INTERCONNECTION OF COMMUNITIES OF PRACTICE

A Web Platform for Knowledge Management

Élise Garrot-Lavoué

Université de Lyon, INSA Lyon, Laboratoire LIESP, 21 Avenue Jean Capelle, F-69621 Villeurbanne, France

Keywords: Interconnection of Communities of Practice, Knowledge Indexation, Contextualised search, Web 2.0,

Human–Computer Interface, Online tutoring.

Abstract: Our works aim at developing a Web platform to connect various Communities of Practice (CoPs) and to

capitalise on all their knowledge. This platform addresses CoPs interested in a same general activity, for

example tutoring. For that purpose, we propose a general model of Interconnection of Communities of

Practice (ICP), based on the concept of Constellation of Practice (CCP) developed by Wenger (1998). The

model of ICP was implemented and has been used to develop the TE-Cap 2 platform which has, as its field

of application, educational tutoring activities. In particular, we propose an indexation and search tool for the

ICP knowledge base. The TE-Cap 2 platform has been used in real conditions. We present the main results

of this descriptive investigation to validate this work.

1 INTRODUCTION

People belong to Communities of Practice (CoPs) at

the local level of their company or institution. The

emergence of such communities occurs when people

have informal discussions, help each other to solve

problems and use this to develop their competencies

and expertise. CoPs centred on a same activity can

have similar practices without being necessarily

aware of it, mainly due to the fact that they do not

belong to the same company or institution. As a

result, every local CoP develops its own practices,

each one reinventing what is certainly being

replicated somewhere else. Our work is illustrated

throughout the article by the example of tutoring,

which we define as the educational monitoring of

learners during courses. CoPs of tutors from

different educational institutions prepare their own

pedagogical contents for their students, and there is

currently no possibility of reusing and sharing them

(Garrot et al., 2009). The result of this is that tutors

lack help in their day-to-day practice, professional

identity and practice sharing.

The problem which is challenging us is the

creation of relation between local CoPs of actors

practicing a same activity so that they exchange their

knowledge and produce more knowledge than

separate communities. We aim at developing a Web

platform to capitalise on all produced knowledge by

contextualising it, so as to make it accessible and

reusable by all members in their working contexts.

Our work is based on the concept of

Constellation of Communities of practice (or CCP)

developed by Wenger (1998). In this article, we first

present the main characteristics of this concept on

which we base our research. We then situate our

works by studying existing knowledge management

systems and social networking services. In the third

section, we propose a model of Interconnection of

Communities of Practice (ICP), as an extension of

the concept of CCP. This model approaches the

actors’ activity according to the point of view of

interconnected practices and considers CoPs’

members to act as the nodes between CoPs to

support knowledge dissemination. In the fourth

section, we present the implementation of the model

of ICP by the development of the TE-Cap 2

platform, meant for CoPs of educational tutors from

different institutions, countries and disciplines who

would tutoring. We finally validate our works by

presenting the main results of a descriptive

investigation.

2 CONSTELLATION OF

COMMUNITIES OF PRACTICE

Explaining that some organisations are too wide to

Garrot-Lavoué É. (2009).

INTERCONNECTION OF COMMUNITIES OF PRACTICE - A Web Platform for Knowledge Management.

In Proceedings of the International Conference on Knowledge Management and Information Sharing, pages 13-20

DOI: 10.5220/0002273500130020

 SciTePress

be considered as CoPs, Wenger (1998) sets out his

vision of these organisations as “Constellation of

Communities of practice” (or “CCP”).

We first define “Communities of Practice” (or

“CoPs”). Communities of Practice gather people

together in an informal way (Lave and Wenger,

1991) because of the fact that they have common

practices, interests and purposes (i.e. to share ideas

and experiences, build common tools, and develop

relations between peers) (Wenger, 1998; Koh and

Kim, 2004). Their members exchange information,

help each other to develop their skills and expertise

and solve problems in an innovative way (Pan and

Leidner, 2003; Snyder et al., 2004). They develop a

community identity around shared knowledge,

common approaches and established practices and

create a shared directory of common resources.

