ENABLING CONTEXT-ADAPTIVE COLLABORATION FOR

KNOWLEDGE-INTENSE PROCESSES

Stephan Lukosch

Faculty of Technology, Policy, and Management, Delft University of Technology

PO box 5015, 2600 GA Delft, The Netherlands

Dirk Veiel, J¨org M. Haake

Faculty of Mathematics and Computer Science, FernUniversit¨at in Hagen, 58084 Hagen, Germany

Keywords:

Shared workspaces, Adaptation, Group context.

Abstract:

Knowledge workers solve complex problems. Their work seems not to be routinisable because of the unique

results and constellation of actors involved. For distributed collaboration knowledge workers need many

different tools, which leads to knowledge dispersed over different locations, artifacts, and systems. Context-

based adaptation can be used to support teams by shared workspace environments best meeting their needs. We

propose an ontology representing context in a shared workspace environment, and a conceptual architecture

for context sensing, reasoning, and adaptation. We report on ﬁrst experiences demonstrating the applicability

of our approach and give an outlook on directions of future work.

1 INTRODUCTION

The design, development and production of new and

innovative products often requires knowledge which

is distributed over a whole company. The design team

knows about the preferences of the customers regard-

ing the look and feel of the product. The develop-

ment department knows about the restrictions (e.g.,

weight, size, budget) and capabilities (e.g., function-

ality, savings, improvements). Finally, the production

department knows how to build and ensure the qual-

ity of the product. All steps, from the design to the

production of a product, depend on the previously ac-

quired knowledge of all participants and are regarded

as knowledge work (Drucker, 1999; Drucker, 1993).

Knowledge workers solve complex problem. Ac-

cording to Sari et al. (Sari et al., 2007) knowledge

work seems to be not routinisable because of the

uniqueness of the results and the unique constellation

of actors involved. When knowledge workers col-

laborate in distributed teams, they need many differ-

ent tools (IT systems; e.g., CAD/CAM environments,

ERP systems, IDEs). This leads to knowledge and in-

formation dispersed over different team members, ar-

tifacts and systems. Teamwork also requires commu-

nication over different systems and artifacts, and leads

to difﬁcult communication and coordination. These

problems make distributed collaboration difﬁcult.

Current approaches to support distributed col-

laboration include multimedia communication sys-

tems, repositories for shared documents and shared

workspaces systems, notiﬁcation systems, shared ed-

itors, shared calendars, and workﬂow systems. These

systems either leave the organization of collaborative

work to the users or ignore the changing needs of

users to include different tools and artifacts. While

shared workspace systems try to combine reposito-

ries with coordination and communication support,

they still require team members to manually main-

tain the organization of social processes, artifacts, and

tools. This includes the integration of the tools. As

a consequence, teams may fail to adapt their shared

workspace to best meet their current needs.

Our idea is to address these problems by enabling

context-adaptive collaboration in an integrated shared

work environment that integrates the different tools

of the team members and adapts the behavior of the

shared work environment according to the group con-

text. Though not routinisable, knowledge work shows

high similarity across very divergent situations (Sari

et al., 2007). We want to identify such recurring situ-

ations and possibilities to improvethe current interac-

tion by analyzing the context and adapting the behav-

Lukosch S., Veiel D. and M. Haake J. (2009).

ENABLING CONTEXT-ADAPTIVE COLLABORATION FOR KNOWLEDGE-INTENSE PROCESSES.

In Proceedings of the 11th International Conference on Enterprise Information Systems - Human-Computer Interaction, pages 34-41

DOI: 10.5220/0001865200340041

 SciTePress

ior of the shared work environment to best meet the

current situation. We suggest that the overhead for

manual adaptation may be decreased by computer-

supported adaptation of shared workspace features.

We assume that such adaptation may positively im-

pact team performance (e.g., by changing affordances

in a way that improvesinteraction, process and/or out-

put quality).

We propose that such adaptation can be based on

context information (i.e. information captured by the

system about individual as well as group use and pref-

erences) and on adaptation rules deﬁning the kind

of adaptation. Adaptation includes modiﬁcation of

the set of applications/services for each user, chang-

ing their respective UI, and changing their content.

