Facilitating Experience Sharing in Groups
Collaborative Trace Reuse and Exploitation
Qiang Li, Marie-Hélène Abel and Jean-Paul Barthès
UMR CNRS 7253 Heudiasyc, Université de Technologie de Compiègne, Centre de Recherches de Royallieu 60205,
Compiègne, France
Keywords:
Collaborative Working Environment, Trace-based System, Collaborative Trace, Collaborative Engineering,
Experience Management.
Abstract:
In the context of a web-based collaborative working environment, any interactive activity among the actors
themselves or between the actor and the system in the collaborative workspace produces a set of Collaborative
Traces (CT). A collaborative trace reflects the group’s working experience from past actions. Indeed, sys-
tematical experience reuse in organizations can sustain the complex project completion and problem solving
by exploiting collaborative traces. This paper proposes a method and fundamental principles to enhance the
exploitation of collaborative traces. Grounded on our previous work that defined a collaborative trace and
proposed a corresponding model, we define a model of complex filter and discuss its possible functionalities
according to the real needs and constrained by technical restrictions. The filter is the primary way for facil-
itating collaborative trace reuse. Using a collaborative platform E-MEMORAe2.0, we apply our model and
validate several complex filters in two practical situations.
1 INTRODUCTION
Due to the rapid changes in information tech-
nology, people can work together using new and
faster web-based collaborative working environment
(CWE) with less restrictions due to time or geo-
graphic position, and even to language or culture.
Such environments can strongly promote and en-
hance different aspects of computer-supported coop-
erative/collaborative work, e.g. the process of or-
ganizational knowledge management, group commu-
nication or decision making. In a typical collabo-
rative workspace, users can send email, edit wikis,
share documents or have a video conference, and
all such interactions with the systems or with other
members of the group leave collaborative traces that
contain information about the collaborative activities
(Li et al., 2012a), (Li et al., 2012b). Indeed, re-
search issues concerning traces is at the intersection
of the three fields of study: Knowledge Manage-
ment(KM), Information Sharing(IS) and Experience
Management(EM).
In this article, we do not intend to discuss the
three important concepts in the IT literature, but ac-
cept the common definitions: information is "process
data" (Zins, 2007), knowledge is "authenticated infor-
mation" (Dretske, 1981), (Machlup, 1980) and expe-
rience is "a special case or a refined form of knowl-
edge in a higher level" (see (Sun and Finnie, 2005)
and (Schneider, 2009)). According to Clauzel and
his colleagues, traces can be considered as a kind
of "knowledge sources" to represent users’ experi-
ences in synchronous collaborative learning activities
(Clauzel et al., 2011). Moreover, Mille and his team
claim that the knowledge of both individual and group
can be captured from the modeled traces (Champin
et al., 2004). Later, they explain that interaction traces
reflect experience more than simple knowledge for
complex task support in computer-mediated environ-
ment. Meanwhile, they propose a framework to as-
sist Trace-Based Systems creation (Laflaquière et al.,
2006). Further, Laflaquière et al. state that the type
of trace from the past interactions can be applied to
measure the personal experience (Laflaquière et al.,
2010).
In the domain of personal experience manage-
ment, the above research works greatly enrich the
trace theory and also provide rich directions for the
practical applications. However, not enough attention
is given to the issue of experience sharing and reuse
for group in collaborative working environment. In
the context of CWE, this issue concerns three aspects
21
Li Q., Abel M. and Barthès J..
Facilitating Experience Sharing in Groups - Collaborative Trace Reuse and Exploitation.
DOI: 10.5220/0004132900210030
In Proceedings of the International Conference on Knowledge Management and Information Sharing (KMIS-2012), pages 21-30
ISBN: 978-989-8565-31-0
Copyright
c
2012 SCITEPRESS (Science and Technology Publications, Lda.)
of group experience management: (i) defining differ-
ent kinds of trace in group; (ii) modeling these traces
with a view to support collaborative work; (iii) ex-
ploiting the defined traces in line with the group and
personal needs. Although the interactive activities in
CWE are numerous and intricate, the main one is col-
laboration. Thus, the trace from collaborative activi-
ties is named "Collaborative Trace (CT)" and defined
as follows: "A Collaborative Trace is a set of traces
that are produced by a user belonging to a group and
aimed at that group" (Li et al., 2012b). This article is
based on our previous research results (the CT defini-
tion and the CT model, refer to (Li et al., 2012a) and
(Li et al., 2012b)), and concentrated in some possible
methods to exploit and reuse the collaborative trace.
This paper is structured as follows: Section 2 de-
scribes some remarkable definitions and projects of
Trace in the literature, and the definition of collabora-
tive trace is reviewed; Section 3 explains the collab-
orative activities in shared workspace and recall the
principals and some basic notations of our proposed
collaborative trace model; Section 4 analyzes the var-
ious possibilities for exploiting collaborative traces in
group spaces and introduces several practical exam-
ples; Section 5 cover the evaluation of our model and
the exploitation process in the collaborative platform:
E-MEMORAe 2.0; Finally, we conclude with a sum-
mary and discuss future work in Section 6.
