

For example, a title search is one index and an au-

thor search is another index. Each phrase of a search

request belongs to exactly one search index. Further-

more, the search request may only contain at most one

phrase for one search index.

We have developed an example ZeeRex document

explaining the DuEPublico search environment which

is presented in section 3.2.1. In section 4 we describe

how to evaluate ZeeRex documents and particularly

<index>

elements to automatically build search di-

alogs in CSCW systems. Please note that the explain

functionality is not yet supported by DuEPublico. In

our prototype the information has to be integrated

manually. To enable CSCW systems to bind search

environments completely automatic, it is mandatory

to support the explain mechanism.

3.2.1 Zeerex Description of Duepublico

As described previously, the focus of this section is on

interpreting the

parts of the ZeeRex de-

scription, since they present the possible search types.

DuEPublico search environments offer a plurality of

search indexes and selection ﬁelds, which can be used

in a search request. Furthermore, conﬁguration ﬁelds

enable the customization of the result presentation.

The search web service supports the same functional-

ities as searching via the DuEPublico web-interface.

For describing the indexed search ﬁelds ”title” and

”author” the explain document contains the following

<index>

element.

<title>Document Title</title>

ICEIS 2007 - International Conference on Enterprise Information Systems

326

<map>

<name set="dc">title</name>

</map>

<index>

<title>Author/Creator</title>

<name set="dc">author</name>

</map>

<map>

<name set="dc">creator</name>

</map>

<index>

</indexInfo>

The indexed search ﬁeld title can be used in search

requests and is one of the possible sort criterions. Be-

cause of these supported functionalities for the search

ﬁeld, the

<index>

element has the attributes

and

sort

. The

<title>

element describes the search

index, which is here the document title. Search re-

quests sent to the web service have to be conform

with one of the supported format of the search en-

vironment. Each search index is mapped to a spe-

ciﬁc term in this format. By enabling search environ-

ments to support various request formats, more than

one

<map>

element can be deﬁned. The relation of the

<map>

element to one format is given by its attribute

set

. In the example mentioned above DuEPublico

accepts only search requests in the Dublin Core (DC)

format (Weibel, 1998). Please note, as the example

show it for the author, different Dublin Core elements

can be mapped to the same search index.

The

<index>

elements for selection ﬁelds and the

conﬁguration ﬁelds should be described similar to the

search ﬁelds. The conﬁguration ﬁelds determine con-

straints for the reply of the search web service, as for

example, the maximum number of results. It is com-

mon practice to characterize materials in a library’s

inventory by classiﬁcations. A media type of the ma-

terial, for example, is determined like animation, au-

dio ﬁle, ﬁgure, text, etc. The ZeeRex DTD lacks de-

scribing selection and conﬁguration ﬁelds. That is

why we extend the ZeeRex document as follows:

<explain xmlns:sEnv=

"http://systemkonvergenz.de/ZeeRexExtension">

...

<title>Media Type</title>

<map>

<name set="dc">format</name>

<sEnv:method>

<sEnv:name>

doListClassification

</sEnv:name>

<sEnv:params>

format

</sEnv:params>

</sEnv:method>

</map>

</index>

<title>Maximum Hits</title>

<map>

<name set="conf">maxResults</name>

<sEnv:selection>

<sEnv:item>50</sEnv:item>

<sEnv:item>200</sEnv:item>

<sEnv:item>1000</sEnv:item>

</sEnv:selection>

</map>

</index>

...

</indexInfo>

...

</explain>

The most important aspect to use selection ﬁelds is

the need to read out the selectable data dynamically.

To support automatic binding the method responsible

for retrieving the values is deﬁned by the usable pa-

rameters in the ZeeRex document.

4 SYSTEM INTERACTION

The ﬁrst step to an integrated search functionality is

the generation of an adequate search dialog. To do

this the explain method has to be called by the CSCW

system, which retrieves a ZeeRex document describ-

ing the supported search ﬁelds, selection ﬁelds, and

conﬁguration ﬁelds of the search environment. The

following section 4.1 explains the evaluation of the

retrieved ZeeRex document by the CSCW system for

building the search dialog. Afterwards section 4.2

presents the process of transforming the user’s inputs

into a search request that can be interpreted by the ad-

dressed search environment. At last, section 4.3 pro-

vides the description of how to prepare the retrieved

search results for the usage in the CSCW system.

4.1 Generating the Search Dialog

As shown in section 3.2.1,

elements

contain information how to specify correct search re-

quests. The CSCW system has to evaluate these el-

ements in order to automatically generate a suitable

search dialog. This is done by parsing the received

ZeeRex document after having invoked the explain

method of the search environment. It is important,

not to transform every

<index>

element into a search

ﬁeld, since

<index>

elements are also used to de-

scribe selection and conﬁguration ﬁelds.