The actual adaptation rules need to evolve over time,

thus requiring means for users to understand ongoing

adaptations and how they can change them. In this

way, we see the (evolving) rules as another part of the

context information.

Considering the above, it is essential to provide a

context model that allows the system to recognize col-

laboration situations which are suitable for adaptation

and which can easily be extended to include further

context factors. Furthermore, it is necessary to pro-

vide a system architecture that can use this context

model and the included adaptation rules to adapt its

runtime behavior in a way that the interaction among

the knowledge workers is improved.

This paper is structured as follows. First, we re-

view existing shared work environments and context-

adaptive systems. Then, we identify the basic ele-

ments of our context model along an example of two

collaborating knowledge workers. In a second step,

we introduce a conceptual architecture that uses our

context model to enable context-adaptive collabora-

tion. In a third step, we validate our context model by

presenting an example for a possible adaptation and

adaptation rule. This example shows how our con-

text model and conceptual architecture can be used

to recognize situations in which adaptation can im-

prove the team interaction and how adaptation rules

can be triggered and executed to actually improve the

interaction within the team. Finally, we report on ﬁrst

experiences before concluding the paper and raising

questions for future work.

2 RELATED WORK

In the following, we ﬁrst review relevant shared work

environments and discuss how these systems deal

with possible adaptations. Then, we take a look

at context-adaptive systems in general. We review

whether the taken approaches are suitable to support

context-adaptive collaboration in knowledge-intense

processes. We discuss the used context models and

whether these models are suitable to model collabo-

ration situations.

BSCW (Appelt and Mambrey, 1999) and CURE

(Haake et al., 2004b; Haake et al., 2004a) are web-

based shared work environments offering a variety

of collaboration services, e.g. document sharing and

communication. CHIPS (Wang and Haake, 2000) is a

cooperative hypermedia system with integrated pro-

cess support. TeamSpace (Geyer et al., 2001) of-

fers support for virtual meetings and integrates syn-

chronous and asynchronous team interaction into a

task-oriented collaboration space. BRIDGE (Farooq

et al., 2005) is a collaboratory that focuses on sup-

porting creative processes and as such integrates a va-

riety of collaboration services. All of the above exam-

ples focus on a speciﬁc application domain. Though

they all offer a variety of services, the systems are

independent of each other and do not allow to inte-

grate additional services. Some of them, e.g. CHIPS

or CURE are highly tailorable, but they do not adapt

their offered functionality to improve the collabora-

tion within a team.

The most prominent examples for context-based

adaptation focus on single users and consider loca-

tion as most relevant context information (e.g. (Schilit

et al., 1994; Abowd et al., 1997; Kindberg et al.,

2002) or focus on learner proﬁles (cf. ITS). Com-

pared to single-user ITS, COLER (de los Angeles

Constantino-Gonz´alez and Suthers, 2003) provides a

software coach for improving collaboration. CoBrA

(Chen et al., 2003; Chen et al., 2004) is an agent-

based architecture that uses shared context knowledge

represented as an ontology to adapt service agents ac-

cording to a users context. Gross and Prinz (Gross and

Prinz, 2004) introduce a context model and a collab-

orative system that supports context-adaptive aware-

ness. The context model consists of events, artifacts,

locations, etc. The main restrictions of their approach

are that the context representation is only used to

update and visualize awareness information and that

only one cooperative application can be used. Ed-

wards (Edwards, 2005) explores the space between

two different context understandings: in CSCW re-

search, people are assumed to be the consumer of

context information; the ubiquitous computing com-

munity has the opinion that systems are the con-

sumers of context information. Intermezzo (Edwards,

2005) tries to ﬁll this gap through the creation of

new higher-level services, but in the end collabora-

tive applications are missing. Rittenbruch describes

an approach to the representation of context of aware-

ENABLING CONTEXT-ADAPTIVE COLLABORATION FOR KNOWLEDGE-INTENSE PROCESSES

ness information but real world examples are miss-

ing (Rittenbruch, 1999). Fuchs (Fuchs, 1999) de-

scribes an integrated synchronous and asynchronous

notiﬁcation service for awareness information called

AREA, but again AREA uses the context representa-

tion only for awareness information. Ahn et al. (Ahn

et al., 2005) introduce a knowledge context model.