We identify three main aspects of the concept of

CCP, on which we base our works so as to develop a

platform to support several Communities of Practice

(CoP), summarised by Figure 1:

 To Favour Interactions among CoPs. Brown

and Duguid (1991) brought the notion of

“communities-of-communities” to develop the

innovation within organisations, considering

that the productions of separate communities

can be increased by exchanges among these

communities. The concept of Constellation of

Communities of Practice (Wenger, 1998)

resumes this idea by directing it on practices.

The advantage to define several communities

around shared practices is to create more

knowledge and to develop more interactions

than in a global community (Pan and Leidner,

2003). An involvement of this vision is to think

about interactions among practices, rather than

to favour information flows.

 To Consider the Boundaries of CoPs as

Places of Creation of Knowledge. The

relations between communities can be

supported by boundary objects (Star and

Griesemer, 1989) and by brokering. Boundary

objects are products of reification and they

constitute the directory of resources shared by

all the communities. Interactions between

communities relate to this knowledge.

“Brokers” belong to multiple communities and

have a role of knowledge import-export

between these communities. According to

Ziovas and Grigoriadou (2007), the

combination of brokering as a product of

participation and the boundary objects as a

product of reification is an effective way to

create relations between CoPs. The meetings

on the boundaries of CoPs arouse interactions

between the members, what makes boundaries

the places of creation of knowledge;

 To Establish a Balance in the Duality

Local/Global. A person belongs to and

involves in one or several CoPs, each bound to

its local practices. But the concept of

constellation approaches the CoPs in a global

point of view, as a set of practices negotiated

with only one shared resources repository.

Every member, as broker, operates the

dissemination of knowledge from a level of

practice to another one. That is why it is

necessary to supply all CoPs with multiple

means of communication between practices

which feed the shared directory (Wenger,

1998).

Figure 1: Modelling of the concept of Constellation of

Communities of Practice (CCP).

3 KNOWLEDGE MANAGEMENT

SYSTEMS AND SOCIAL

NETWORKING SERVICES

In this section, we situate our works with regard to

KM systems and social networking services so as to

show that we cannot use existing complete solutions.

A KM system has to support the KM process

following three stages (Von Krogh, 1999): capturing

knowledge, sharing and transferring knowledge,

generating new knowledge. The KM platform of a

company is aimed at its organisational entities, what

implies that:

 These systems are not designed to CoPs which

do not correspond to traditional organisational

entities;

 The proposed computer tools are the only means

for the employees to communicate remotely;

they thus have to use them if they want to

exchange their practices;

 The employees meet during meetings within

their organisational entities, so weave relations

except the platform.

KMIS 2009 - International Conference on Knowledge Management and Information Sharing

 The employees belong to organisational entities

for which they already have a feeling of

membership.

Since our works concern actors who do not

necessary belong to the same institution or the same

company, we cannot use an existing KM platform.

The most important difficulty to overcome is to

arouse interactions between persons except any

frame imposed by an organisation. For that purpose,

it is necessary to bring them to become aware that

they have shared practices and to provide the

available means to get in touch with people from

different CoPs.

Some Web 2.0 applications as Facebook or

MySpace are social networking services which

“connect you with the people around you”. They are

very good examples of services which aim at

connecting people who have common interests.

Some social networking services are for more

professional vocation, such as LinkedIn and Viadeo.

But these sites are used for socialisation and to meet

people. A consequence is that the tools offered to

classify and to search for knowledge are not adapted

to CoPs. Indeed, they often rest on collective

categorisation in the form of tag clouds (O'Reilly,

2005) (folksonomies) or on full text search. But this

system of ‘tagging’ lacks structuring (Guy and

Tonkin, 2006). Within the framework of a CoP, we

consider it is necessary to bring a knowledge

organization to help users to index and search for

knowledge. Tags systems work well for

communities of interest where the users want to

navigate within the application without precise

intention. But these systems are not really adapted to

CoPs where the users search for resources bound to

working experiences. Users must be able to find a

testimony, a discussion, an ‘expert’ or other

resources (document, Web link…) very quickly, so

that they can use it in their practice.