The type of an

<index>

element can be identi-

ﬁed referring to the children of the

<map>

element.

According to the possible differentiation between se-

lection, conﬁguration, and search ﬁelds, the CSCW

system can generate the search dialog as follows:

For each search ﬁeld a corresponding input ﬁeld

will be assimilated into the search dialog. After-

ENABLING CSCW SYSTEMS TO AUTOMATICALLY BIND EXTERNAL KNOWLEDGE BASES

327

wards the conﬁguration ﬁelds will be included into

the search dialog as choice ﬁelds. For every

<item>

element an adequate choice option for the generated

choice ﬁeld is inserted. Selection ﬁelds present dy-

namic data for which it is not reasonable to store it

in the manually conﬁgured ZeeRex document. Ac-

cordingly, a digital library should provide a method

to query entries for a dynamic information. DuEPub-

lico, for example, provides such a service for re-

ceiving classiﬁcations of the available materials (see

section 3.2.1) which is called by the CSCW system.

The ZeeRex document contains all information that is

necessary to query the classiﬁcation entries by deﬁn-

ing the name and the parameters of the responsible

method. With the newly gained information about

supported classiﬁcations the CSCW system can adds

choice ﬁelds with appropriate choice options into the

search dialog.

4.2 Building the Search Request

The ZeeRex document is used to generate an adequate

search dialog for the particular search environment

described in section 4.1. It also contains information

to build search requests which can be interpreted by

the search environment. Since arbitrary search envi-

ronments do not support the same query language, the

CSCW system has to create different queries for dif-

ferent search environments.

The ZeeRex document deﬁnes for each ﬁeld the

variable to which the input has to be mapped during

the search request. As described in section 3.2.1, a

search term may be addressed by different terms of

distinguished or equal request formats. DuEPublico

only supports requests in the DC-format, which was

extended by the

conf

-set for conﬁguration options.

The CSCW system has to extract the non-empty ﬁelds

of the user’s submitted search dialog and integrate ad-

equate search phrases into the search request. These

phrases are of the form of

fieldNames Operator

userInput

. The useable ﬁeld names are listed in the

ZeeRex document in the

<name>

-element.

After the search request is built, the CSCW system

invokes the web service to execute the search. In ad-

dition to the search request some search environments

require the transmission of authentication information

for the usage of the search functionality, as well as for

access to the inventory.

4.3 Handling the Search Results

The reply of the search web service is an XML doc-

ument which contains the results in a speciﬁc format.

DuEPublico, for example, sends the results in the DC-

format. If a search environment returns the results in

a different format as the one used by the CSCW sys-

tem in the search request, it is crucial that the CSCW

system can realize the mapping of the received terms

to the requested phrases with the help of the ZeeRex

document (adequate

name

-elements have to exist).

The CSCW system transforms the received results

with an XSLT stylesheet into a suitable presentation

for the user in the CSCW system. In the transforma-

tion process additional functionalities of the CSCW

system can be integrated. In the open sTeam system

the backup-functionality was integrated into the result

presentation in order to enable users to store found

material directly in knowledge spaces.

5 PARALLEL SEARCH IN

DIFFERENT ENVIRONMENTS

The scenarios presented in section 1.1 show the users’

demand for performing speciﬁc searches in differ-

ent search environments. Often, sending the same

search request to different search environments at the

same time, (including or excluding the local CSCW

search) will simplify searching, making it more efﬁ-

cient, since search results from different search en-

vironments are available at the same time and thus,

can be compared. The user-friendliest solution is to

present all results on a single page, which also en-

ables the CSCW system to sort the results according

to their relevance and independent from their source.

In section 4, the handling of one search environ-

ment was described, the concept of which has to be

adjusted to support parallel searches in different envi-

ronments. Then the ﬁrst step for the CSCW system

is to present a selection of useable search environ-

ments in which the user can select the ones to which

his search request should be sent. After the user has

selected the search environments, an adequate search

dialog is presented. This search dialog may only con-

tain search ﬁelds, selection ﬁelds, and conﬁguration

ﬁelds which are supported by all selected search en-

vironments. It is crucial to present the intersection of

the sets of supported ﬁelds in order to return only cor-

rect results. This will be explained in the following:

A search request should be sent, for example, to

DuEPublico and Google. If the search dialog presents

all the input ﬁelds supported by any of the search en-

vironments, the user may add the author information.

The ordinary Google search does not support author

searching. In contract, if the CSCW system adds the

inserted name of the author into the keyword search,

Google will ﬁnd also documents in which the author

is mentioned but of which he is not the author. Ac-

ICEIS 2007 - International Conference on Enterprise Information Systems

328

cordingly, the Google search results are not totally

correct and listing them in the same context as the

DuEPublico results will not be valid.

