Content uploaded by Juhani Iivari
Author content
All content in this area was uploaded by Juhani Iivari on Dec 21, 2020
Content may be subject to copyright.
Pergamon Information S.vstems Vol. 21, No. 7. pp. 551-575, 1996
Copyright Q 1996 Elsevier Science Ltd
PII: SO306-4379(96)00028-2 Printed in Great Britain. All rights reserved
0306-4379!96 $15.00 + 0.00
ANALYZING INFORMATION SYSTEMS DEVELOPMENT:
A COMPARISON AND ANALYSIS OF EIGHT IS DEVELOPMENT
APPR.OACHES+
JUHANI IIVARI’ and RUDY HIRSCHHEIM’
‘Department of Information Processing Science, University of Oulu. FIN-90571 Oulu, Finland
2College of Business Administration, University of Houston, Houston, Texas 77204-6283, U.S.A
(Received 10 September 1994; in,$nul revised from 9 Aqust 1996)
Abstract - This paper analyses two fundamental assumptions associated with the analysis and design of
information systems:: (1) the assumed organizational role of information systems, and (2) the view of information
requirements. In the case of the first assumption, it distinguishes three alternatives: a techmcal view, a
sociotechnical view, and a social view. In the case of the second assumption, again three alternatives are explored:
an objective view, a subjective view, and an intersubjective view. The paper points out the importance of these
assumptions from the viewpoint of IS development through the analysis of eight IS development approaches:
Information Modelling, Decision Support Systems, the Socio-Technical Approach, the Infological Approach, the
Interactionist approach, the Speech Act-based approach, Soft Systems Methodology and the Scandinavian Trade
Unionist approach. The first four are established traditions and the last four, newer and more emerging as IS de-
velopment approaches. The analysis shows that the first two established traditions have a technical-mechanistic
view of the organizational role of information systems, the view of the socio-technical tradition being socio-tech-
nicaI and the infological approach reflecting all three views. Most of the emerging approaches emphasize the social
nature of information systems. In the case of information requirements, the differences between the established and
emerging approaches are not as striking. While the objective and subjective views dominate the established
traditions, only the Speech Act-based approach and Soft Systems Methodology among the emerging approaches
seem to emphasize the intersubjective nature of information requirements. Copyright 0 1996 Elsevier Science Ltd
Key words: Information Systems, Systems Development, Systems Design, Systems Development Approaches,
Systems Development Methodologies, Information Requirements, Requirements Analysis
1. INTRODUCTION
Behind the approaches that are used to develop information systems (IS), lie a number of implicit and
explicit assumptions and views. Although the alternative assumptions and views guide the IS developer in
the choice of various analysis, design, and implementation options and hence have important
consequences for the development of successful systems, only rarely do they appear to be critically
reflected upon or challenged [ 11,451. The need to examine the underlying assumptions of IS development
(ISD)’ has begun to be recognized in the academic literature. Hirschheim and Klein [62], for example,
note the important role that assumptions play in developing information systems. The purpose of this
paper is to contribute to the debate on the underlying assumptions associated with ISD by looking at two
’ Recommended by Peri Loucopoulos
’ ISD refers to the analysis, design, technical implementation (construction), organizational implementation (institutionalization)
and subsequent evolution (enhancement, maintenance) ([32], p. 611) of information systems.
552 JUHANI IIVARI AND RUDY HIRSCHHEIM
sets of assumptions: one dealing with the perceived organizational role of IS; the other, fundamental
beliefs about what constitutes or determines information requirements. These two sets of assumptions are
explored within the context of eight ISD approaches. The choice of these approaches as well as the
assumptions are described in some detail later in the paper. We start by offering our definition of IS which
then guides our choice of assumptions to be explored.
An information system is interpreted here as a computer-supported system which provides a set of
people (users) with information on specified topics of interest in a certain organizational context (cf. [59]).
According to this interpretation, one can distinguish three major aspects in the modelling of information
systems: organizational context and users (host organization), topics of interest to the users (Universe of
Discourse), and computers (technology). In accordance with this view, the paper is based on a distinction
between three major levels of modelling or abstraction in IS development: the organizational level, the
language level, and the technical level [70, 951. Lyytinen [95], claims that these levels or “contexts” are
exhaustive, and argues their hierarchical order as follows: “Technology, or in general, the physical world,
is the basis for the language context, because language is always represented in some material carrier. On
the other hand language is necessary for any organized social action that comes into focus in the
organization context” (p. 11).