To sum up, we can use neither complete KM

solutions nor existing social networking services but

we can use existing components. We adopt one of

the Web 2.0 principles: “innovation in assembly”

(O'Reilly, 2005). When there are a lot of basic

components, it is possible to create value by

assembling them in a new way. We chose to develop

a platform partially composed of existing Web 2.0

tools (Wenger et al., 2005), available as well for KM

systems as for social networking services, to

capitalise knowledge and get in touch with people.

Other part of the system consists of a knowledge

indexation and search tool specifically developed to

answer specific needs of CoPs, based on the model

on Interconnection of Communities of Practice

depicted in next section.

4 MODEL OF

INTERCONNECTION OF

COMMUNITIES OF PRACTICE

The concept of CCP is based on the assumption that

considering a global community as a set of

interconnected CoPs increase member participation

and creation of knowledge. Furthermore, this vision

of an organisation takes into account as much the

local level of every CoP as the global level formed

by all the CoPs. We adopt this approach to develop a

model of Interconnection of CoPs (ICP) which

proposes to approach a general activity according to

multiple points of view depending on actors’

practices. The development of the Web platform Te-

Cap 2, depicted in section 5, is based on this model.

4.1 General Model of ICP

In the case of informal professions, such as tutoring,

it is difficult to define exactly the field of practice of

the actors. Actors’ activities can be seen as a set of

different practices which are similar in some points.

For example, tutors’ roles can be different as their

interventions could be punctual or long-lasting; the

learning session could be computer mediated or not

and the learners’ activity could be individual or

collective. But some roles are shared by some of

these contexts. We propose that this group of actors

should be seen not as an endogenous entity defined

by a field of practice, but rather as a set of CoPs

supported by a Web platform where individual

members acting as nodes of interconnected practices

are the connection points (see Figure 2). We suggest

developing this concept that we have named

Interconnection of Communities of Practice (ICP).

This model aims at making existing local CoPs of

actors (e.g. within an educational institution), who

are engaging in the same general activity (i.e.

tutoring), to get connected. This model also proposes

active support for the dissemination of knowledge

from CoP to CoP.

Figure 2: General model of Interconnection of

Communities of Practice.

INTERCONNECTION OF COMMUNITIES OF PRACTICE - A Web Platform for Knowledge Management

At an individual level, an actor’s activity can be

approached according to multiple points of view

depending on the working context. In the ICP

model, a CoP corresponds to the elementary level of

actors’ practice. The CoPs to which they belong are

defined by their working context. At a general level,

an ICP is composed of all the elementary CoPs

defined by all the actors who participate in the Web

platform. We could see it as a single community of

actors practicing a same activity, brought together

on the same platform; a group which can be

approached from multiple points of view and

accessed through multiple entry points.

For example (see Figure 3), Tutor 1, working in

the industrial engineering department of the

University A in France who is monitoring a

collective project about maintenance can belong to

five different CoPs: tutors who monitor collective

activities, tutors who are interested in maintenance,

tutors who monitor educational projects, tutors of

the industrial engineering department and tutors of

the University A. Tutor 2 from another educational

institution, for example University B in Canada, can

belong to several CoPs, some of which Tutor 1 may

also belong to.

Figure 3: Tutors as nodes of Interconnection of CoPs.

These two tutors, from different countries, will be

put in touch since their working context can be

approached according to several similar points of

view, which imply that they belong to same CoPs.

Tutor 3 will be put in touch with both tutors because

he belongs to the same educational institution and

the same department as Tutor 1 and because he

monitors the same type of activity as Tutor 2. So this

example illustrates the fact that it is the tutors who

are the nodes of Interconnection of CoPs. In this

example, tutors’ activity can be approached from

several points of view: the context of the activity

(collective, distance), the learning situation (project

based learning, courses), the discipline

(maintenance), the curriculum (industrial

engineering) and the educational institution

(universities). These points of view are categories of

CoP and we propose in section 4.3 an approach to

define a model of actors’ practices, which implies

determining all the categories of CoPs and which

CoPs correspond to a given activity.