2 DEFINITION OF A TRACE
In the world of nature, usually, a trace is a mark, an in-
dication or an object denoting the existence or passing
of activities, e.g. a series of animal footprints in the
wood. It strongly relies on the effective "actions" and
the surrounding "environment". As a stretch of the
original meaning, in the domain of computer science,
a trace always comes from the observation of the in-
teractive activities between the user and the system.
Almost a decade ago, Mille and his colleagues pro-
posed an approach named Musette (Modeling USEs
and Tasks for Tracing Experience, see more details in
(Champin et al., 2003), (Champin et al., 2004)). The
objective of Musette is to "capture a user trace ac-
cording to a general use model describing the objects
and relations handled by the user of the computer sys-
tem". In this case, primitive trace is collected and an-
alyzed as a "task-neural knowledge base" for the ex-
perience reusing to support user’s reflexivity. What’s
more, a generic framework for experience modeling
and experience management are mentioned and dis-
cussed in details (Champin et al., 2003), (Champin
et al., 2004). Indeed, trace
1
is considered as a variable
or a tool to measure the user’s experience for the past
interactions.
On the basis of their results, Laflaquière and his
colleagues found that the trace can be used to solve
some of the important issues in experience manage-
ment, e.g. "the activity reflexivity and the experi-
ence reuse". They defined a trace as "temporal se-
quences of observed items". Besides, a framework
was introduced to support the Trace Based Systems
(TBS) that focused on the processing of trace ex-
ploitation (Laflaquière et al., 2006). With minor vari-
ance, Clauzel and his colleagues defining an interac-
tion trace as: "histories of users’ actions collected in
real time from their interactions with the software"
(Clauzel et al., 2009). Zarka et al. defined a trace
of interaction as " a record of the actions performed
by a user on a system, in other words, a trace is a
story of the user’s actions, step by step" (Zarka et al.,
2011). In the TRAIS project (Personalized and Col-
laborative Trails of Digital and Non-Digital Learning
Objects)
2
, the researchers analyze a trace that can be
considered as a sequence of actions in an hypermedia
environment to identify the users’ overall objective.
From another perspective, Settouti et al. defined a nu-
merical trace as a "trace of the activity of a user who
uses a tool to carry out this activity saved on a nu-
merical medium" (Settouti et al., 2009). They applied
the framework of trace-based system in Technology-
Enhanced Learning (TEL) Systems that can meet the
needs of personal services.
From the definition of interaction trace, we intro-
duced the new concept of Collaborative Trace (CT)
and recall its definition: "A Collaborative Trace is a
set of traces that are produced by a user belonging to
a group and aimed at that group" (Li et al., 2012b). In
the following section, a brief summary of our collab-
orative trace model (Li et al., 2012a) (Li et al., 2012b)
is provided with some basic notations.
3 COLLABORATIVE ACTIONS IN
GROUP SHARED SPACE
3.1 Collaborative Activities
A collaborative working environment (CWE) repre-
sents a kind of computer-supported working environ-
ment that "consists of a network of spatially dispersed
actors (either humans or not) that play different roles
1
In this article, we do not make a difference between
trace, interaction trace and trace of interaction unless ex-
plained in particular.
2
http://www.noe-kaleidoscope.org/telearc/
KMIS2012-InternationalConferenceonKnowledgeManagementandInformationSharing
22
and cooperate to achieve a common goal"(Angelaccio
and D’Ambrogio, 2007). It stems from the concept
of "virtual workspaces" (see (Schaffers et al., 2006))
and can be used to assist both the individual work and
the cooperative work, e.g., e-work and e-professional
(Prinz et al., 2006b). With various information and
communication technologies and tools, group users
could conduct their collaborative work through the
CWE (Prinz et al., 2006a). Actually, very basic fac-
tors found in CWE facilitate knowledge and informa-
tion sharing in group (Patel et al., 2012).
In software engineering principally, collaborative
activities can be divided into four types: "Mandatory,
Called, Ad hoc, and Individual" (Robillard and Ro-
billard, 2000), e.g. the scheduled video conference,
sending e-mail, or document management. For a typ-
ically CWE, most of these activities happen in the col-
laborative workspace (shared workspace) (Martinez-
Carreras et al., 2007). With the popularity of Inter-
net and the development of wireless technology, there
are less limitations due to time or space for partici-
pants, therefore, CWE inherits and extends this idea
from the design theory of groupware. In the early re-
search stages, a shared workspace is designed as "a
form of an electronic white-board" that helps collabo-
rators draw or write (Whittaker et al., 1993). With-
out a doubt, communication (e.g. video and audio
conference) and information sharing (e.g. exchanges
of messages or files) are one of the elementary parts
of shared workspace functions. In addition to this,
during the past decade, knowledge management (e.g.
document management, group wikis and task man-
agement) and application sharing (work in the same
application in real-time) have expanded the function-
ality of CWE and added new features to the shared
workspace.