Based on this context model they implement the vir-

tual workgroup support system (VWSS). One draw-

back of their solution is that their knowledge context

model has to be extended for other application do-

mains. The Semantic Workspace Organizer (SWO)

(Prinz and Zaman, 2005) is an extension of BSCW. It

analyzes user activities and textual documents inside

the shared workspace to suggest appropriate locations

for new document upload and for document search.

The ECOSPACE project aims at providing an inte-

grated collaborative work environment (Prinz et al.,

2006; Martnez-Carreras et al., 2007). For that pur-

pose, ECOSPACE uses a service-oriented architecture

and provides a series of collaboration services for or-

chestration and choreography. The orchestration and

choreography is based on a ontology which still has

to be described (Martnez-Carreras et al., 2007; Von-

rueden and Prinz, 2007).

The above approaches focus on context represen-

tations and adaptations which are used in speciﬁc do-

mains, e.g., single-user systems or ITS, or on sub-

domains in the ﬁeld of CSCW, e.g. awareness or

knowledge management. Adaptation based on group

context and for multiple users of a cooperative sys-

tem is intended only by ECOSPACE, but the required

context model is still an open issue. Similarly, only

ECOSPACE supports the integration of different col-

laboration services within the same shared work envi-

ronment. In summary, current approaches do not suf-

ﬁciently support a context-based adaptation of shared

work environments.

3 A CONTEXT MODEL FOR

KNOWLEDGE-INTENSE

PROCESSES

Dey et al. deﬁne context as any information used to

characterize a situation of an entity where entities

may be any object, person or place providing infor-

mation about the interaction between a user and an

application (Dey et al., 2001). With this deﬁnition,

any information may help characterizing the situation

of the interaction’s participants because it is part of

the context itself. For our purposes, we can narrow

this deﬁnition so that context includes all information

which is necessary or helpful to adapt a cooperative

workspace to better ﬁt the needs of a collaborating

team. This implies that the context contains informa-

tion about the team as well as about the current col-

laborative situation.

There exist different approaches to model con-

text. These approaches range from simple key-value

models over graphical models up to sophisticated on-

tology based models, which support validation and

reasoning (Strang and Linnhoff-Popien, 2004). Nel-

son et al. as well as Levitt et al. hypothesized that

routines are key determents of organizational perfor-

mance (Nelson and Winter, 1982; Levitt and March,

1988). To discover such routines, to identify adapta-

tions, and to ensure a consistent context model, we

need reasoning mechanisms and model our context as

an ontology using OWL

In the following, we describe and explore the ba-

sic concepts and relations within a global collabo-

ration space using an example of two collaborating

knowledge workers. We use the following conven-

tion for the names of concepts: <name of the con-

cept>:<name of the instance> (e.g. Actor :alice). In

some cases we use an index to describe the relation-

ship of one instance to another more precisely (e.g.

Role : Author

alice

Figure 1: Basic interactions within a sample global collab-

oration space.

Figure 1 illustrates the basic interactions within a

sample global collaboration space. A global collabo-

ration space contains all actors, user workspaces, col-

laborative applications, services and artifacts needed

to carry out a collaborative project between the in-

volved actors. In our sample environment Actor:alice

and Actor:bob are members of a team represented as

Team:Project

(cf. Figure 1) that uses several collabo-

rative applications to design and describe a new prod-

uct. The concept Role deﬁnes a set of possible actions

that actors can execute within a collaborative appli-

cation. Each collaborative application deﬁnes a set

of supported actions. A subset or the whole set of

these possible action is assigned to a role. Thereby,

different roles with different access rights can be de-

http://www.w3.org/2004/OWL/

ICEIS 2009 - International Conference on Enterprise Information Systems

ﬁned. Each actor can have multiple roles. The role

for a collaborative application is assigned to a user

by the user’s workspace. As Figure 1 shows, Ac-

tor:alice has the role Role:Designer

alice

while using

Collaborative Application:Image Editor

alice

, and the

role Role:Editor

alice

while using Collaborative Appli-

cation:Text Editor

alice

. We distinguish between these

two roles to illustrate that each role usually implies

different activities and uses speciﬁc (i.e. domain spe-

ciﬁc) applications to operate on artifacts. Actor:bob

uses the Collaborative Application:Text Editor

bob

and

has the role Role:Reviewer

bob

. This role implies a re-

stricted action set that is available to operate on ar-

tifacts. Usually, reviewers add comments to the text

but are not allowed to edit or delete the text. While

using the Collaborative Application:Rich Text Edi-

tor

bob

, Actor:bob has the role Role:Author

bob

Figure 2: Basic interactions between user workspaces and

shared model.

