This document presents a review of patterns in collaborative work. It begins with an introduction on the role of patterns in facilitating collaboration. It then provides an overview and surveys various pattern approaches relevant to collaboration, discussing their characteristics. The approaches are evaluated based on characteristics like their model, language, categorization, target users, and ability to support collaboration, interactions, knowledge, and be event-aware. The results argue for approaches that can directly assist participants in automatic ways and develop tools to proactively recommend corrective actions. It also discusses the Collaboration Patterns Assistant, which uses an ontological model to capture collaboration patterns and provide pattern recommendations and assistance to users.
1. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
ACM 2010 International Conference on Supporting Group Work
A Review of Patterns in Collaborative Work
Yiannis Verginadis
Nikos Papageorgiou
Dimitris Apostolou
Gregoris Mentzas
Information Management Unit
Institute of Communication and Computer Systems
National Technical University of Athens
2. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Overview of presentation
Introduction
Survey of Patterns Approaches
Discussion
Collaboration Patterns Assistant
3. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Introduction
Collaboration is a critical business function that demands skills
and knowledge, spanning a wide range of domains (including
social, business and technical domains).
Due to the dynamic nature of Group Working environments, the
reuse of segments of collaborative work (Collaboration Patterns)
can constitute an advantage.
With this Review, we try to evaluate and categorize relevant
research and commercial pattern efforts in order to detect any
shortcomings or possible improvements / new directions of research
towards facilitating collaboration through patterns.
4. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Concept of Patterns
the word Pattern has appeared almost entirely due to the work of
Christopher Alexander in architecture.
he defined a pattern as a “morphological law that explains how to
design an artifact in order to solve a problem in a specific context”.
[Alexander et al., 1977]
the first notable publication in the context of software engineering
was the book "Design Patterns: Elements of Reusable Object-
Oriented Software" by the so-called Gang of Four [Gamma et al.,
1995], that advanced the popularity of patterns in computer science.
5. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
The Role of Patterns in Facilitating Collaboration
Although organizations seem to invest increasingly in
collaboration tools, they still lack knowledge on how to leverage
such tools effectively.
Collaborating partners may find the collaboration tools easy to
operate, but they cannot typically use their full potential.
Moreover, they cannot readily identify what the best collaboration
practice is, when to embark on such a practice and how to
facilitate it using available tools.
To alleviate problems in collaborative work and to encourage the
sustained use of collaboration tools in organizations, patterns have
been exploited as models for repeatable processes for recurring
high-value collaborative tasks.
6. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Overview of presentation
Introduction
Survey of Patterns Approaches
Discussion
Collaboration Patterns Assistant
7. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Survey of Patterns Approaches Relevant to
Collaboration (1/2)
Collaboration Patterns in Virtual Communities (CoP), (de Moor et
al., 2006)
Collaboration Engineering
Thinklets, (Briggs, 2003), (Kolfschoten et al., 2006)
Collaboration Patterns in Knowledge Management, (Qureshi et al., 2004)
Event Patterns, (Chakravarty et al., 2008), (Barros & Singh, 2007),
(Rapide, 1997)
Workflow and Process Patterns
Workflow Patterns, (van der Aalst et al.,2003, 2005)
Patterns for Business Process Management, (Atwood, 2006)
Patterns for Business Object Frameworks, (Barros et al., 2004)
Patterns for Inter-Organisational Business Processes, (Norta et al., 2006),
Grefen (2006)
MIT Process Handbook Pattern Model, (Malone et al., 2003)
Action Patterns in Business Process Models, (Smirnov et al., 2009)
Activity Patterns, (IBM, 2005), (Geyer et al., 2006)
Interaction Patterns
Service Interaction Patterns, (Barros et al., 2005)
Collaborative Interactive Applications Patterns, (Molina et al., 2006)
Interaction Patterns from SNA, (Dustdar et al., 2007)
Action Patterns in Virtual Collaboration, (Biuk-Aghai et al., 2005)
8. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Survey of Patterns Approaches Relevant to
Collaboration (2/2)
Patterns for Collaborative Working Environments, (Pattberg and
Flügge, 2007), (Eace, 2005)
Knowledge Flow Patterns, (Sarnikar et al., 2007)
e-Business Patterns (Zhao et al., 2007)
CPats (Verginadis et al., 2009)
Pattern Languages & Ontologies
Pattern Neighbourhoods, ConEngine, (Arevalo et al., 2004)
Pattern Language Mark-up Language (PLML), (Fincher et al., 2003)
Patterns Relations in CoPE, (Schuemmer, 2003)
GAMA, (Schuemmer and Lukosch, 2007)
Task Pattern Markup Language, (Gaffar et al., 2004)
Usability Patterns (Henninger and Ashokkumar, 2006)
Business Function Object Patterns, (OMG, 2004)
9. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Overview of presentation
Introduction
Survey of Patterns Approaches
Discussion
Collaboration Patterns Assistant
11. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Characteristics Used for the Review (1/5)
Model
for portraying a formal or informal modeling mechanism for patterns
We found that
the majority of researchers and commercial developers use a structured
descriptive mechanism (e.g.tabular format) for imprinting patterns,
just three efforts use only text for describing patterns (i.e. Patterns for BOF),
only six efforts present a complete model using a structured descriptive
mechanism combined with an ontology (e.g. Cpats).