All interactions or actions that concern diverse
functionalities of CWE in the shared workspace can
be recorded by traces. Thus, a trace model is neces-
sary and strongly required in the process of experi-
ence management. It is not only the historical list that
shows the user’s past actions, but also the previous
"experiences" to help perceive and interpret clearly
his interactions with systems. The trace model that
was proposed by Clauzel and his colleagues (Clauzel
et al., 2009) for the project ITHACA represents and
visualizes traces in the context of synchronous col-
laborative learning platforms. To address similar is-
sues, Lafifi (Lafifi et al., 2010) and his colleagues in-
troduced a trace model for the project SYCATA that
they concentrated on the whole architecture of the col-
laborative learning system. In a different approach,
the trace model was proposed by Sehaba (Sehaba,
2011) and deals with the transformation process for
the adaptation of the shared trace in accordance with
the user’s profile. For CWE, a collaborative trace
model could facilitate analyze and reuse knowledge
and experience in group. It focuses on the activi-
ties that involve or engage the collaborators in group
shared workspace.
3.2 Collaborative Trace Model
Before explaining our model, a simple example is in-
troduced. Suppose that in an established CWE, some
engineers collaborate within a project. John finds a
crucial problem that may be helpful for all the group
members. So, first of all, he sends a mail to the group
(every member in this group), then creates a new en-
try on this issue in group’s wikis, and finally shares his
solution (a pdf document) in the group workspace. In
the meantime, Tom and Peter, whose views are simi-
lar but different from John’s on this problem, both re-
quest a video conference with John in the reply email.
John receives the emails and agrees on a video con-
ference with Tom and Peter. At last, they obtain a
satisfactory answer for this problem in the subgroup
meeting.
Thinking back the meaning of an interaction trace,
apparently, there exist three basic factors concern-
ing the trace: (i) "Emitter" who produces the trace;
(ii) "Receiver" who obtains the trace (the target of
the trace); (iii) "A property and a corresponding val-
ues", that are the elements of the active environment
where the trace is generated and utilized. In practical
web-based CWE, the definition of "Emitter" and "Re-
ceiver" depends on the structure of the collaboration
group. A collaboration group is generally defined as a
set of some users with a same collaborative objective:
g
j
= {u
i
,u
k
,...,u
m
}
and may contain several subgroups and independent
users. Moreover, a single user can be considered as
a special type of collaborative group (a group of one
person): g
0
i
= {u
i
}.
Therefore, a trace is formally defined as:
t
k
i, j
=< E
i
,D
j
,Q
k
>
where t
k
i, j
is the kth trace sent by the ith Emitters E
i
(a
set of users), and received by the jth Receivers D
j
(a
set of users), and Q
k
is a subset of pairs of the set Q,
each element including a property and a value. Differ-
ent situations of Emitters and Receivers lead to iden-
tify three types of traces (Li et al., 2012a).
3.2.1 Collaborative Trace
A collaborative trace can be regarded as a type of trace
that satisfies the conditions:
E
i
= g
0
i
= {u
i
}
FacilitatingExperienceSharinginGroups-CollaborativeTraceReuseandExploitation
23
and
D
j
6= g
0
i
This means the trace is the result or the effect of
an operation that has been done by a "Emitter" and
then flows to another user or to a group. In partic-
ular, we can classify different types of collaborative
traces based on the relations between "Emitter" and
"Receiver":
i) The trace is produced and transferred within a
group:
u
i
∈ g
k
,D
j
⊂ g
k
That is to say, the emitter is belonging to the receivers
group. However, the relations between D
j
and g
k
in-
dicate that there exist two types of sub-situations:
• (a) The trace is between the subgroups:
u
i
∈ g
k
,D
j
⊆ g
k
For instance: a member sends an email to several
group members that constitute a subgroup. This
type of trace records the collaborative activities in
subgroups.
• (b) The trace is inside the whole group:
u
i
∈ g
k
,D
j
= g
k
For example: a member announces the details of
the project schedule in group (that a message is
sent to all the group members). All the group’s
activities are recorded by this type of trace.
ii) The trace is between two groups:
∃g
k
,u
i
/∈ g
k
and
D
j
⊆ g
k
For instance: some groups work together for a
project, and documents or messages are shared be-
tween groups. In Figure 1, we can clearly see the dif-
ferences between such traces.
Figure 1: Example of two types of collaborative trace.
In order to analyze and reuse collaborative traces,
a filter is applied as a tool or a pattern in the CT
model. The basic component of a filter is the extrac-
tor (operators to access some part of the trace), then
the elementary filters, and last, the complex filter (a
combination of elementary filters) (Li et al., 2012b).