4.2 The Reasons for using ICP instead

of CCP

We based the model of ICP on the model of CCP

since they suggest both considering wide

organisations as a set of communities of practice

which have common characteristics (Wenger 1998):

 They share members: the ICP members belong

to several CoPs, each corresponding to a point

of view of their working context;

 They share artefacts: the ICP members

participate on the same Web platform;

 They have access to the same resources: the

ICP members have access to the shared

directory of resources stored in the platform

database.

However, an organisation defined as an

Interconnection of CoPs (supported by a Web

platform and composed of individual members who

act as nodes of interconnected practices) does not

form a Constellation of CoPs as defined by Wenger:

 Contrary to a CCP, the CoPs of an ICP do not

share historic roots on which the mutual

engagement of the members could base itself.

The ICP members do not know apart the

platform on which they join. This difference is

fundamental because it raises the difficulty

bringing persons who do not know each other

to interact, what requires supporting a high

level of sociability on the platform.

 In a CCP, the CoPs have interconnected

projects which connect them whereas an ICP

consist of actors practicing a same general

activity who want to exchange on their

practices with others, the community emerging

by “propagation”. So that members are

interested in the practices of the others, it is

important to bring them to be aware that they

have rather close practices which they can

share.

 Contrary to a CCP, the ICP members do not

belong necessarily to the same institution. Since

we aim at supporting exchanges as well in

members’ local working context as at the

general level of the activity, it is necessary that

KMIS 2009 - International Conference on Knowledge Management and Information Sharing

there are actors of various institutions.

 The CoPs of a CCP are in close proximity to

each other, in particular geographically,

whereas an ICP is constituted of persons who

meet themselves on a Web platform and can

thus be from countries of the whole world. This

model does not thus include geographical

proximity.

So, we propose a new model of ICP to represent a

close but different type of organisation which could

be seen as:

 An extension of the model of CCP in the sense

that the conditions are less restricting. We

showed that only three conditions on seven put

by Wenger (1998) are necessary to validate the

existence of an ICP.

 A transposition of the model of CCP in the

sense that it concerns persons gathered by a

Web platform and not by a given institution or

company.

4.3 Management and Dissemination of

the ICP Knowledge

The ICP resources are stored in a database according

to a hierarchical classification composed of subjects

based on a model of actors’ practices. In the case of

tutoring, resources correspond to explicit knowledge

(documents and Web links) and tacit knowledge

shared among members (e.g. exchanges of

experience, stories, and discussions). We built a

model of tutors’ practices which defines at most four

levels. The first level corresponds to the main factors

which differentiate actors’ practices (e.g. educational

institution, curriculum, discipline, activity) and are

the main categories of CoP. Each category is divided

into subcategories and so on. The terminal nodes

correspond to CoPs. This taxonomy of tutoring has

been developed by an iterative process, based on

interviews with six tutors (first development cycle)

and on results of an experiment of a first prototype

(second development cycle). The classification

cannot be exhaustive because it is only a base which

will evolve through modifications and additions

made by the ICP members themselves.

When creating a resource (message, document,

Web link), the author decide that it belongs to one or

several CoPs by associating the name of the CoP

(subject in the lowest level of the classification) with

the resource. When they find a resource (result of a

search), members can also associate new subjects

with this resource so as to spread it to new CoPs.

They can either associate the name of a CoP to

spread the resource to only a single CoP, or

associate it to the name of a category of CoPs

(subjects at higher levels in the classification) to

spread the resource to all child CoPs. Indeed, Child

CoPs (hierarchically lower level CoPs) inherit all the

resources of a category of CoPs. So, ICP members’

participation not only consists of creating new

resources but also of creating links between these

resources according to their relevance to the CoPs.

This relevance is estimated by members themselves

who consider a resource to be useful or interesting

for a CoP. The supply of a resource to a CoP can

lead to a debate on this resource and possibly to the

creation of new resources for this CoP. Events

reported in a precise context can lead to experience

sharing (solutions, cases, scenarios), being used as a

base to generate rules or recommendations which

become global knowledge within the ICP.

