Participatory Design: Notes from a classic article by Christiane Floyd
et al
Randy Trigg
May 8, 1995
Floyd, C., Mehl, W.-M., Reisin, F.-M., Schmidt, G., &
Wolf, G. (1989). Out of Scandinavia: Alternative Approaches to Software
Design and System Development. Human-Computer Interaction, 4(4),
253-350
This is a wonderfully informative article, especially
in terms of the history and roots of the so-called "Scandinavian approach"
and several of its best-known projects.
Section 1 inquires as to whether there even is
such a thing as a Scandinavian approach:
We came to the conclusion that there is a Scandinavian Approach, but
that it is not necessarily a tangible phenomenon for the Scandinavians
themselves. The reason for this is that, in our opinion, it consists in
certain common features of the different schools, which we consider to
be intrinsic, but which are taken for granted in their own cultural setting
to the extent that they are not consciously recognized as such. Conversely,
the Scandinavians are able to put their finger on the differences between
the respective approaches, and in some cases these are the subjects of
controversy and discussion. (p. 258)Section 2 looks very briefly at Scandinavian history
and culture:Central European observers like ourselves are struck by the fundamental
will to establish a social consensus and to bring about a deliberate activation
of human resources with a view toward shaping a society that offers a high
quality of life to all its members and largely compensates social injustices,
that is, to create an egalitarian society. (p. 260, original emphasis)
The section proposes a variety of factors that contribute
to the particular stance toward technology and design found in Scandinavia.
These include:
strong trade unions and associations of employers,
and a high degree of cooperativeness between them;
importance of and investment in adult education in Denmark
(see also Susanne's PDC'94 paper);
the predominance of design as a skill and also
a feature of the social infrastructure:
Design aims at reconciling different, someimes conflicting concerns,
making optimal use of available resources in order to create high quality
products or systems at a reasonable cost. (p. 261-262)
various highly developed forms of cooperation: between
researchers in different fields (e.g. computer science and humanities or
organization theory); between researchers and industry around "pilot projects"
(p. 262); between researchers and trade unions; between trade unions and
employers; and finally, in the narrower context of system development,
between designers and users of technology.
a "fundamentally positive attitude" (p. 262) to science
and technology. In Scandinavia, the goal is, "to use technology to meaningfully
support social systems, while safeguarding human and social values." (p.
262)
and finally,
The Scandinavian Approach has been largely shaped by nonmilitary needs
[in contrast to the US] and is explicitly geared to the use of computers
for the benefit of human beings, in pursuance of the overriding goals of
humanization and democratization (p. 264)Section 3 discusses the roots of system design
approaches in Scandinavia. According to Figure 1 (p. 265) and the accompanying
discussion, the Scandinavian "school" is similar to socio-technical school
in that both "support human strengths", but differs in its aim to "contribute
to values and self-image of all affected" as opposed to "contribute to
control by and values of system owners". [We need to check this (and the
more in depth discussion in Section 4.1) against other historians of the
socio-technical approach.] According to the discussion on p. 267-268, the
sociotechnical approach led to a "harmony perspective" which was "sharply
criticized" by the trade unions. There's was a "conflict perspective" in
keeping with the understood sense of labor-management negotiations. The
result of the trade unions' collaboration with Kristin Nygaard & co
in Norway was what came to be known as the collective resource approach:[This] alternative methodology propounded by the trade-union-oriented
scientists was based essentially on consciously value-guided approaches
to design, elaborating and employing innovative concepts of systems development
and on the use of computer technology in organizations. Particular importance
is also attached to the idea of training schemes. These are designed to
familiarize workers with the new technologies, in order to safeguard and
to improve their qualification levels. (p. 268)
The subsection "Historical Shifts" lists several changes
in the methods used by Scandinavian system developers during this period.
Particularly important is the issue of user participation:
Cooperation between developers and users is considered a crucial factor
and is given methodological support. User participation was a demand already
recognized by the sociotechnical and other system-oriented approaches.