As a matter of fact, the most important part is the de-
sign of elementary filters. It can be considered as a
predicate testing the value associated with a specific
property. There may be many elementary filters asso-
ciated with a single property. Formally, an elementary
filter is defined as:
ξ : V ×V → B, where B = {true, f alse}
With the elementary filters, we can analyze a
set of particular traces according to our needs:
{t |ξ
k
j
(α(t, p
j
),v
m
)}, where α is the value extractor,
and ξ
k
j
is one of the operators associated with prop-
erty p
j
. v
m
is a reference value. For example: we’d
like to find the traces that mention female members in
the group. We apply
ξ
member
sex
≡ f emale − member(α(t, sex), f emale)
Concisely, a collaborative trace model is a triple
structure: (G,Q,Ξ), where G is the set of users:
G = {g
j
}, that for ∀E
i
⊂ G,∀D
j
⊂ G, they meet the
conditions: E
i
= g
0
i
= {u
i
} and D
j
6= g
0
i
. Q is a set in
which each element includes a property and a value:
Q = P × V = {< p
l
,v
m
>}. P is a set of properties
(attributes of environment) and V is a set of values :
p
l
∈ P and v
m
∈ V . Z is a set of elementary filters:
Ξ = {ξ}. Indeed, the processes of programming can
be greatly simplified by the formulaic model of col-
laborative trace.
4 EXPLOITATION OF
COLLABORATIVE TRACES
Continuing the example above: (i) Naturally, the
email that is sent to the group by John is stored in
the group shared workspace, but has it been read by
all the members in group or just by a single person?
Same question for the shared pdf document: did they
open and view it or not? (ii) If Tom or Peter were
absent, it would affect the results of the video confer-
ence with John? In other words: do Tom and Peter
have the same competence on this problem and any
one of them could be substituted for the other? (iii)
Actually, John, Tom and Peter collaborate together
and can be regarded as a subgroup. Were the others in
the group satisfied by their answers to the problem? Is
the new added entry in the group wikis really helpful
for their project? In CWE, such questions are com-
mon but difficult to answer. They are directly relevant
to the issue of CT exploitation.
As we explained in Section 3, collaborative traces
record past interactive activities in a group shared
workspace and can be used as tools to enhance an ap-
plication, to generate adaptive scenarios and to assist
members in their collaborative tasks. In general, the
KMIS2012-InternationalConferenceonKnowledgeManagementandInformationSharing
24
collaborative activities produce more information and
knowledge than personal states. Therefore it may cre-
ate a large number of CTs in the group space. Ele-
mentary filters are limited, when screening and ana-
lyzing a large amount of CTs against actual demands.
A complex filter is thus proposed and designed to help
addressing this issue. It is defined as a logical com-
bination of elements of Ξ ( Ξ is the set of elementary
filters, Ξ = {ξ}).
Thus,
ζ : T × Ξ ×P ×V → B
An example of group collaborative trace would be
{t | t ∈ CT
i,l
∧ ξ
k
j
(α(t, p
j
),v
l
) ∧... ∧ ξ
n
m
(α(t, p
m
),v
s
)}
This allows selecting for example traces emitted by
a user, mentioning the concept of "culture", or traces
sent to a particular group during a specific week, or
traces of messages sent by a specific user to a specific
group, etc.
Three foundational parts constitute a primary
framework of trace-based systems ((Laflaquière et al.,
2006) and (Laflaquière et al., 2010)): (i) Collection;
(ii) Transformation; and (iii) Presentation. One can
clearly see this architecture Figure 2:
• Collection: this process uses diverse sensors and
collectors, in a web-based CWE, the main data
consists of text documents, hypertext documents,
linked structures, server logs, browser logs and so
on. The level of capture determines the diversity
of the values. Collecting can be done on-line or
off-line;
• Transformation: this part includes three func-
tions: filtration, calculation and analysis. The
output (CT) from the first process can be classi-
fied, analyzed, merged and edited automatically
or manually. The programming language and
practical system environment (e.g. the number of
users, the hardware support, etc.) directly affect
the efficiency and accuracy;
• Presentation: the last process utilizes the out-
comes from the transformation procedure. The
object is to explain the users’ finished "interactive
activities" and assist them in their future work.
To make the modeled traces easier to understand
and reuse, the interface design and the mode of
presentation (for instance: visualization, audio or
video) require serious consideration.
In CWE, the exploitation of collaborative traces is
principally focused on the transformation and the pre-
sentation process. Since CT signifies the collaborative
experience, it is an important issue concerning expe-
rience reuse in EM theory (for the general experience
reuse, refer to (Bergmann, 2002) and (Tautz, 2001)).
Figure 2: A primary framework of trace-based system.