5 THE TE-CAP 2 PLATFORM

We have developed the TE-Cap 2 (Tutoring

Experience Capitalisation) platform according to a

co-adaptive approach based on an iterative process

including three development cycles. Each cycle rests

on the development of a prototype, on its evaluation

by the users by means of interviews or experiments

and on the analysis users’ activity (Garrot et al.,

2009). This approach aimed at making users’ needs

emerge, at leading users to explicit these needs. The

platform specifications evolved according to these

emerging needs. We were particularly interested in

developing a knowledge indexation and search tool

for an ICP. We describe this tool in the following

section.

5.1 User Profile Management

The knowledge indexation and search tool is based

on the user profiles used to personalise subjects

proposed to them. Users define their profile by

filling several fields corresponding to categories of

CoPs of the hierarchical classification. Values given

to fields define CoPs and imply tutors’ membership

of these CoPs. The profile is composed of three

main characteristics: identity profile, working

context and secondary interests. The working

context is about all the CoPs directly bound to

actors’ working context. The secondary interests are

about all the CoPs which are not directly bound to

their working context but which could interest them

(give access to other resources able to interest them

and to profiles of other people who share similar

practices or experiences).

INTERCONNECTION OF COMMUNITIES OF PRACTICE - A Web Platform for Knowledge Management

Figure 4: Knowledge search tool.

As a tool provided for the use of members of a CoP

in their daily practice, this one offers them fast

access to the relevant resources for them by two

means (see Figure 4):

 A link between the search interface and the

profile allows users to only see the subjects

from the classification which concern users and

which interest them according to their profile.

So users only have access to the resources of

the CoPs to which they declare themselves to

belong and can create resources only for these

CoPs.

 Users have the possibility, according to their

intention when connecting to the platform, to

apply a filter to display on the classification

interface only those subjects bound to their

working context or to their secondary interests.

In their daily practice, it is advisable to offer

users at first only those subjects which concern

their direct working context, this being the most

efficient. If users do not find the information

they look for in their direct working context,

they must be able to extend the search to the

other subjects of interest bound to their activity.

In this manner they can find interesting

‘unexpected’ resources, which they can then

bring into CoPs in which they have a central

role.

5.2 Knowledge Indexation and Search

Tool

The knowledge search and indexation tool,

illustrated by Figure 4, rests on the classification

built for the ICP. The main panel (at the centre of

the screenshot), composed of three tabs, allows easy

and fast navigation between the results of the search

and the classification. The tab ‘Search’ gives the

possibility of navigating within the classification and

of selecting search subjects. These subjects are

represented in the form of bubbles, to bring

conviviality and attractiveness to the interface. Users

can navigate in the classification by a ‘double-click’

on a bubble which explodes it into more bubbles

representing the sub-subjects. When reaching the

last level (corresponding to the CoPs), subjects are

represented in the form of a combo box allowing a

multiple selection. Users can return to a superior

level thanks to the navigation path. The platform

proposes the same interface to search for posted

messages and for member profiles, by separating

them by the way of two tabs. In this way users can,

at every search, consult the profiles of found

members and ‘discover’ people who have similar

practices or who offer expertise.

The secondary panel (on the right of the

screenshot) gives the possibility of storing the

subjects chosen for the search (by a drag and drop

from the main panel). The subjects in this column

are always visible when users navigate in the tabs of

the main panel and from one request to another.

Once in the “search column” users can deselect or

select a subject (so as to refine or to widen the

search), delete a subject by sliding the bubble

outside the column and move bubbles inside the

column to choose a preferred order. This principle of

category selection can be compared to carts on

commercial Web sites. This original human

computer interaction has been chosen to promote

navigation within the classification and to simplify

the selection of items.

KMIS 2009 - International Conference on Knowledge Management and Information Sharing

The indexing of an initiating message (starting a

discussion) is made according to the following

principle: users classify the message according to its

context (bound subjects) at the same time as they

write it. This principle aims at leading them to

reflect upon the experience they relate. To facilitate

this action, an interface in the form of tabs ensures

an easy navigation, at any time, between the writing

and the indexing of a message. The selected subjects

in the classification column are then associated with

the message, meaning that this resource belongs to

the CoPs or categories of CoPs. Every user can

associate the discussion with new subjects so as to

spread the resource from one CoP to another one

and from one level to another. Regulation is carried

out by the author of the initiating message who has

the right to remove the subjects which they do not

consider relevant for the discussion.

