Abstract
This paper describes a research effort to investigate
cross-cultural collaboration in software development.
The work is based on a model of collaborative design
that calls for stakeholders (including developers and
end-users) to iteratively construct an understanding
of design problems and potential solutions through
the construction and refinement of design
representations. The Evolving Artifact (EVA)
software design environment has been implemented
to support this process. EVA has been used in a
development project in a regional telephone
company. A case study of this project will be
analyzed to yield guidelines and design principles for
constructing representations for mutual
understanding.
Keywords:
collaborative design, workplace
cultures, work-oriented design, software
development, EVA, functional objects, prototyping
Introduction
My research views development of work-oriented [3]
software as a collaborative design task between
software developers and clients (i.e., end-users).
Developers and end-users belong to different
workplace cultures [2], each of which depend on
shared representations, norms, rules, and languages.
While shared conventions of meaning permit a
conceptual coordination within a given culture, they
also form boundaries that can hamper
communication across different cultures. The
essential challenges for cross-cultural collaborative
software development are to transcend the respective
cultural boundaries and to achieve a growing
understanding. The two main questions addressed by
my research are:
1) What kind of design representations can support
cross-cultural collaborative design of work-
oriented systems?
2) How can a growing understanding be
represented in a software artifact?
RELATED APPROACHES
Techniques and methods for large-scale, contract-
based software development are not oriented toward
user involvement [4], and therefore are not
appropriate to support the design of highly
interactive, work-oriented systems. These approaches
assume that analysts or other user representatives can
produce requirements documents that precisely and
completely reflect the needs of end-users. However,
this is an unreasonable assumption because end-users
do not always know their needs, and because their
task knowledge is often tacit [8]. Furthermore, end-
users are not always aware of the available
technological options [5].
Participatory design approaches aim to increase
usability of systems through collaboration between
developers and end-users [1]. My work shares this
goal. In particular, I focus on tool support
to ground collaboration with
computationaldesign representations, to drive an
evolutionary design process with the communication
that takes place around the representations, and to
capture the communication in an evolving artifact
along with the representations themselves.
MODELING COLLABORATIVE DESIGN
The action-reflection-critique model [6] is proposed
as a model for collaborative design. The model
synthesizes design theory with theories about how
people collaborate to solve problems. Design theory
suggests that designers solve problems incrementally
by creating explicit design representations which
"talk back" to the designer [9], revealing new
understandings of the problem. Human-human
collaborative problem solving theory suggests that
people construct and maintain an understanding
through dialog, in which meanings are accrued
incrementally, along with evidence of what has been
understood so far. The action-reflection-critique
model calls for stakeholders to incrementally develop
an understanding of design problems and potential
solutions through construction of external design
representations that ground communication of design
intent and design critiques. Mutual understanding is
constructed through creation, discussion and
modification of design representations.
REPRESENTATIONS FOR MUTUAL
UNDERSTANDING
Communication of design ideas through external
design representations is difficult when the
communication partners come from different
workplace cultures. Bødker and Pedersen [2]
describeworkplace cultures as systems of meanings
"hidden behind or in the various artifacts, symbols,
work routines, and established patterns of
cooperation." Importantly, meanings are not an
intrinsic property of artifacts and symbols. Instead,
they are ascribed by the people who share the culture.
Design representations must therefore be meaningful
across cultural boundaries to talk back to all
stakeholders and to foster a growing understanding of
the design issues.
System documentation is often not meaningful to
end-users because it is abstract and tends to be
written in the developer's language. Prototypes, on
the other hand, can be meaningful across cultural
boundaries because they are concrete design
representations that are interpretable by both
developers and end-users. When prototypes and
documentation are combined, the context for
understanding each is increased. Prototypes clarify
and instantiate documentation, while documentation
provides context for interpreting prototypes.
THE EVA ENVIRONMENT
The EVA (for Evolving Artifact) development
environment [7] has been implemented as a part of
this research to support the construction, integration,
and evolution of representations for mutual
understanding. EVA supports the creation of a novel
type of software artifact that integrates hypertext-
based documentation (including system
requirements, design rationale, usage scenarios, user
critiques) with prototypes. Hypertext is characterized
by mouse-sensitive links that allow navigation from
one piece of text to another. The EVA environment
provides an additional type of link that, when
selected, causes a prototype to appear on the screen
(in effect, navigating to the prototype). When the
user is finished interacting with the prototype, it
disappears from the screen and the user is once again
in the hypertext.
Using EVA, all stakeholders can create hypertext-
based documentation describing the application
domain and system requirements. Developers create
prototypes and embed them within the context of
appropriate documentation. End-users can then
access and interact with the prototypes to experience
what the final system might be like, and record their
comments and critiques in the hypertext.
Documentation and prototypes are then modified
accordingly and the iterative process continues. The
system design process is thus driven by
communication between developers and end-users,
and grounded by the evolving software artifact.
FUTURE WORK
EVA has been used to develop an application
prototype in a regional telephone company. This case
study will be analyzed to yield guidelines and design
principles for constructing design representations for
mutual understanding. Of particular interest are:
- What are the limitations and challenges of an
evolutionary approach to collaborative software
development?
- What tool support is necessary to help end-users
and developers to better deal with these
challenges?
References
1. Communications of the ACM. Special Issue on
Participatory Design, 36, 4 (June 1993).
2. K. Bødker, J. Pedersen. Workplace Cultures:
Looking at Artifacts, Symbols and Practices. In
J. Greenbaum, M. Kyng (Eds.), Design at Work:
Cooperative Design of Computer Systems,
Lawrence Erlbaum Associates, Hillsdale, NJ,
1991, Chap. 6, pp. 121-136.
3. P. Ehn. Work-Oriented Design of Computer
Artifacts. Almquist & Wiksell International,
Stockholm, Sweden, 1988.
4. J. Grudin. Interactive Systems: Bridging the
Gaps Between Developers and Users. Computer
24, 4 (April 1991), 59-69.
5. F. Kensing, A. Munk-Madsen. PD: Structure in
the Toolbox. Communications of the ACM 36, 4
(June 1993), 78-85.
6. K. Nakakoji, G. Fischer. Intertwining
Knowledge Delivery and Elicitation: A Process
Model for Human-Computer Collaboration in
Design. Knowledge-Based Systems Journal
(1995). (forthcoming).
7. J. Ostwald, B. Burns, A. Morch. The Evolving
Artifact Approach to System Building. Working
Notes of the AAAI 1992 Workshop on Design
Rationale Capture and Use, AAAI, San Jose,
CA, July, 1992, pp. 207-214.
8. M. Polanyi. The Tacit Dimension. Doubleday,
Garden City, NY, 1966.
9. D. Schoen. The Reflective Practitioner: How
Professionals Think in Action. Basic Books,
New York, 1983.