CWE, like other application scenarios, e.g. electronic
commerce, diagnosis of complex technical equipment
or electronics design, has the following characteris-
tics:
• Knowledge Intensive: Collaborative knowledge,
e.g., about group project (e.g., project descrip-
tion and budgeting, task management, human re-
sources, re-set target), group member (e.g., back-
ground, competence and character etc.) and group
management (e.g., leadership and hierarchical re-
lationships) directly influences every phrase and
is enriched with group needs;
• Vague Collaboration Description: the goal of
group collaboration are often vague, incompletely
specified or even fickle. To clarify the problem
and the objective, regular meetings are recom-
mended for all the group members.
• Large Collaboration/Solution Space: the more
possible collaborations and solutions there are,
the larger the space would be in CWE and a single
collaboration or solution is not enough for a com-
plex project. Normally, these solutions depend on
the quantity of tasks and involved people;
• Group Size: different kinds of people (e.g., en-
gineers, experts or manager) are needed in every
process of problem solving and act in a collabo-
rative task. However, for the this issue, most of
the research works examine small size (Steiner,
1972) and (Ellis et al., 1991). A great challenge
for CWE is the large size of collaborative groups;
• Highly Dynamic: the rapid change and develop-
ment of technology has a great effect on the re-
newal of knowledge, the people involved, the po-
tential collaboration, the working style and so on.
Like the sketched situations above, a complex
project is heavily based on collaborative experience.
Collaborative traces sharing and reuse enable helping
FacilitatingExperienceSharinginGroups-CollaborativeTraceReuseandExploitation
25
individuals and groups to avoid making same mis-
takes over again. To understand the process of ex-
ploiting collaborative traces, four basic scenarios are
introduced as follows:
• Record and Analyze the Finished Collabora-
tive Activities: this scenario can be characterized
as "a dictionary of group collaborative activities
in accordance with the chronological order". All
the members in a group could distinctly see their
interactions and the corresponding results in the
group shared space, e.g. the usage status of shared
documents or the sent email may be opened by the
others;
• Assist Group Future Work: in this case, the fil-
tered CTs can be distinguished as "a guide" for
the future collaborations in groups. For example,
if a task that failed in several ways, we can avoid
doing the same mistake in the future. Besides,
some potential collaborations may be found by
their similar CTs, e.g. the comparable preference
of shared documents or entries in group wikis;
• Support Group Decision Making: in this situ-
ation, collaborators can review all their past de-
cisions with their consequences and the project
progress in the group. They can make a better
decision with such classified CTs. For instance:
during Tendering, the analysis of customer RFP
(Request For Proposal) within collaborators;
• Enrich Group Knowledge: in this circum-
stance, CTs reflect the needs and preferences
of groups, with recommendation strategies, new
knowledge is gained and shared in the collabora-
tive workspace. For example: from the preferred
books, links or videos, it is easy to recommend
more with similar topics.
In CWE, the benefits of CT exploitation in the
mentioned scenarios are beyond our expectations.
The major advantages can be summarized as below:
• Shorter Project Completion Time/Cycle: e.g.
the cost of time or group work efficiency;
• Improve Project Quality: e.g. from the reuse of
CTs, we would make less mistakes but have more
potent collaborations.
• Reduce Project Expenditure: e.g. from the
analysis of CTs, it is not difficult to identify col-
laborators’ contributions and attitudes.
5 APPLICATION
In this section, we evaluate our CT model and several
complex filters in a web-based collaborative platform
E-MEMORAe2.0 (Leblanc and Abel, 2008). Within
the MEMORAe approach (Abel and Leblanc, 2008),
E-MEMORAe2.0 (Figure 3) is combined with: (i)
Models stem from knowledge engineering to support
Knowledge Management; (ii) Semantic Web tech-
nologies to facilitate sharing and interoperability; (iii)
Web 2.0 technologies to promote the social processes.
Via this platform, both the fields of organizational col-
laboration and expertise can be enhanced by means
of ontologies that define knowledge in organization
(Abel and Leblanc, 2009). In a shared workspace,
the users can exchange messages, edit and annotate
shared documents, write wikis, share calendar and
so on. For the personal use, the user can navigate
through the shared ontologies; moreover he can also
organize and capitalize the resources (e.g. documents,
links and etc.). Up to now, within the range of this
platform, only two kinds of personal interactions are
recorded: (i) the access to concepts; (ii) the access to
resources; then presented in the "Navigation history."
In order to facilitate collaborative activities, appar-
ently, the personal traces are limited and weak. The
application of the CT model and of complex filters
directly meet this issue and are easy to use.
In our application, firstly, the CTs are stored in
accordance with the CT model conditions; then the
queries are done through the designed complex fil-
ters; lastly, the results are presented in a chart or
graph. Summarizing our case: the collaborative group
is formed by four members: Qiang, Étienne, Marie-
Hélène and Jean-Paul, formally, g
1
= {u
1
,u
2
,u
3
,u
4
}.