It was, nevertheless, later criticized as being inadequate, particularly
by adherents of the collective resource approach, because it failed to
provide sufficient codetermination rights for users and also ignored the
issue of user qualification for the participation process. Today, precedence
is given to mutual learning with guaranteed rights of codetermination,
also covering conflicts between different interest groups. Special training
schemes have been developed for this purpose. Various forms of prototyping
are used to provide technical support for the process of mutual learning,
to help progressively qualify users, and to adapt sofware to meet the needs
of specific user communities. (p. 269)
The subsection "The Systems Perspective - and Beyond" describes
the "systems perspective" whereby the computer is seen as part of a larger
system (including for example, humans and other "components"). The authors
show Nygaard's roots via Langefors in this perspective, but the ways that
Nygaard in his work around SIMULA moved beyond Langefors. Indeed,
The more common alternative in Scandinavia is to abandon the system
perspective altogether, and to use other concepts for describing the relationship
between human beings and computers. (p. 272)
The remainder of Section 3 discusses the "design view" (as
opposed to "production view") that characterizes system development in
Scandinavia. Features include a shunning of didactic methods and increased
autonomy for the developers,
In contrast [to the software engineering tradition], Scandinavian
approaches are based on software development as it really is (i.e., they
proceed from experiments on real program development processes and empirical
studies of real-life project situations). They take account of the problems
and tasks encountered here in their totality, and they view methods as
aids for the benefit of workers, enabling them to behave intelligently
in the development situation and use their skills to produce high quality
software. (p. 274)
The authors trace these ideas from Naur's early studies of
programming in the 70's and Mathiassen's doctoral thesis in 1982 and show
how they lead to three fundamental principles:
Mutual learning: Users and developers alike are reliant on a mutual
process of learning and communicating.
Designing by doing: Early experimentation and testing, such as using
prototyping and promoting communication and learning processes.
Interest-governed codetermination by the trade unions in the conception,
design, and control of computer-based work processes. (p. 275)
They go on to briefly describe prototyping as it is intended
to be used in Scandinavia and more generally the "process-oriented" approach
to system development.
Section 4 gives more background including a historical
line that goes from socio-technical to the collective resource approach,
and ends with capsule summaries of several of the historically important
projects.
The first subsection places the start of the socio-technical
approach with Tavistock just after WWII and then passes it to Norway in
the 60's, and from there to Sweden. In the 70's it was developed jointly
by Mumford et al and some Scandinavians at the Manchester Business School.
One of the key concepts is open systems, by which
the socio-technicals seem to mean systems that interact with an environment,
exchanging materials and maintaining a "steady state." Crucially, these
systems include both technical and social subsystems:
The central paradigm of the sociotechnical approach to work organization
consists of (a) the postulation of a causal relationship between job satisfaction
and work productivity and (b) the recognition that technology must be compatible
with organizational and social needs, if both of these are to be increased.
(p. 277)
Socio-tech recognized the importance of self-determination
and the problems of work alienation, but was less concerned with "the general
socioeconomic conditions affecting [work]." (p. 278)
The Norwegians took up socio-tech at the end of the 50's
as one means of enabling industrial democracy. The authors describe the
early course of adoption in Norway and the particular questions that were
taken up, for example, "How are the new technologies to be embedded in
the existing organizational structure so as to increase both work quality
and productivity, while at the same time ensuring greater job satisfaction?"
(p. 279) An assumption called the harmony thesis, the idea "that
each organization ... is governed by a common overriding goal or 'mission'."
(p. 280) can be said to characterize the work during this period.
Existing conflicts of goals between different groups or individuals
must be identified, and these can, by structural changes in the organization,
be settled in such a way that the stability of the sociotechnical system
and its effectiveness with respect to the common goal are restored. (p.
280)
The second subsection (4.2) introduces the collective
resource approach in part as a reaction (at least by trade unions in
Sweden) to the socio-technical strategy of focusing on the interests of
individual workers rather than developing collective solutions (p. 281).
In Norway, the Iron and Metalworkers union also was concerned over the
socio-technical approach's reluctance to confront the "need for changes
in social and economic policy" (p. 281) and the conflicts that must follow.
They speculated that the early participation experiments eventually lost
attraction for the workers because they missed, for example, the workers'
status in the company as a whole.
In the trade unions' view, a process of democratization in the interest
of the workers had to entail a change in existing power structures and
hierarchies (i.e., a change in the distribution of power, both at the shop-floor
level and in society as a whole). And they concluded that what was needed
was an interest-related strategy, taking into account the specific historical,
sociopolitical, and sociocultural conditions in Scandinavia, that would
enable the trade unions to play an independent part in the process of democratization.