The significance of the technical level is axiomatic since information systems are defined as computer-
based. The language level emphasizes that an information system defines a formalized language to be
used to communicate about some Universe of Discourse (UoD). The related problems of information
requirements analysis (e.g. [126] and conceptual schema design [3] have been the subjects of active
research since the early 1970’s. The recent growth in Business Process Reengineering or Redesign [60]
has increased the interest in the organizational level in ISD. BPR implies that information systems are
highly organization sensitive and that IS development and organizational development are likely to be
very interwoven.
Despite the general agreement about the three levels, there is considerable variety in the way an
information system is conceived at these levels. An information system at the language level may be
viewed: (1) as a system that objectively describes the state of some part of reality - the Universe of
Discourse (e.g. [%I); (2) it may be seen as a system to satisfy user’s information requirements that may be
very specific to each user, his or her cognitive style and other personality traits (e.g. [78]); or (3) it may
been viewed as a formalization of language shared by the users (e.g. [54]). These three views are
respectively characterized as ‘objective’, ‘subjective’ and ‘intersubjective’ in the following’.
An information system at the organizational level may be considered a technical artifact or tool (e.g. [39])
just as a hammer, for example, which is socially (organizationally) produced and used, but which does not
embody any deeper social (organizational) meaning. Such a view we term “technical”. Alternatively, an
information system may be considered an artifact which entails inherent social and organizational aspects
(e.g. [108, 1301). This we characterize as a “social view”. An intermediate position between these two
extremes, termed the “socio-technical view”, emphasizes that an information system comprises both a
technical subsystem and a social subsystem that should be designed jointly [ 16]* .
The two sets of views are examined in more detail in sections 2 and 3. Section 2 distinguishes three
alternative views of the organizational role of an information system: a technical view, a sociotechnical
view, and a social view. Similarly, section 3 distinguishes three views of information requirements: an
objective view, a subjective view, and an intersubjective view. The ‘organizational role’ views discussed
in section 2 are related to the organizational level, while the ‘information requirements’ views articulated
in section 3 relate to the language level.
’ The distinction between the subjective and intersubjective views has some similarity with the ‘personal view’ and ‘common view’
of the agreement dimension in the requirements engineering (RE) framework of Pohl [113] which we will return to later in the
paper.
* The relevance of the organizational level is the most distinctive aspect of IS development approaches when compared with the
software engineering (SE) tradition reviewed by Pohl [113]. Even though there is a long history of the analysis of organizational
views or perspectives from the viewpoint of IS development (e.g. [84]), there is dearth of comparative research into the
organizational role of information systems in ISD approaches.
Analysing Information Systems Development 553
The paper applies the two sets of views to the analysis and comparison of eight ISD approaches. An “ISD
approach” in this paper is interpreted as a class of specific ISD methodologies or methods which share a
number of common features. This distinction will be elaborated in section 4.1’. Four of these approaches
are established traditions of ISD - Information Modelling (IM), Decision Support Systems (DSS), the
Sociotechnical Design Approach (STD), and the Scandinavian Infological Approach - and four are more
emerging - the Interactionist approach, the Speech Act (SA)-based approach, Soft Systems Methodology
(SSM), and the Scandinavian Trade Unionist approach. The research approach used in the selection and
analysis of these approaches is explained in section 4. The results of the analysis and comparison of the
approaches are described in section 5, while the paper as a whole, is summarized in section 6.
In addition to the analysis of the eight ISD approaches, showing their similarities and differences, the
paper argues the need for attention to the practical implications of the alternative views. The alternative
views, we suggest, guide the IS developer in the choice of concrete analysis, design, and implementation
options. The analysis also shows the considerable plurality of assumptions adopted collectively by the
eight approaches.
2. THE ORGANIZATIONAL ROLE OF AN INFORMATION SYSTEM
The organizational role of an information system is concerned with the relationship and function of an
IS within its organizational context. In the case of the organizational role of an information system, we
distinguish three views - technical, sociotechnical and social - on a continuum describing the relationship
between the information system and its host organization [73]*. A technical view regards an information
system predominantly as a technical artifact, and assumes that its connections with its organizational
environment can be reduced to well-defined inputs and outputs and ergonomic interface questions. The
social view considers an information system primarily as an organizational and social system; an
information system is seen as an integral, constitutive part of organizational communication, control,
coordination, cooperation and work arrangements and not only as a separate support system for these or-
ganizational activities. In the more theoretical terms of structuration theory [50], an information system as
a social system can be characterized as an embodiment of interpretive schemas, facilties of coordination
and organizational/social norms [108, 1281. The sociotechnical view is based on the assumption of
interdependent subsystems, the technical subsystem and the social subsystem which are designed jointly
(e.g. [61, 1281). It therefore is interpreted to form an intermediate value on the continuum technical vs.
social.