They cooperate in a project called "Trace". As
shown Figure 4, the group has four members and two
subgroups. Recall the general framework of trace-
based systems, the proposed model and filters are
mainly in the transformation and presentation parts.
Three factors of CT are collected and stored in a
MySQL database: the list "per_id" from the table
"mem_personne" is used to identify the members (e.g.
the E
i
and D
j
); the values and properties are decided
by the needs of practical scenarios (refer to differ-
ent methods of CT exploitation), but the "time" and
"date" of past interactions is determined as the "In-
dex" of CT (geographical position could be another
choice). We discuss respectively the two cases ("Con-
cepts" and "Resources") in the following part:
• For the Case: "Concepts," the three compo-
nents of collaborative trace are written as: (i)
E
i
="The administrator (one of the members who
is in charge of building the ontology, e.g. cre-
ation, connection of concepts etc.)"; (ii) D
j
="All
the group members" (g
1
), e.g. every member
can view and check the shared ontology in the
group workspace; (iii) Apart from "time" and
KMIS2012-InternationalConferenceonKnowledgeManagementandInformationSharing
26
Figure 3: The collaborative platform E-MEMORAe2.0 (in French).
"date" (formally, which can be written as v
1
),
the frequency/times (that can be considered their
preferences) is intended as another value (v
2
) for
the property (p
1
) "the concerned concepts in the
group shared ontology"; the upper part of Figure
5 shows the three most consulted concepts: "Col-
laborative Trace Definition", "Experience Man-
agement", "Collaborative Trace Model" during
one month (from 01/12/2012 to 02/12/2012).
The most relevant concept is "Experience Man-
agement" and the least is "Collaborative Trace
Model". The lower chart presents the trail of
the most consulted concept in time. In 06/Feb,
the group examined this concept three times, but
only once in 12/Jan. Therefore, we can clearly
obtain their preference and attention within this
ontology. The group’s attention for this concept
changes and is reduced after a period of time. This
phenomenon may imply that the group has an ex-
tensive understanding and shares common con-
clusions about this concept. Here, the complex
filter (ζ
1
) is designed to analyze the frequency and
the change of consulted concepts with time varia-
tion;
• For the Case: "Resources", (i) the "Emitter" (E
i
)
and "Receiver" (D
j
= g
1
) are the same as in the
above situation, but the traced object is the shared
file (two categories: one involving pdf and doc
documents, the other concluding video and im-
ages) that concerns the three concepts in the ontol-
ogy. (ii) The property (p
2
) is "the shared files for
the most checked concept in the group shared on-
Figure 4: The Group Structure.
Figure 5: An example of the case "Concepts".
tology". Besides, for the values, one (v
3
) is the sit-
uation of shared files (file types and quantity) and
FacilitatingExperienceSharinginGroups-CollaborativeTraceReuseandExploitation
27
another (v
4
) is the frequency/times of the service
situation for each type of the file. As shown in
Figure 6, the upper chart demonstrates the quan-
tity of each type file that has been shared in group
workspace during one month (same as the case
"Concepts": from 01/12/2012 to 02/12/2012).
We can see that the most frequently consulted
concept (from the first case), Experience Manage-
ment (EM), is connected with several files: three
pdf, three doc, one video and two images in this
period. And the least interesting concept, "Col-
laborative Trace Model," has the most connected
files: four pdf, four doc, two videos and three im-
ages. The lower figure presents the state of service
for the three shared pdf documents that is associ-
ated with the concept EM. The "frequency" signi-
fies the number of times the file has been opened
("open this file"). For the "PDF2" ("Note I of
EM"), it is obvious to see that Étienne (u
2
) had
a lack of interest and has never opened this file.
However, it was certainly read several times by
Qiang (u
1
) and Jean-Paul (u
4
). Moreover, Marie-
Hélène (u
3
) is more interested in "PDF3" ("EM
vs. KM") than in other documents ("Note I of
EM" or "Trace and EM"). In this case, the com-
plex filter (ζ
2
) is used to help observe, compare
and analyze the group’s preference and members’
contributions in collaborative workspace.
Figure 6: An example of the case "Resources".
As a consequence of the filtered CTs, some poten-
tial collaborative relations that tightly rely on their
"preferences" and "contributions" will be rec-
ommended within group members, for example:
Qiang (u
1
) and Jean-Paul (u
4
) collaborate with the
subject of "PDF2". Furthermore, the competence
or knowledge background within group members
can be identified with more complex filter, e.g.
from the similarity of the shared files. It is helpful
to allocate the tasks or replace a member in some
particular situation. For instance if we are missing
an expert in a group, we could propose another
expert for this task. Without a doubt, the group’s
knowledge is enriched by these shared files and
the ontology in the group workspace. Using the
filtered CTs, we could understand the service state
of the shared knowledge, e.g. the level of knowl-
edge usage and the type of knowledge requested
in the group.