Figure 3: Decomposition of collaborative applications us-

ing the model-view-controller pattern.

Figure 2 illustrates that each actor has its own user

workspace (User Workspace:UW

alice

and User

Workspace:UW

bob

). As each user’s workspace

can be conﬁgured differently, a user workspace

deﬁnes a set of available applications. This set of

applications is an attribute of each user workspace.

Each of the two actors (Actor:alice and Actor:bob)

interacts with two collaborative applications that

reside in the corresponding user workspace. Actions,

e.g. AddShape:alice or OpenText:bob, are used

to describe an interaction between actors and the

collaborative applications. When interacting with

a collaborative application, actors perform actions

which the applications are capable of and are allowed

by their role.

A collaborative application uses services, e.g.

Service:LineTo or Service:SetContent, to operate

on artifacts, e.g. Artifact:Image or Artifact:Text

Document. Artifacts may be extended with data and

services for coordination of collaborative work, e.g.

locking information). All actors and artifacts are

potentially located in space and time.

A collaborative application implements the

model-view-controller (MVC) paradigm (Krasner

and Pope, 1988) (cf. Figure 3). Actors interact with

the collaboration application by performing actions

allowed by their roles. These actions are received

by the corresponding controller components of the

application. In Figure 3, Actor:alice performs an

EditText:alice. This action is received by the Con-

troller:Text Document

alice

. The controller then uses

the Service:setContent to modify the Artifact:Text

Document. As in MVC, the shared model then

uses the Service:Notiﬁer to notify registered view

components, e.g. View:Text Document

alice

, about

modiﬁcation in the model. By receiving such a notiﬁ-

cation the registered view component can update the

displayed information.

Figure 4: Basic ontology concepts and relations.

Figure 4 summarizes the above model of a global col-

laboration space and shows the basic ontology con-

cepts and their relations. We modeled this ontol-

ogy using OWL. In addition to the above, we in-

troduced additional concepts, like Communication,

Shared Editing, or Awareness, to classify the func-

ENABLING CONTEXT-ADAPTIVE COLLABORATION FOR KNOWLEDGE-INTENSE PROCESSES

tionality a Collaborative Application offers. For

overview reasons, we only show here three high-level

concepts, but we distinguish further high-levelas well

as additional sub-level concepts.

4 CONCEPTUAL

ARCHITECTURE

Due to the openness of the real world, we may either

have to deal with evolving context dimensions or live

with a closed set of context dimensions. In both cases

new questions arise. How would users be able to deal

with evolving context dimensions? How can we en-

able users to build social solutions around a system

with ﬁxed context dimensions? Though our current

prototype uses a closed set of context dimensions, our

conceptual architecture allows integrating services for

extending context dimensions.

Figure 5: Conceptual architecture.

Our conceptual architecture consists of four layers:

Application User Interfaces (UIs), Adaptation Server,

Collaboration Services, and Shared Model (cf. Figure

5).

A ﬂexible adaptive system executes a cycle of

1. User interaction

2. Sensing user activities

3. Adapting system behavior

4. Modifying adaptation knowledge (e.g., if users

want to change adaptation rules)

We use the sensing functionality to illustrate the in-

teractions between the different components of our

architecture. The UI-part of a collaborative applica-

tion (Application UIs) is used to collect relevant in-

formation about the interaction between the user and

the application. This information (including the ser-

vice calls) is forwarded to the Service Call & Inter-

action Sensing component of the Adaptation Server,

which extracts relevant information and updates Con-

text Knowledge via the Context Knowledge Service

Provider. This information is passed to the Collab-

orative Application Manager (CAM), which triggers

the services of the corresponding Collaborative Ap-

plication. Applications can use several basic services

from different Service Providers to implement the ap-

plication logic. Thereby, we can integrate different

services into an application and adapt application be-

havior across different service providers.