5.3 Classification Evolution

Users can make the resource classification evolve

through their participation on the platform, so as to

lead to a classification using a vocabulary which

gradually moves closer to the actors’ practices. For

that purpose, the interface gives at any time the

possibility of adding a new subject to the

classification, be it when filling in a profile, when

classifying a resource, when searching a resource or

when consulting a resource. The subjects used are

recorded which allows for example the deletion of

those considered useless. Unused subjects are later

deleted, meaning that they were not adapted to the

actors’ field of practice or not located at the right

level of the classification. This evolution of subjects

is necessary so that the classification made a priori

becomes closer to the reality of actors’ practices and

can follow the evolution of actors’ uses and

practices. It is also an important point for ensuring a

coherence of all the CoPs forming the ICP and for

offering a common identity to all the members.

6 A DESCRIPTIVE

INVESTIGATION

We conducted a descriptive investigation in real

conditions, from 25 February 2008 to July 5 2008.

Our role consisted of encouraging registered tutors

to participate by sending regular newsletters. The

Web address of TE-Cap 2 was disseminated to

several communities of tutors (ATIEF, t@d,

PALETTE) and to virtual campus (VCiel, FORSE,

E-Miage, Téluq, Master UTICEF, did@cTIC, FLE).

We also sent an email to the users of the first

prototype TE-Cap. We wanted to develop the

community around this existing core, hoping that

they would encourage new users to participate.

Discussion threads created during the first study

were kept to be used as a base for new discussions.

To help in the understanding of the how the platform

works, we posted online demonstration videos: one

general one and three specific ones (how to do a

search, to write a message and to fill in the profile).

This study aimed at validating the TE-Cap 2

platform as a support for the interconnection of

CoPs of tutors. We defined indicators to measure

sociability, levels of knowledge creation and sharing

and utility of the platform (Garrot, 2008). Results

come from three types of data: use tracks, answers to

a questionnaire and usability tests.

42 persons from nine francophone countries

registered on TE-Cap 2. We present in this paper

only the main results regarding the indexation and

search tool. First of all, the answers to the

questionnaire show that our aim to put local CoPs

and online CoPs (general CoPs) into a relation

answers an existing need. Indeed, tutors look for

information as much at the local level of their course

(eight answers to the questionnaire) as at a more

general level such as tutors’ roles (twelve answers),

technical and educational tools and resources

(twelve answers), learners (ten answers) or learning

scenarios (eight answers). Although quite a few

messages were written (fifteen) more (twenty-seven)

users simply viewed discussions. This is explained

by the fact that, according to the answers to the

questionnaire, the users registered as much out of

curiosity regarding a new tool as out of a desire to

really participate. Furthermore, participation in a

community will always be lower priority than

teaching or tutoring. A positive result is the rather

large number of subjects added to the classification

(45), which implies a significant evolution in the

classification and thus an appropriation by the users.

Finally, usability tests carried out with three tutors

according to a scenario, highlight the fact that the

indexation and search interfaces of TE-Cap 2 are

very easy to use and effective. But the use of these

interfaces requires a learning stage, as is normal for

an innovative interface which proposes new features.

Furthermore, twenty-three users did not fill in or did

not use their profile which, we must assume, means

they not did not see the interest or did not take the

time (it requires 5 to 10 minutes). The emphasised

reason according to the questionnaire responses was

that they did not understand the link between the

INTERCONNECTION OF COMMUNITIES OF PRACTICE - A Web Platform for Knowledge Management

profile and the proposed classification. It would be

necessary to explain this link better so that they

could see the interest. The help brought by the

videos was either not sufficient or not adapted and

an improvement could be the addition of a

contextual help or a software companion.