In the E-MEMORAe2.0 platform, the group col-
laborative working experiences are modeled and
reused by the application of collaborative traces. CTs
model and the complex filters can also be applied to
other ends, like in supporting the Tendering process
(in railway applications) (Penciuc et al., 2010) and the
organizational Content Management (Deparis et al.,
2011). Moreover, our model can be expanded to dif-
ferent collaborative platforms, e.g. in an agent-based
CWE, or collaborative learning systems.
6 CONCLUSIONS
In this paper, we proposed using a complex filter ap-
proach to facilitate group experience management and
support collaborative works in the context of web-
based CWE (Section IV). This approach has been de-
veloped from the results of our previous work: Col-
laborative Trace Definition and Model. A literature
review of traces and some basic notations of our CT
model were introduced and discussed in Section II
and Section III. Furthermore, to validate this method
and some principles of exploitation of CTs, two typi-
cally use cases based on the collaborative platform E-
MEMORAe2.0 were compared and explained in Sec-
tion V. Exploiting collaborative traces concerns sev-
eral critical issues in the different fields of EM study,
which can be summarized into three principal points:
(i) Experience reuse in collaborative groups, not only
for personal usage; (ii) Analysis and modeling of the
finished/past collaborative activities and interactions;
(iii) Enriching organizational knowledge and support-
ing the Knowledge Management process in CWE. Us-
ing the CT model and the complex filters are also in-
teresting in other fields, such as: multi-agent systems,
or social systems.
Actually, there exists many other collaborative in-
teractions in the platform E-MEMORAe 2.0, for ex-
ample: in a group shared work space, that adds con-
cepts or resources, creates comments in wikis, etc.
We are interested to test our model and filter for more
types of collaborative interactions. Besides, in the
coming autumn semester, we will apply our applica-
tion to the collaborative learning scenarios via plat-
KMIS2012-InternationalConferenceonKnowledgeManagementandInformationSharing
28
form E-MEORAe 2.0 in the University of Technol-
ogy of Compiègne. It will be used to support the stu-
dents’ collaborative activities that can be traced and
analyzed by our application.
REFERENCES
Abel, M.-H. and Leblanc, A. (2008). An operationnaliza-
tion of the connections between e-learning and knowl-
edge management: the memorae approach. In Pro-
ceedings of the 6th IEEE International Conferences
on Human System Learning, Toulouse, France, pages
93–99.
Abel, M.-H. and Leblanc, A. (2009). A web plaform for
innovation process facilitation. In IC3K 2009 In-
ternational Joint Conference on Knowledge Discov-
ery, Knowledge Engineering and Knowledge Manage-
ment, pages 141–146, Madeira Portugal. ACM.
Angelaccio, M. and D’Ambrogio, A. (2007). A model
transformation framework to boost productivity and
creativity in collaborative working environments. In-
ternational Conference on Collaborative Computing:
Networking, Applications and Worksharing, 0:464–
472.
Bergmann, R. (2002). Experience management: founda-
tions, development methodology, and internet-based
applications. Springer-Verlag, Berlin, Heidelberg.
Champin, P.-A., Prié, Y., and Mille, A. (2003). MUSETTE:
Modeling USEs and Tasks for Tracing Experience. In
Workshop 5 ’From Structured Cases to Unstructured
Problem Solving Episodes For Experience-Based As-
sistance’, ICCBR’03, pages 279–286.
Champin, P.-A., Prié, Y., and Mille, A. (2004). MUSETTE
: a framework for Knowledge from Experience. In
EGC’04, RNTI-E-2 (article court), pages 129–134.
Cepadues Edition.
Clauzel, D., Sehaba, K., and Prié, Y. (2011). Enhancing
synchronous collaboration by using interactive visu-
alisation of modelled traces. Simulation Modelling
Practice and Theory, 19(1):84–97.
Clauzel, D., Sehaba, K., and Prié, Y. (2009). Modelling
and visualising traces for reflexivity in synchronous
collaborative systems. In Proceedings of the 2009
International Conference on Intelligent Networking
and Collaborative Systems, INCOS ’09, pages 16–23,
Washington, DC, USA. IEEE Computer Society.
Deparis, E., Abel, M.-H., Lortal, G., and Mattioli, J. (2011).
Knowledge capitalization in an organization social
network. In Filipe, J. and Liu, K., editors, Proceed-
ings of the International Conference on Knowledge
Management and Information Sharing, pages 217–
222, Paris, France. SciTePress.
Dretske, F. I. (1981). Knowledge and the Flow of Informa-
tion, volume 61. MIT Press.
Ellis, C. A., Gibbs, S. J., and Rein, G. (1991). Group-
ware: some issues and experiences. Commun. ACM,
34(1):39–58.