Next, we describe the adaptation functionality.

Based on the event log, the Service Call & Interaction

Sensing component prioritizes the events and adds all

information to a queue. If a high priority event is

present, the number of events exceeds a predeﬁned

threshold value, or a timeout occurs (i.e., no further

events came in for a certain time) the Reasoner is trig-

gered. The Reasoner uses the current Context Knowl-

edge, which is represented by instances of our on-

tology concepts and relations between those, via the

Context Knowledge Service Provider and the current

Adaptation Knowledge, which e.g. includes adapta-

tion rules or axioms, via the Adaptation Knowledge

Service Provider to ﬁnd relevant adaptation rules. Ex-

ecuting these adaptation rules will potentially modify

the set of applications/services for each user, change

their respective UI, and/or change their content in or-

der to improve team interaction. The information

about the modiﬁcations/changes is passed to the CAM

to update the current service conﬁguration. The Noti-

ﬁer sends the adaptation information to the subscribed

UI-parts of a Collaborative Application. The UI-parts

then refresh their view according to the current ser-

vice conﬁguration in the CAM and/or use some Adap-

tation Functions through the Adaptation Functional-

ity Service Provider to modify the content of the UI.

Adaptation rules are initially deﬁned by users

or developers using the Collaborative Adaptation

Knowledge Editor (CAKE). Rules need to match the

needs of the team, thus requiring means for users to

understand ongoing adaptations (traceability, reﬂec-

tion, understandability of changes) and how they can

change them.

The CAKE is used to support the modiﬁcation

of adaptation knowledge. It allows users to access,

review, edit, and create new user-deﬁned adaptation

rules. CAKE is a collaborative application using our

architecture meaning that it can be adapted as well.

CAKE uses the Adaptation and Context Knowledge

Service Providers to access and modify the corre-

sponding adaptation information.

ICEIS 2009 - International Conference on Enterprise Information Systems

5 ADAPTATION EXAMPLE

As a scenario consider that a team consisting of Alice

and Bob synchronously collaborates on a shared text

document. We assume, that Alice and Bob created lo-

cal workspaces. Alice then created a shared text doc-

ument and opened a shared text editor to work on a

design document. Bob later opened the same shared

text document to review the current state of the de-

sign. The team members have different roles high-

lighting their task within the team.

This is expressed in Figure 6 by Role concepts

(e.g., Author) connected to possible actions (e.g.,

OpenText, EditText, Annotate) related to an applica-

tion concept (e.g., TextEditor). We assume that the

assignment of roles (instances of Role concepts) to

concrete Actors (e.g. Bob, Alice) was done initially

when creating application instances (i.e., when open-

ing a shared text editor instance on a concrete docu-

ment). Team members can now synchronously access

and edit document parts, which can lead to conﬂicts

or potential collaboration situations. To improve the

interaction within team, each user’s workspace could

be adapted by providing additional awareness infor-

mation or establishing a communication possibility.

Figure 6: Context graph.

In Figure 6 two actors access the same Artifact:Text

Document. These actors have different roles which

allow the actors to perform different actions. This is

expressed by the relation of roles to actions, which

are supported by the collaborative applications within

the user’s workspace. As an example, consider Ac-

tor:alice in the Role:Author. The Role:Author is re-

lated to all actions supported by the application TextE-

ditor.

Please note, concepts are shown as ovals whereas in-

stances of ontology concepts are shown as rectangles.

Though Actor:alice has the Role:Author and Ac-

tor:bob has the Role:Reviewer, there can be a chance

for collaboration. Both actors access the Artifact:Text

Document and thus are part of the context of the arti-

fact as well as of each other. Actor:bob, in the role of a

reviewer, might want to ask questions for clariﬁcation

for which he has to directly contact an author of the

document. A corresponding adaptation rule checks

whether both actors have the possibility to commu-

nicate with each other. For this purpose, the adapta-

tion rule checks the available applications within each

user’s workspace and whether the current roles allow

the actors to communicate with each other.

