Integrating Awareness Sources in Heterogeneous
Collaboration Environments
Vijayanand Bharadwaj, Y.V. Ramana Reddy, Sumitra Reddy
Concurrent Engineering Research Center
Lane Dept. of Computer Science and Electrical Engg.
West Virginia University
Morgantown, W.V., U.S.A, 26505
Abstract. Collaboration in heterogeneous environments involves dealing with
variety
of information sources that generate information that users need to be
aware of. Users must be empowered to tailor the quality of awareness informa-
tion. Heterogeneity of sources and media adversely affects the quality of group
awareness. We propose a solution in terms of integrating the sources at the in-
formation level and provide a model for the same. We discuss our progress in
designing the model, its utility and benefits. We believe that such a unifying
framework can increase the effectiveness of group awareness in supporting co-
ordination and execution of collaborative work.
1 Introduction
Collaboration in heterogeneous environments has become a necessity. Heterogeneity
is in the plethora of applications, supported by hardware and communication infra-
structure of varying capabilities. Added to this is user mobility. Awareness among
such groups is essential. Many groupware applications are designed to support group
awareness but often most groups use general-purpose applications not designed to
support awareness. Awareness propagation is effective if the appropriate amount of
information, relevant to the user’s sphere of activity is delivered in an unobtrusive
manner, without compromising the privacy and security constraints of the group. A
key design goal of awareness systems is to empower its users with awareness charac-
terized by a high degree of the above “quality factors”.
The heterogeneity of sources and mediums that generate awareness inform
ation
deeply impacts the quality factors. We examine this relation in detail, and the specific
problems. As part of the solution, we propose a means to effectively integrate aware-
ness information by means of an awareness model. The next section describes the
effect of heterogeneity on the quality factors. Section 3 outlines the solutions. Details
of our model are in Section 4, with an example. We conclude with validation steps.
Bharadwaj V., V. Ramana Reddy Y. and Reddy S. (2005).
Integrating Awareness Sources in Heterogeneous Collaboration Environments.
In Proceedings of the 2nd International Workshop on Computer Supported Activity Coordination, pages 18-27
DOI: 10.5220/0002577600180027
Copyright
c
SciTePress
2 Awareness in Heterogeneous Environments
2.1 Awareness in Group Work
Awareness has been well researched in information technology and social sciences.
Gross et al. [9] provide a good comparison of both perspectives. Various forms of
awareness and terminology have been defined. However we find that three aspects of
awareness in general are closely related. These are Quality of Awareness, the Aware-
ness Information Characteristics and Awareness Sources & Mediums.
1. Quality of Awareness: The quality of awareness impacts its role in group work
coordination and execution. Quality can be characterized by factors such as:
Relevancy: How relevant is the awareness information to my sphere of work, both
current and for the future (planning and coordination)? Information Overload: Am I
aware of the appropriate amount? Am I being inundated with too much or not receiv-
ing enough? Obtrusiveness: How distracting is this information to my current activ-
ity? Is it interrupting my current activity? Privacy: Can privacy in the form of access
control be established to ensure that only one with appropriate permissions is aware
of classified information?
2. Awareness Information Characteristics:
Type (What am I aware of): Is this about an individual’s activity or location, actions
on an artifact, conversation transcript, streaming video of a meeting, or a combination
of the above? Form: Is it text, audio, visual? Volume (How much of it am I aware
of): Am I aware of every email exchanged by the group or just a synopsis; is it a long
videoconference session? Do I receive a notification for every action taken on an
artifact? Frequency (How often do I get that information): In a highly interactive chat
session that I am monitoring, do I want to receive every line typed?
3. Sources and Medium:
Sources (Which is the information source): Email and IM messages, sensors relaying
location, an artifact (actions on it), camera, microphone, user’s keystrokes are all
sources of awareness information, capturing various aspects of work that other users
would be interested in. Sources impact the Type, Form, Volume and Frequency. For
instance a sensor-based application will communicate awareness information only
when the appropriate stimulus occurs and at the same frequency. High rate of key-
strokes can result in high frequency and volume of information relayed.
Medium (How do I obtain information): Wired and Wireless networks, closed-circuit
video, telephone (landlines, cellular) are all media. The capabilities of the Medium
impact Form, Volume and Frequency. High bandwidth network could provide
streaming video at much better quality than a dial-up connection.
Quality factors are evaluated from the perspective of the consumer of awareness and
how they influence his work. Awareness Characteristics are about the information
and so absolute. However the perceived quality of awareness does depend on the
characteristics, which in turn are determined by Sources and Mediums.