Further results will be obtained only by a use by

a large number of persons and over a longer time

period. It is only in these conditions that the

platform and the proposed tools can be expected to

reveal their potential.

7 CONCLUSIONS

In this paper, we defined a general model of

Interconnection of CoPs, based on the concept of

Constellations of CoPs. This model aims at

supporting knowledge sharing and dissemination for

local CoPs interested in a same general activity, in

our case tutoring. We validated the implementation

of this model by the development of the TE-Cap 2

platform. This platform was designed to connect

several CoPs centred on same general activity and to

manage their knowledge. We conducted a

descriptive investigation lasting several months with

tutors from various disciplines and countries. The

results of usability tests demonstrated the ease of use

and the utility of the proposed tool, although not all

the offered possibilities were taken up, as

highlighted by use tracks.

The aim of this study was not to observe the

emergence of an Interconnection of Communities of

Practice because it was unachievable in only four

months. So as to observe such emergence, we plan

to conduct another type of study, across a long-term

period and with the addition of a software

companion to facilitate the understanding of the

innovative interface. It would also be interesting to

address other communities than that of tutors or

teachers who often tend towards rather

individualistic professional behaviour and who are

not always used to sharing.

REFERENCES

Brown, J.S. and Duguid, P., 1991. Organizational learning

and communities of practice. Organization Science,

2(1), 40-57.

Garrot, E., George, S. and Prévôt, P., 2007. The

Development of TE-Cap: an Assistance Environment

for Online Tutors. 2nd European Conference on

Technology Enhanced Learning (EC-TEL 2007),

Lecture Notes in Computer Science, Springer, Crete,

Greece, 481-486.

Garrot, E., George, S. and Prévôt, P., 2009. Supporting a

Virtual Community of Tutors in Experience

Capitalizing. International Journal of Web Based

Communities, 5(3), 407-427.

Guy, M. and Tonkin, E., 2006. Folksonomies: Tidying up

Tags? D-Lib Magazine, 12(1). Available at:

http://www.dlib.org/dlib/january06/guy/01guy.html.

Koh, J. and Kim, Y., 2004. Knowledge sharing in virtual

communities: an e-business perspective. Expert

Systems with Applications, 26(2), 155–166.

Lave, J. and Wenger, E., 1991. Situated Learning.

Legitimate Peripheral Participation. Cambridge, UK:

Cambridge University Press.

O'Reilly, T., 2005. What Is Web 2.0: Design Patterns and

Business Models for the Next Generation of Software.

O'Reilly Media. Available at: http://www.oreillynet.com

/pub/a/oreilly/tim/news/2005/09/30/what-is-web-20.html.

Pan, S. and Leidner, D., 2003. Bridging Communities of

Practice with Information Technology in Pursuit of

Global Knowledge Sharing. Journal of Strategic

Information Systems, 12, 71-88.

Snyder, W.M., Wenger, E. and de Sousa, B.X., 2004.

Communities of Practice in Government: Leveraging

Knowledge for Performance. The Public Manager,

32(4), 17-21.

Star, S.L. and Griesemer, J.R., 1989. Institutional Ecology,

`Translations' and Boundary Objects: Amateurs and

Professionals in Berkeley's Museum of Vertebrate

Zoology, 1907-39. Social Studies of Science, 19(3),

387-420.

Von Krogh, G., 1999. Developing a knowledge-based

theory of the firm, St. Gallen, Switzerland: University

of St. Gallen. Available at: http://

www.dialogonleaders

hip.org/docs/vonKrogh-1999.pdf.

Wenger, E., 1998. Communities of practice: Learning,

meaning, and identity, Cambridge: Cambridge

University Press.

Wenger, E., White, N., Smith, J.D., Rowe, K. and

CEFRIO, 2005. Technology for Communities. Guide

to the implementation and leadership of intentional

communities of practice. Work, learning and

networked. 71-94.

Ziovas, S. and Grigoriadou, M., 2007. Boundary Crossing

and Knowledge Sharing in a Web-Based Community.

IADIS Web Based Communities Conference.

Salamanca, Spain, 248-256.

KMIS 2009 - International Conference on Knowledge Management and Information Sharing