CSCW

user

Citeseer API

DuEPublico

Queryservice

Google

Search API

CSCW system binds several external

search environments

search in

returns

hits

ZeeRex

Figure 2: Binding Arbitrary Search Environments.

To generate a search dialog presenting only the in-

tersection of the supported functionalities of several

search environments, their ZeeRex documents have

to be compared. Since titles and formats may differ,

the intersection can only be identiﬁed if all search en-

vironments use the same

for equal index ﬁelds. In

this case the CSCW system will be able to ﬁnd equal-

ities of search ﬁelds by comparing the

-attributes.

After the user has submitted the ﬁlled search di-

alog, the CSCW system builds adequate search re-

quest(s) for the selected environments. This step re-

quires the generation of several search requests if the

search environments support different formats. When

the search results have been received by the CSCW

system, adequate XSLT transformations are executed

and the elements can be sorted if the required rele-

vance information has been transmitted by the search

environments or is evaluated by the CSCW system it-

self.

6 CONCLUSION

A user-friendly and efﬁcient solution for searching in

search environments, such as digital libraries, and for

using adequate results in a CSCW system requires the

avoidance of media breach. Therefore, many search

environments already deploy web services for invok-

ing their search functionalities. The presented ap-

proach developed a standardized web service to in-

tegrate different search sources through a self expla-

nation mechanism of web services. Coupling CSCW

systems with various search environments will extend

the functional range of a CSCW system and support

users in their individual work. As a result, an auto-

matic binding of search environments in CSCW sys-

tems is necessary, as manual integration is too time-

consuming, and therefore would limit the number of

available search environments in a CSCW system will

be strongly limited.

In this paper we have presented a prototype which

supports automatic binding of search environments.

For the automation, the search environments have to

support an explain method to describe their function-

alities. This method returns an individual ZeeRex

document, which has been extended in this paper to

satisfy the requirements for the automatic binding.

REFERENCES

Bopp, T., Hampel, T., Hinn, R., L

utzenkirchen, F., Prpitsch,

C., and Richter, H. (2006a). Alltagstaugliche medien-

nutzung erfordert systemkonvergenz in aus- und weit-

erbildung. In E-Learning - alltagstaugliche Innova-

tion?, volume 38 of Medien in der Wissenschaft, pages

87–96.

Bopp, T., Hinn, R., and Hampel, T. (2006b). A service-

oriented infrastructure for collaborative learning in

virtual knowledge spaces. In Education for the 21st

Century: Impact of ICT and Digital Resources, vol-

ume 210 of International Federation for Information

Processing, pages 35–44. Springer.

Gollan, H., Luetzenkirchen, F., and Nastoll, D. (1999). Mi-

less - a learning and teaching server for multi-media

documents. In Lecture Notes in Control and Informa-

tion Sciences: Workshop on Wide Area Networks and

High Performance Computing. Springer, London.

Lagoze, C. and de Sompel, H. V. (2001). The open

archives initiative: building a low-barrier interop-

erability framework. In Proceedings of the 1st

ACM/IEEE-CS joint Conference on Digital Libraries,

pages 54–62, New York, NY, USA. ACM Press.

Payette, S. and Lagoze, C. (1998). Flexible and exten-

sible digital object and repository architecture (fe-

dora). In ECDL ’98: Proceedings of the Second Euro-

pean Conference on Research and Advanced Technol-

ogy for Digital Libraries, pages 41–59, London, UK.

Springer-Verlag.

Petinot, Y., Giles, C., Bhatnagar, V., Teregowda, P., and

Councill, I. (2004). Citeseer-api: Towards seamless

resource location and interlinking for digital libraries.

In Proceesdings of the 13th ACM Conference on Infor-

mation and Knowledge Management, pages 553–561.

ACM Press.

Smith, M. (2002). Dspace: An institutional repository from

the mit libraries and hewlett packard laboratories. In

ECDL ’02: Proceedings of the 6th European Confer-

ence on Research and Advanced Technology for Digi-

tal Libraries, pages 543–549, London, UK. Springer-

Verlag.

SRW (2005). SRW/U - search / retrieve web services.

http://www.loc.gov/standards/sru/

Weibel, S. (1998). The dublin core: A simple content

description format for electronic resources. NFAIS

Newsletter, 40(7):117–119.

ZeeRex (2004). Zeerex: The explainable ’explain’ service.

http://explain.z3950.org/

ENABLING CSCW SYSTEMS TO AUTOMATICALLY BIND EXTERNAL KNOWLEDGE BASES

329