19
2.2 Impact of Heterogeneity on Awareness Quality
Let us consider a scenario with three groups creating documentation independently.
The team leader (TL) responsible for coordinating work among them is off-site as
part of duty. Each group uses a collaborative editor, and at close of business, emails
the TL their drafts. Upon review, the TL sends her comments to each group using
both email and/or phone. Conflicting schedules, work commitments and lack of infra-
structure keep her from joining their editing sessions. Relaying instructions at the end
of the day is not very efficient. Sections in the documents depend on the content of
the others. When inconsistencies are detected they have to be reworked (email is
checked the next day), leading to delays. The ability to simultaneously monitor the
status of three groups would be very valuable as well as knowing which parts they
were working on and their decisions in the process. Even if the actual artifact being
edited cannot be seen, the above “meta-information” is essential. Thus “knowing of”
something can be valuable even if actual artifacts cannot be accessed. In another
scenario a group in one location, needs to confer with a group at another location.
They engage in video conferencing. They may record their decisions on a whiteboard.
How could the TL be aware of this activity remotely without access to the video ses-
sion? In other words we see the use of applications (some not designed to support
awareness), running on infrastructure of non-uniform capabilities, being used in un-
anticipated situations i.e. heterogeneity at all levels. The TL cannot access relevant
information as her capabilities limit the type, volume and frequency. Thus heteroge-
neity impacts the quality of awareness information and its effectiveness. We are
interested in how users could control the quality factors of awareness in such envi-
ronments.
3 Addressing the Impact of Heterogeneity
The approach is to integrate the various sources and mediums of awareness in a man-
ner most suitable to the group effort. This integration has to occur at two levels.
3.1 Integrating Sources and Mediums
There needs to be mechanisms to tie various sources of information that exist in a
group effort. Systems such as email servers, group editors, applications such as word
processors, sensors and devices such as cameras, microphones, all generate informa-
tion of different types. Consumers of this information use web browsers, hand-held
devices, and applications (that are themselves sources) to obtain awareness. Commu-
nication occurs over different mediums on non-uniform infrastructure. Thus there
needs to be an “Awareness Middleware” that can bind all the above together. Spe-
cific interfaces would be needed to connect the sources to the middleware making
integration as easy as “plugging-in” the source. The use of non-electronic artifacts in
many collaboration environments poses a challenging integration problem. The mid-
dleware and interfaces would have to conform to security, reliability and perform-
20
ance. Notable pioneering efforts [13], [8], [1], [2], [7] have been made to create such
“awareness frameworks” for awareness propagation in groups.
3.2 Integrating Information
Different types of information generated (email messages, sensor coordinates, stream-
ing video) have no absolute relation, though related within the context of the group’s
activities. They have varied source specific characteristics (Form, Volume, Fre-
quency). These need to be woven to make the composite picture that is awareness.
This integration must occur at the information level. Consumers of awareness must be
able to determine quality and control it. Specifically, the implications of the aware-
ness requirements on information integration are:
1.There must be a transformation (mapping) of the source-specific nomenclature of
information generated, to the group’s terminology and definitions, as users have no
knowledge of the former. One way is by providing meta-tags to the information.
2.To retrieve all relevant information based on user’s sphere of activity, there must be
ways to relate information with respect to the common context, which is the group’s
work. Thus users can look for awareness information they need from their work per-
spective as opposed to querying about the source.
3. The integration process must be straightforward. Change in various factors is in-
herent in group work [4] and a simple process would help change management.
4. There must be means to express the Awareness Characteristics by which users can
determine and tailor the quality factors using mechanisms such as filtration.
5. Apart from manual selection, there could be agents providing context-sensitive
awareness. Mechanisms to enable such agents to match the user’s profile, and work
context to available information are required. This can support enhanced forms of
awareness such as “intersubjectivity” ( “i know, that you know, that I know”)[1] .
6. Since users’ capabilities to acquire awareness may not support all formats and
exact artifacts, there must be means to obtain meta-information about it.
7. Obtaining awareness would begin with searching and/or browsing for it. A com-
plete picture of information should be available to select from.
8. Historical awareness [12] of various aspects of the group’s work is often essential
and there must be means to retrieve such information.
Thus integration involves information transformation, relating different types and
adding meta-information description to enable the above features. It is much more
than creating databases, labeling and storing data. We propose a model as a basis for
integration. This provides a common logical framework thereby decreasing the com-
plexity in the integration process. Meta-information would be part of this framework.