In order to represent the tools available to all users

of the team, the respective information must be in-

cluded in the context graph as shown in Figure 6.

Here, the concept Role:Participant connected to the

Chat concept speciﬁes that an instance of a chat ap-

plication supports the OpenChat and SendMsg ac-

tions. The isCapableOf relationship from both user

workspaces to the concept Chat and from the Chat

concept to the set of possible actions expresses that

both actors may actually use a chat application. Thus,

in the current state, Actor:alice and Actor:bob can

both use a Chat as the corresponding concept is re-

lated to their user workspace.

In the above scenario the Service Call & Inter-

action Sensing component receives the EditText:alice

action. Based on this event, the Service Call & Inter-

action Sensing component evaluates the condition of

all adaptation rules. Since the condition of the follow-

ing rule is fulﬁlled, the rule will be executed:

teams := getActiveTeams("Actor:alice");

FOREACH (team : teams) DO {

availableCommAppl := availableAppl(team,

"Communication");

// There exists at least one

// communication application available

// to all team members AND there is a

// common shared artifact.

IF (isNotEmpty(availableCommAppl) AND

isNotEmpty(sharedArtifact(team)))

THEN {

// Select a communication tool and

// open it for all members.

selectedApplication :=

selectOneFrom(availableCommAppl);

openForAll(selectedApplication, team);

}

In this adaptation rule,

availableAppl

returns a set

of Collaborative Applications which support Commu-

nication and are connected to all members of the team

in their current context (cf. Figure 4 for correspond-

ENABLING CONTEXT-ADAPTIVE COLLABORATION FOR KNOWLEDGE-INTENSE PROCESSES

ing concepts and relations).

sharedArtifact

returns

a set of artifacts for which the whole team belongs to

the context of each artifact and

selectOneFrom

se-

lects from a set of context elements the one which has

been used most by the collaborating team members.

The corresponding information is stored as activation

value within the context graph. The activation value

is calculated for each concept in the graph and is up-

dated by a special learning algorithm. Here a Chat ap-

plication will be opened for Alice and Bob and a chat

session will be established between them, as a Chat

application the only application available for Commu-

nication within both users’ workspaces.

6 EXPERIENCES

The current prototype manages our context model us-

ing the semantic web framework Jena

. The Adap-

tation Server is based on Equinox

and realizes all

components as so-called bundles in OSGi

. In the

prototype, we integrated two Collaborative Appli-

cations. Firstly, we extended CURE (Haake et al.,

2004b; Haake et al., 2004a) to provide service classes

for access right, document, user, and workspace man-

agement as well as asynchronous awareness and com-

munication. Secondly, we use ActiveMQ

to develop

and integrate support for the synchronous awareness

and communication. The UI-parts for these service

classes are implemented as plug-ins for Eclipse

We split up the evaluation in three phases. In the

ﬁrst phase, we tested our approach by implementing

the conceptual architecture, integrating two collabo-

rative applications, and conducting functional tests.

Currently, we are integrating further applications, like

Microsoft Exchange or the Java Content Repository,

to extend the number of available application types in

our context model.

In the second phase, we have setup a test environ-

ment and conducted expert walkthroughs in different

work scenarios, e.g. collaborative planning or writ-

ing. Feedback from these experts indicates both that

adaptation is in general accepted as useful but often

users also want to understand why a speciﬁc adapta-

tion was performed. Some users also reported that

they dislike automatic adaptation in general, as they

want to stay in complete control of their workspace.

However, even these users would appreciate sugges-

tions for adaptation, i.e. semi-automatic adaptations.

http://jena.sourceforge.net/

http://www.eclipse.org/equinox/

http://www.osgi.org

http://activemq.apache.org

http://www.eclipse.org/

In the third phase, we will evaluate our architecture

with the integrated collaborative applications in real-

world settings that are based on knowledge-intense

processes as they can be found in collaborative mod-

eling or design. We will also collect more feedback on

the existing adaptation rules and identify best prac-

tices for collaboration leading to further adaptation

rules.