Language
for capturing the approach taken for encoding the description of
patterns
We found that
five of the approaches examined do not provide a formal language (e.g.
Thinklets)
the majority introduces a formal language (e.g. Pattern Language Mark-up
Language ) often based on an ontology (e.g. Activity Patterns),
less than half of the approaches use a graphical notation to give a visual
overview of the designed patterns (e.g. Business Function Object Patterns)
to collaboration designers, facilitators and users.
12. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Characteristics Used for the Review (2/5)
Categorization
for pattern approaches: based on pattern objective/function and based
on pattern granularity
We found that
in most approaches categorization is based on collaboration objective or
function (e.g. Collaboration Patterns in Virtual Communities),
seven approaches present a categorization based on pattern granularity
(e.g. Usability Patterns),
just four approaches try to adopt both styles of categorization (e.g. Patterns
for Collaborative Working Environments).
Collaboration
is a key consideration of our survey.
We found that
the majority of the examined approaches consider collaboration as their
main focus; in some cases collaboration may even span across
organizational boundaries,
nine of them are considered to be able to support collaborations (e.g.
Workflow Patterns) but have not been motivated by the collaboration needs
at least in their early phases of their work.
13. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Characteristics Used for the Review (3/5)
Interaction
to depict whether an approach supports human tasks, services or both.
We found that
most of the examined approaches focus on patterns on a service interaction
level (e.g. Service Interaction Patterns),
seven approaches focus on human-to-human interaction (e.g. Patterns in
CoPE) without taking under consideration the services that may be involved
in a collaborative work,
only six approaches try to cover human and service interactions
simultaneously (e.g. e-Business Patterns) in an effort to establish a
complete supporting mechanism for facilitating both humans and services in
the context of collaboration.
Target User
to focus on the actual beneficiary of the pattern
We found that
the focus of the majority of the approaches is on the designer or the
facilitator of the collaboration (e.g. Patterns for Inter-Organisational
Business Processes),
only five of the pattern approaches that were examined, focus both on the
designer and the participant (e.g. CPats).
14. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Characteristics Used for the Review (4/5)
Contribution
in order to portray the vision on how patterns should be used.
We found that
The majority of efforts aim to support pattern mining as ‘guide of best
practices’ or as comparative elements between the design and the actual
collaborative work (e.g. MIT Process Handbook),
Only five approaches involve patterns that can actually be executed by
some kind of engine (e.g. Workflow Patterns, CPats).
Distribution
to capture the boundaries of collaboration.
most approaches cover interactions or collaborations that are located both
inside a group or an organization and can also span these limited
boundaries (e.g. Activity Patterns),
just five efforts are focused mainly on intra-organisational/group interactions
(e.g. Patterns in CoPE).
15. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Characteristics Used for the Review (5/5)
Knowledge
to position efforts based on their focus to acquire, distribute and
manage knowledge based aspects of collaboration (e.g. knowledge
flows, knowledge artefacts) with the use of patterns.
We found that
one third of the examined approaches (e.g. Cpats) is able to capture
knowledge related to collaborations,
one third can support aspects of knowledge based collaborations (but it was
not their main focus) ,
the remaining one third cannot integrate knowledge elements in their
approach (e.g. Service Interaction Patterns).
Event-Aware
for capturing whether the approach can be aware of the
collaboration state by means of event detection and processing
We found that
the majority of approaches do not provide an event-aware
mechanism,
only four of them take under consideration the change of states in
order to recommend or detect a pattern (e.g. CPats).
16. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Results (1/2)
We have detected two broad dimensions of work in existing
efforts.
The first dimension distinguishes between approaches that
detect/mine patterns in order to identify differentiations from
established best practices and propose corrective actions, mainly to the
designer/facilitator, and approaches that aim to directly assist
participants.
The second dimension distinguishes between approaches that require
manual intervention in order to be operable and ones that can provide
automatic support to participants.
17. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Results (2/2)
We argue on shifting the attention towards
directly assisting participants in automatic ways and
developing new tools that can proactively recommend
corrective actions in ongoing collaborations.
ontological structures of collaboration patterns that will:
allow the specification of sound and complete conceptual
models
enable the processing of rich social interactions and facilitate
additional functionalities, such as mining and recommendation of
appropriate collaboration tasks
to allow the full exploitation potential of the social semantic web
18. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Overview of presentation
Introduction
Survey of Patterns Approaches
Discussion
Collaboration Patterns Assistant
19. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
Collaboration Patterns (CPats)
For a solid and unambiguous description of Collaboration
Patterns, we present a model that captures the important
aspects of a CPat.
22. Information Management Unit / ICCS of NTUA www.imu.iccs.gr
CPA Screenshots
CPats the User
Selected to
Participate in
Events Related
to All the Active
CPats of a VO
According to the
Assignments by Main CPA Area for:
the CPat Initiator Collaboration (based on
Actionlist or Wf), CPat
Configuration, CPat
According to the Details
Collaboration State
Collaboration
(Pre-Conditions &
Triggers) Activities Pending for
a Specific Participant
per CPat