The Awareness Frameworks reviewed emphasized awareness propagation. Sources
and information were integrated in some fashion to facilitate context-awareness and
user filtration. As sources to be used cannot be anticipated, our emphasis is on a
generic framework. Awareness quality, simplified integration process and adaptabil-
ity to changes in group work being issues of interest that we aim to address.
21
4 Model-Based Information Integration
4.1 The Awareness Model
Early work in modeling awareness has used spatial metaphors [3]. Medium, aura,
awareness, focus, nimbus and adapter were concepts used to model and enable inter-
action, determining how entities behaved in virtual worlds. This was generalized to
CSCW environments by a model [14] where a pool of objects and the relation be-
tween them was considered as the space. Users acting on these objects gain mutual
awareness due to their interaction. Graph theory was used to express notions like
strength of awareness. Another non-spatial model emphasized the effect of awareness
on behavior of the objects [15]. This was based on the reaction-diffusion metaphor.
Some efforts have been towards specific forms of awareness such as a model for
presence awareness using concepts such as location, presentity, watcher and vicinity
[6]. Another seeks to model cooperative awareness using three abstractions viz.
events, places and communities [11]. Inspired by the above we borrow some of the
key concepts and constructs. Our emphasis is on integrating the information sources
with respect to the quality factors of awareness. Figure 1 illustrates the model.
Focus of Attention (Focus): This top-level concept represents the focus of an aware-
ness consumer. A user’s Focus consists of all active sources that are providing aware-
ness. A Source is any entity that provides information (e.g. person, location, artifact).
A Focus is a unified view of all its sources and the corresponding events and interac-
tions occurring due to them (e.g. person’s actions, communication, actions on an
artifact). For instance the Focus could provide awareness about an activity in terms
of email messages exchanged by the users (sources) or awareness about people enter-
ing a room, modifying an artifact and leaving. A Focus can shift over time. The type
and number of sources win a Focus can change. Each Source has a corresponding
Medium, which delivers information. Its characteristics dictate the quality of aware-
ness information. Meta-information attributes describing awareness characteristics
qualify the Focus, its Sources and Medium and are essential to determine and tailor
quality. These attributes are:
Description: About what the Focus is providing.
Identifier: A unique identifier as there could be multiple Foci.
Start Time and End Time: Indicate the duration for which the Focus was active.
Source & Medium List: A list of sources and corresponding medium in the Focus.
(Source1, Medium1), (Source2,Medium2),….. …(SourceN, Medium N).
Source: In addition to its own Description and ID fields, each Source entity has:
Start Time: The time the source comes within the current Focus.
End Time: The time the Focus excludes it.
Foci List: Each source element may have multiple foci it is obtaining awareness from
(Focus1, Focus 2….FocusN) .
Information Content: This describes the information generated by the Source. Its
attributes are:
Type: About the Information (natural language and/or keywords).
22
Form: Text, Audio, Video stream, others.
Frequency: How often is the source generating the information. Could be in the form
of discrete notifications or continuous stream.
Total Volume: How much has been generated so far (or recorded).
Content: The actual information being generated, according to a content-specific
schema depending on the source. For example, actions taken by sources, actions on
artifacts, video streams, email messages and chat sessions would all be content.
Medium: Corresponding to each Source is its Medium. In addition to its own De-
scription and ID fields, each Medium entity has:
Medium Specific Characteristics: A set of attributes about the specific medium. For
example, the network characteristics for a wired or wireless computer network, cellu-
lar phone network or closed circuit television network would be of interest since they
would indicate how much of information could be delivered and in what manner.
Fig. 1. Awareness Model
4.2 Levels of Awareness
An individual’s Focus consists of the sources that are actively providing awareness.
Users could have multiple Foci, each providing information about different aspects of
the group’s work. There are other sources that the user could access but are not cur-
23
rently part of his Foci. The user is aware of their characteristics but not receiving
information content. This entire set of all active and potential sources is his Source
Superset, which the user can access based on his role and access control policies.
Sources outside his Superset are not accessible and invisible to him. Figure 2 illus-
trates this hierarchy. Finer levels could be enforced depending on the needs, what is
important is that our model seeks to provide such control.