7 CONCLUSIONS

In this paper, we introduced an explicit representation

of context and a conceptual architecture for context

sensing, reasoning, and adaptation in an integrated

shared work environment. We integrated ﬁrst collab-

orative applications to show the feasibility of our ap-

proach.

Our approach exceeds current approaches by

deﬁning an ontology-based context model for shared

workspaces and an architecture which allows integrat-

ing a varietyof services to adapt the shared workspace

to best meet the current needs of collaborating users.

First experiences show the usefulness as well as prob-

lems caused by the adaptation.

The ﬁrst experiences especially highlight the im-

portance of traceability support for the users. Such a

support has to allow users understand why an adap-

tation was performed. For that purpose, we will as a

ﬁrst step include an ACTIVITY LOG as well as AC-

TIVITY INDICATOR (Sch¨ummer and Lukosch, 2007).

But as in addition some users reported that they want

to keep the full control of their workspace, we will

also focus on enabling users to deﬁne their own adap-

tation rules as well as supporting a process to negoti-

ate adaptations and adaptation rules.

REFERENCES

Abowd, G. D., Atkeson, C. G., Hong, J., Long, S., Kooper,

R., and Pinkerton, M. (1997). Cyberguide: a mo-

bile context-aware tour guide. Wireless Networks,

3(5):421–433.

Ahn, H. J., Lee, H. J., Cho, K., and Park, S. J. (2005). Utiliz-

ing knowledge context in virtual collaborative work.

Decision Support Systems, 39(4):563–582.

Appelt, W. and Mambrey, P. (1999). Experiences with the

BSCW shared workspace system as the backbone of a

virtual learning environment for students. In Proceed-

ings of ED-MEDIA99.

Chen, H., Finin, T. W., and Joshi, A. (2003). Using owl in a

pervasive computing broker. In Craneﬁeld, S., Finin,

T. W., Tamma, V. A. M., and Willmott, S., editors,

ICEIS 2009 - International Conference on Enterprise Information Systems

Proceedings of the Workshop on Ontologies in Agent

Systems (OAS 2003), pages 9–16.

Chen, H., Finin, T. W., and Joshi, A. (2004). Semantic web

in the context broker architecture. In Proceedings of

the Second IEEE International Conference on Perva-

sive Computing and Communications (PerCom 2004),

pages 277–286. IEEE Computer Society.

de los Angeles Constantino-Gonz´alez, M. and Suthers,

D. D. (2003). Automated coaching of collaboration

based on workspace analysis: Evaluation and impli-

cations for future learning environments. In Proceed-

ings of the 36th Hawai‘i International Conference on

the System Sciences (HICSS-36). IEEE Press.

Dey, A. K., Abowd, G. D., and Salber, D. (2001). A concep-

tual framework and a toolkit for supporting the rapid

prototyping of context-aware applications. Human-

Computer Interaction, 16(2, 3, & 4):97–166.

Drucker, P. F. (1993). Concept of Corporation. Transaction

Publishers.

Drucker, P. F. (1999). Management Challenges for the 21st

Century. HarperCollins Publishers, New York, USA.

Edwards, W. K. (2005). Putting computing in con-

text: An infrastructure to support extensible context-

enhanced collaborative applications. ACM Transac-

tions Computer-Human Interaction, 12(4):446–474.

Farooq, U., Carroll, J. M., and Ganoe, C. H. (2005). Sup-

porting creativity in distributed scientiﬁc communi-

ties. In GROUP ’05: Proceedings of the 2005 in-

ternational ACM SIGGROUP conference on Support-

ing group work, pages 217–226, New York, NY, USA.

ACM.

Fuchs, L. (1999). AREA: a cross-application notiﬁcation

service for groupware. In ECSCW’99: Proceedings

of the sixth conference on European Conference on

Computer Supported Cooperative Work, pages 61–80.

Kluwer Academic Publishers.

Geyer, W., Richter, H., Fuchs, L., Frauenhofer, T., Dai-

javad, S., and Poltrock, S. (2001). A team collabo-

ration space supporting capture and access of virtual

meetings. In Proceedings of the 2001 International

ACM SIGGROUP Conference on Supporting Group

Work, pages 188–196, Boulder, Colorado, USA. ACM

Press, New York, NY, USA.