(p. 281)
The goals of the collective resource approach included: redefining
the role of trade unions in planning and introducing new technologies and
work processes; "activating" trade union groups (committees?) at the company
level for safeguarding worker interests; encouraging and supporting union-initiated
R&D projects and "interest-governed, action-oriented" training schemes
aimed at increasing democratization (p. 282).
Three of the first major projects were NJMF (Norsk Jern
og MetallarbeiderForbund), DEMOS (Democratic Control and Planning in Working
Life: On computers, trade unions, and industrial democracy project) in
Sweden, and DUE (Demokrati, Udvikling og EDB) in Denmark. Though these
are characterized as "first generation" projects, at least DUE was formulated
in part as a reaction/follow-up to the lessons learned in NJMF (p. 284).
The "second generation" projects were oriented explicitly
toward involving union members in design of new technology (p. 286),
most famous among these was Utopia (see Section xx).
Given that the collective resource approach was distinctively
Scandinavian (unlike socio-technical) especially in its focus on industrial
democracy and the involvement of unions, Section 4 concludes with a discussion
of the particular theoretical assumptions underlying the approach. The
understanding of labor processes draws on Marxist conceptions of "use value"
and "division of labor". Into this is fit a "design process" that complements
"use process" as part of a total labor process (p. 287-288). The two processes
are essentially mutually dependent and constitutive, "The design of computer
support for labor processes must therefore be carried out with the users;
it cannot be performed either for or by them." (p. 288). This also leads
to a focus on training and independent union-directed R&D:
Clear identification of ... conflicts of interest... should not be
seen as inhibitions, but rather as prerequisites to cooperation on technological
and economic changes. An important prerequisite for trade union participation
in management design processes is a parallel and independent process of
accumulation of knowledge on the part of the unions. Independent trade
union research and development is therefore seen as a necessary complement
to traditional strategies of systems design. (p. 288)Section 5, System-Oriented Approaches, covers two
main methods of system design practiced in Scandinavia, ISAC and object-oriented.
The ISAC methodology was devised by Mats Langefors in the early 60's. One
distinction ISAC makes is between information systems and data systems
(the former involving "human knowledge"). The other distinction is between
user-oriented and computer-oriented activities, achieved by, "... calling
on system developers to confine themselves initially to user- and use-oriented
design decisions, postponing design decisions influenced by technical factors
until as late a stage as possible." (p. 290) The five basic work steps
of the methodology are: change analysis, activity studies, information
analysis, data system design, equipment adaptation. However, due to problems
with the change analysis step (too time consuming, imprecise selection
criteria, etc.), a new step is added at the start, enterprise diagnosis,
meant to uncover the goals at the enterprise level. Activity studies are
meant to limit the scope of the information system according to the actual
activities being carried out in the organization. (I could imagine an argument
for doing this in parallel with enterprise diagnosis.) Floyd et al explicitly
give their own appraisal of this approach:In our opinion, ISAC is a rather conventional approach, viewing, as
it does, the development of information systems from a result- or documentation-oriented
perspective. Less conventional, however, is the way the methodology is
applied in projects, in which a methodology is viewed from a wider perspective
that takes into account the communication and learning processes of the
people involved. (p. 294)
The authors trace the use of object-oriented system description
methodologies in Scandinavia back to the o-o programming language SIMULA
(their reference is to 1973). Nygaard's lectures at Aarhus University followed
in which he proposed combining the new language DELTA with a system description
method, "to create a tool for participative systems design that could be
used in various kinds of communication situations." (p. 294)
An essential aspect of this approach is that the authors make explicit
their conscious decision to view something as a system with respect to
a particular purpose. According to this view, systems do not exist as such;
rather, we choose to look upon certain subareas of reality as systems ...
This emphasizes the importance of the perspective of those involved (i.e.,
the way in which they mentally structure the development situation on the
basis of their experience, their expectations, or their personal values).
Consequently, systems do not exist in themselves, but are dependent on
an inquirer who seeks information about, or proposes decisions with respect
to, the considered system. (p. 295)
This work has continued more recently with the BETA programming
language, its programming environment, and research in system descriptions
associated with BETA.
This page is produced by CPSR members Randy
Trigg and Andrew Clement
[others?]. Comments, suggestions and pointers to non-commercial resources
are welcome.
Member login
Not a member yet?