The user’s view of all sources, media, and Foci in his Supersets with correspond-
ing meta-information is his “Awareness Map”, a concept inspired by [10]. On select-
ing an element of the map, one can zoom-in to get details about the content as well as
source and medium characteristics. Such a view helps ascertaining, who else is aware
of what I am aware or? Who is aware of me? intersubjectivity, and so on, which is
valuable in coordination. Context-sensitive awareness sources could be added manu-
ally or programmatically based on the user’s profile of work. Access control policies
limiting what users can access can be established with this view. A super user would
have unrestricted view of all users and their Source Supersets.
Fig. 2. Awareness Map
4.3 Illustration of Awareness Model
Figure 3 illustrates the scenario considered in section 2.2. The generic architecture of
an Awareness Framework is based on our review of related work that revealed com-
mon components. Applications (sources) generate awareness information, which is
communicated over certain media. They are bound to a central entity (by middleware)
that manages the information flow. Application specific interfaces are necessary to
connect them to the medium (very often computer networks). Users activities using
these applications are of interest to other users. These consumers access this informa-
24
tion using applications that are bound to the framework too. In our case the central
entity is the Awareness Model. We assume the collaborative editors used by the
groups are physically integrated. The TL accesses the AM through a portal applica-
tion. The collaborative editor is seen as one of the sources along with its characteris-
tics in the TL’s Source Superset. The TL adds it to his Focus and chooses details such
as when to be notified if certain events occur in the editing session, how often and so
on. This is an example of choosing the relevancy and frequency of information based
on the displayed characteristics. Being on a low bandwidth connection, the TL
chooses text-based notifications only. Though all events from the editing session are
communicated, the AM would relay information to the TL based on her preferences.
The TL being aware of the status could communicate with the groups through
email/telephone/IM or even the editor application itself assuming it has such capabili-
ties. We assume the editor communicates details such as, which changes were made
by whom, when and so on making it part of the Information Content. The TL should
be able to view this information as it occurs, or later. For e.g. the TL should be able to
query to see all the changes made by a particular group member. The editing groups
know they are connected to the AM through their Awareness Map views. They know
that the TL is aware of their activities. Thus there is mutual awareness. Let us assume
that one of the groups engages in a videoconference with another group as part of the
process. The user TL could expand his focus to include this source if necessary. The
TL may not be able to access the session but may be able to access a text transcript
that is recorded and stored in the AM after the conference. The AM is not meant to be
a data bottleneck; rather it is analogous to a lens over the awareness capabilities of the
group bringing into focus the pertinent aspects. Users may have direct lines of com-
munication with the source for actual data transfer.
5 Validation and Future Steps
Further work involves completing the model specification. The model will be vali-
dated with realistic collaboration scenarios and refined. Evaluating how useful the
model is in enabling the users to tailor the quality factors is essential. Mechanisms
used in integrating the information, displaying sources, supporting user browsing and
querying based on the quality are being developed. Also providing context-sensitive
awareness based on the user’s sphere of activity, profile and the current state of the
project is of interest. Another interest is evaluating the model in scenarios where there
exists a certain level of awareness in the group, with intervals requiring a “height-
ened awareness” as in emergency rooms and call centers [5]. Among the multiple
foci that a user has, some could diminish in “strength” when not important and others
could come into prominence during such periods. The Awareness Map concept will
be developed completely and a proto-type implemented.
25
Fig. 3. Awareness Model Illustration
6 Conclusion
Heterogeneous sources of awareness and media have a significant influence on the
quality of awareness in groups. We examined the major issues in empowering users
with the ability to tailor the quality of awareness and established the need for integra-
tion of sources at the information level. We proposed a generic unifying model to
provide the composite picture that awareness is. We aim to demonstrate the utility of
the model in tailoring awareness quality, simplifying the integration process and deal-
ing with changes in group work.
Acknowledgements
The authors would like to thank Dr. Srinivas Kankanahalli (U.S. Dept. of Education),
Dr V. Jagannathan, (West Virginia University) and Dr. Ranjan Sen (Microsoft Cor-
poration) for their valuable suggestions in this effort.
References
1. Anderson, K. M., Bouvin, N. O.: Supporting Project Awareness on the WWW with the
iScent Framework. In Proceedings of the International Workshop on Awareness and the
WWW, Part of the 2000 ACM Conference on Computer-Supported Cooperative Work.