Gross, T. and Prinz, W. (2004). Modelling shared contexts

in cooperative environments: Concept, implementa-

tion, and evaluation. Computer Supported Coopera-

tive Work (CSCW), 13(3):283–303.

Haake, J. M., Haake, A., Sch¨ummer, T., Bourimi, M.,

and Landgraf, B. (2004a). End-user controlled group

formation and access rights management in a shared

workspace system. In CSCW ’04: Proceedings of the

2004 ACM conference on Computer supported coop-

erative work, pages 554–563. ACM Press, New York,

NY, USA.

Haake, J. M., Sch¨ummer, T., Haake, A., Bourimi, M., and

Landgraf, B. (2004b). Supporting ﬂexible collabora-

tive distance learning in the CURE platform. In Pro-

ceedings of the Hawaii International Conference On

System Sciences (HICSS-37). IEEE Press.

Kindberg, T., Barton, J., Morgan, J., Becker, G., Caswell,

D., Debaty, P., Gopal, G., Frid, M., Krishnan, V., Mor-

ris, H., Schettino, J., Serra, B., and Spasojevic, M.

(2002). People, places, things: web presence for the

real world. Mobile Network Applications, 7(5):365–

376.

Krasner, G. E. and Pope, S. T. (1988). A cookbook for using

the model-view-controller user interface paradigm in

Smalltalk-80. Journal of Object-Oriented Program-

ming, 1(3):26–49.

Levitt, B. and March, J. G. (1988). Organizational learning.

Annual Review of Sociology, 14(1):319–338.

Martnez-Carreras, M. A., Ruiz-Martnez, A., Gmez-

Skarmeta, F., and Prinz, W. (2007). Designing a

generic collaborative working environment. In IEEE

International Conference on Web Services (ICWS

2007), pages 1080–1087.

Nelson, R. R. and Winter, S. G. (1982). An evolutionary

theory of economic change. Harvard University Press.

Prinz, W., Loh, H., Pallot, M., Schaffers, H., Skarmeta,

A., and Decker, S. (2006). ECOSPACE – towards

an integrated collaboration space for eprofessionals.

In International Conference on Collaborative Com-

puting: Networking, Applications and Worksharing,

pages 39–45.

Prinz, W. and Zaman, B. (2005). Proactive support for the

organization of shared workspaces using activity pat-

terns and content analysis. In GROUP ’05: Proceed-

ings of the 2005 international ACM SIGGROUP con-

ference on Supporting group work, pages 246–255.

ACM, New York, NY, USA.

Rittenbruch, M. (1999). Atmosphere: towards context-

selective awareness mechanisms. In HCI (2), pages

328–332.

Sari, B., Katzy, B., and Loeh, H. (2007). Coordination rou-

tines behind knowledge intensive work processes. In

Proceedings of the 13th International Conference on

Concurrent Enterprising.

Schilit, B., Adams, N., and Want, R. (1994). Context-aware

computing applications. In First Annual Workshop

on Mobile Computing Systems and Applications (WM-

CSA).

Sch¨ummer, T. and Lukosch, S. (2007). Patterns for

Computer-Mediated Interaction. John Wiley & Sons,

Ltd.

Strang, T. and Linnhoff-Popien, C. (2004). A context mod-

eling survey. In First International Workshop on Ad-

vanced Context Modelling, Reasoning And Manage-

ment.

Vonrueden, M. and Prinz, W. (2007). Distributed document

contexts in cooperation systems. In Kokinov, B. N.,

Richardson, D. C., Roth-Berghofer, T., and Vieu, L.,

editors, Modeling and Using Context, 6th Interna-

tional and Interdisciplinary Conference, CONTEXT

2007, LNCS 4635, pages 507–516. Springer-Verlag

Berlin Heidelberg.

Wang, W. and Haake, J. M. (2000). Tailoring Groupware:

The Cooperative Hypermedia Approach. Computer

Supported Cooperative Work, 9(1):123–146.

ENABLING CONTEXT-ADAPTIVE COLLABORATION FOR KNOWLEDGE-INTENSE PROCESSES