Lafifi, Y., Gouasmi, N., Halimi, K., Herkas, W., Salhi, N.,
and Ghodbani, A. (2010). Trace-based collaborative
learning system. Journal of Computing and Informa-
tion Technology, 18(3):1–14.
Laflaquière, J., Mille, A., Ollagnier-Beldame, M., and Prié,
Y. (2010). Modeled traces for systems allowing reflec-
tion on personal experience. International Journal of
Human-Computer Studies, page 12p.
Laflaquière, J., Settouti, L. S., Prié, Y., and Mille, A. (2006).
Trace-based framework for experience management
and engineering. Lecture Notes in Computer Science,
4251(4251):1171–1178.
Leblanc, A. and Abel, M.-H. (2008). E-memorae2.0: an
e-learning environment as learners communities sup-
port. International Journal of Computer Science and
Applications, Special Issue on New Trends on AI Tech-
niques for Educational Technologies, 5(1):108–123.
Li, Q., Abel, M.-H., and Barthès, J.-P. (2012a). A model of
collaborative trace to enrich group experience. In The
5th International Conference of the World Summit on
the Knowledge Society, Accepted, Rome, Italy.
Li, Q., Abel, M.-H., and Barthès, J.-P. (2012b). Shar-
ing working experience: Using a model of collabo-
rative traces. In The 16th International Conference
on Computer Supported Cooperative Work in Design,
Accepted, Wuhan, China.
Machlup, F. (1980). Knowledge, its creation, distribu-
tion, and economic significance. Princeton University
Press.
Martinez-Carreras, M. A., Ruiz-Martinez, A., Gomez-
Skarmeta, F., and Prinz, W. (2007). Designing a
Generic Collaborative Working Environment. In IEEE
International Conference on Web Services, ICWS
2007., pages 1080–1087.
Patel, H., Pettitt, M., and Wilson, J. R. (2012). Factors of
collaborative working: A framework for a collabora-
tion model. Applied Ergonomics, 43(1):1 – 26.
Penciuc, D., Abel, M.-H., and Van Den Abeele, D. (2010).
Requirements and modelling of a workspace for tacit
knowledge management in railway product develop-
ment. In Proceedings of the International Conference
on Knowledge Management and Information Sharing
International Conference on Knowledge Management
and Information Sharing, KMIS 2010, pages 61–70,
Valence Espagne.
Prinz, W., Dustdar, S., and Ballesteros, I. L. (2006a). New
Collaborative Working Environments 2020. European
Commission Information Society and Media.
Prinz, W., Loh, H., Pallot, M., Schaffers, H., Skarmeta, A.,
and Decker, S. (2006b). Ecospace - towards an inte-
grated collaboration space for eprofessionals. In In-
ternational Conference on Collaborative Computing
Networking Applications and Worksharing, pages 1–
7.
Robillard, P. N. and Robillard, M. P. (2000). Types of col-
laborative work in software engineering. Journal of
Systems and Software, 53(3):219 – 224.
Schaffers, H., Brodit, T., and Pallot, M. (2006). The future
workspace: mobile and collaborative working per-
spectives. Telematica Instituut.
FacilitatingExperienceSharinginGroups-CollaborativeTraceReuseandExploitation
29
Schneider, K. (2009). Experience and Knowledge Manage-
ment in Software Engineering. Springer.
Sehaba, K. (2011). Adaptation of shared traces in e-learning
environment. 2011 IEEE 11th International Confer-
ence on Advanced Learning Technologies, pages 103–
104.
Settouti, L. S., Prié, Y., Marty, J.-C., and Mille, A.
(2009). A Trace-Based System for Technology-
Enhanced Learning Systems Personalisation . In
The 9th IEEE International Conference on Advanced
Learning Technologies.
Steiner, I. D. (1972). Group process and productivity. Aca-
demic Press.
Sun, Z. and Finnie, G. R. (2005). Experience management
in knowledge management. In International Confer-
ence on Knowledge-Based and Intelligent Information
& Engineering Systems, KES (1), pages 979–986.
Tautz, C. (2001). Customizing software engineering expe-
rience management systems to organizational needs.
Fraunhofer IRB Verlag, Stuttgart.
Whittaker, S., Geelhoed, E., and Robinson, E. (1993).
Shared workspaces: how do they work and when are
they useful? International Journal of Man-Machine
Studies, 39(5):813 – 842.
Zarka, R., Cordier, A., Egyed-Zsigmond, E., and Mille, A.
(2011). Trace replay with change propagation impact
in client/server applications. In Ingénierie des con-
naissances (IC 2011), Sciences exactes et naturelles,
pages 607–622. Publibook.
Zins, C. (2007). Conceptual approaches for defining data,
information, and knowledge. Journal of the American
Society for Information Science, 58(January):479–
493.
KMIS2012-InternationalConferenceonKnowledgeManagementandInformationSharing
30