Philadelphia, PA, USA. (December 2-6, 2000)
26
2. Bardram, J. E., Hansen, T.R.: The AWARE architecture: supporting context-mediated
social awareness in mobile cooperation. In Proceedings of the 2004 ACM conference on
Computer supported cooperative, CSCW’04, Chicago, Illinois, USA. (November 6-10,
2004) 192-201
3. Benford, S.D. and Fahlén, L.E.: A Spatial Model of Interaction in Large Virtual Environ-
ments. In: De Michelis, G., Simone, C. and Schmidt, K. (eds.): Proceeding of the 3rd Euro-
pean Conference on Computer Supported Cooperative Work ECSCW’93. Dordrecht: Klu-
wer Academic Publishers, Milano, Italy (1993) 109-124
4. Bharadwaj, V., Reddy, Y.V.R., Kankanahalli, S., Reddy, S., Selliah, S., Yu, J: Evaluating
Adaptability in Frameworks that Support Morphing Collaboration Patterns. In Proceedings
of the 13th IEEE International Workshops on Enabling Technologies: Infrastructure for
Collaborative Enterprises (WETICE'04), University of Modena and Reggio Emilia, Italy
(June 14 - 16, 2004) 186-191
5. Blandford, A., Wong, B.L.W.: Situation awareness in emergency medical dispatch. In
Intnational Journal of Human-Computer Studies, Volume 61, Issue 4, (2004) 421–452
6. Christein, H., Schulthess, P.: A General Purpose Model for Presence Awareness. In: Plaice
J. et al. (eds.): Distributed Communities on the Web: 4th International Workshop, DCW
2002, Sydney, Australia, April 3-5, 2002. Revised Papers. LNCS Volume 2468, Springer-
Verlag Berlin Heidelberg (2002) 22-34
7. Domingos, H.J., Preguica, N., Martins, J.L.: Coordination and Awareness Support for
Adaptive CSCW Sessions. In Proceedings of Fourth International Workshop on Group-
ware, CRIWG’98, Búzios, Rio de Janeiro, Brazil. (September 9-11, 1998) 21-37
8. Gross, T., Specht, M.: Awareness in Context-Aware Information Systems. In: Oberquelle,
H., Oppermann, R. and Krause, J. (eds.): Mensch & Computer - 1. Fachuebergreifende
Konferenz (Mar. 5-8, Bad Honnef, Germany), Teubner. (2001) 173-182.
9. Gross, T., Stary, C., Totter, A.: User-Centered Awareness in Computer-Supported Coop-
erative Work-Systems: Structured Embedding of Findings from Social Sciences. Interna-
tional Journal of Human-Computer Interaction (to appear)
http://www.uni-weimar.de/~gross/publ/IJHCI_gross_et_al_awareness.pdf (February 2005)
10. Gross, T., Wirsam, W., Graether, W.: AwarenessMaps: visualizing awareness in shared
workspaces. In Proceedings of the Conference on Human Factors in Computing Systems,
CHI’03, extended abstract. Ft. Lauderdale, Florida, USA, (2003) 784-785
11. Leiva-Lobos, E. P., Covarrubias, E.: The 3-Ontology: A Framework to Place Cooperative
Awareness. In: Haake, J.M. and Pino, J.A. (eds.): Groupware: Design, Implementation and
Use: 8th International Workshop, CRIWG 2002, La Serena, Chile, September 1-4, 2002.
Proceedings, Lecture Notes in Computer Science, Volume 2440, Springer-Verlag Berlin
Heidelberg (January 2002) 189-199
12. Nutter, D., Boldyreff, C.: Historical Awareness Support and Its Evaluation in Collaborative
Software Engineering. In Proceedings of the Twelfth International Workshop on Enabling
Technologies: Infrastructure for Collaborative Enterprises (WETICE’03), Linz, Austria.
(June 09 - 11, 2003) 171-176
13. Prinz, W. NESSIE: An Awareness Environment for Cooperative Settings. In Proceedings
of the Sixth European Conference on Computer-Supported Cooperative Work - ECSCW'99
(Sept. 12-16, Copenhagen, Denmark). Kluwer Academic Publishers, Dortrecht, NL, (1999)
391-410.
14. Rodden, T.: Populating the Application: A Model of Awareness for Cooperative Applica-
tions. In: Proceedings of the Conference on Computer Supported Cooperative Work
CSCW’96. ACM Press, Boston (1996) 87-96
15. Simone, C., Bandini, S.: Compositional features for promoting awareness within and across
cooperative applications. In Proceedings of the International ACM SIGGROUP Confer-
ence on Supporting Group Work: The Integration Challenge, GROUP’97, Phoenix, Ari-
zona, United States (1997), 358 - 367
27