2.1 Theoretical Underpinnings of the Three Views
Figure 1.a analyzes the three views with regard to ontology, i.e. assumptions concerning so-
cial/organizational reality, on two continua: realism vs. nominalism [21], and voluntarism vs. determinism
in organizational and IS development (cf. human nature in [21]). The realist view implies that reality is
predictable and at least in principle manipulable, prescribable and designable, whereas a nominalist view
sees the social/organizational phenomena largely as emergent and cannot be directly designed (cf. 11271
for the distinction between prescribed and emergent organizations). In the case of voluntarism and
determinism of organizational and IS development. the latter emphasizes the importance of social inertia,
’ It may seem that these eight ISD approaches are incomparable. We claim, however, that the distinction between “approach” and
“methodology”obviates this problem of incomparability (see section 4.1).
* livari and Keroia [73] used the adjectives ‘mechanistic’ and ‘organic’ instead of ‘technical’ and ‘social’, adopting them from Burns
and Stalker [20]. This seemed to cause misunderstanding, even though the terms were used in a similar spirit in the sense that the
mechanistic view of the organizational role of an information system assumes that one can define a clear boundary between the
information system and the rest of the the organization, corresponding to clearly defined, deliniated and separated tasks of the
mechanistic organizational form [20]. In the case of the organic (social) view the boundary is much more fuzzy and a matter of
definition. The risk of misunderstanding is well illustrated by the fact that Bums and Stalker assume the orgamc organizational
form to be more innovative and adaptive than the mechanistic form. In our case of the organic/social vs. mechanist&technical view
of the organizational role of an information system, the conclusion is just the opposite because of social inertia.
554 JUHANI BVAFU AND RUDY HIRSCHHEIM
which may be related to both the prescribed and emergent aspects of organizations. The voluntarist view,
on the other hand, does not recognize or emphasize this kind of inertia. The diagonal, intersecting the first
and third quadrants of the figure, forms a continuum ‘design optimism vs. design pessimism’ as a synthesis
of the two dimensions. Design optimism assumes that only existing resource constraints limit the de-
velopment and acquisition of desired information systems, and that the systems can be implemented
without difficulty, if the quality of the system is good enough and people behave ‘rationally’ (cf. below).
Design pessimism, in contrast, assumes no such ease. People are assumed not to behave rationally and the
social nature of ISD makes any design exercise “wicked”.
In this basic framework we interpret the technical view to have a clear design optimist position,
characterized with a strong realistic ontology and a strong voluntarist view of ISD. The sociotechnical
view is more moderate, but nevertheless is characterized by design optimism. The social view is the most
complicated, since it intersects all the four quadrants of Figure 1.a. It especially emphasizes, however, the
emergent aspects of organizational phenomena and social inertia, when compared to the preceding two
alternative views. The emergent aspects appear as a result of social interaction.
Design
pesrnlsm Determinism I”
org. and IS
de”eblxne”+
Voluntorism In
org. and IS D&hl
Emergent Design&Pa
Figure 1 .o
Collective
orie4ltatlon
Design
Figure 1 .b
hdlvidud
orhtatbn
Fig. 1: Theoretical Underpinnings of the Three Views of the Organizational Role of an Information System
The extreme positions in the above interpretation are consistent with Kling and Scacchi’s [84],
distinction between the discrete-entity model and the web model and with Iacono and Kling’s [67]
distinction between the tool view and the institutional view; the former on the dimension ‘realism vs.
nominalism’ and the latter on the dimension ‘voluntarism vs. determinism’ in organizational and IS
development. Kling and Scacchi introduce four organizational perspectives: formal-rational, structural,
political, and interactionist, in their characterization of the discrete entity and web models (pp. 23-24). The
formal-rational perspective is most characteristic of the discrete-entity model and the interactionist
perspective to the web model [84]. This is consistent with Figure 1 .a where the discrete-entity model is
associated with the technical view and the web model with the social vied.
In the dichotomy of tool view vs. institutional view, Iacono and Kling [67] emphasize that tools are
assumed to be easy to change and replace and the personal freedom of their use, and in the case of the
institutional view, the social inertia and history dependence of the systems. This is also consistent with our
f The structural and political perspectives can be applied in framing both the discrete-entity and web models, even though the
application differs in these two cases. It is not possible to analyze the correspondence of these intermediate perspectives with our
framework in this paper.
Anaiysing Information Systems Development 55.5
interpretation of the technical and social views on the dimension ‘voluntarism vs. determinism’ in
organizational and IS development+.
We distinguish the sociotechnical view as an intermediate position between the technical and social
views, differing in this respect from the dichotomies of the discrete-entity model vs. web model and tool
view vs. institutional view. Iacono and Kling [67] seem to associate the sociotechnical approach more
with the institutional than the tool view (p. 102). We interpret in Figure l.a, however, that it is closer to
the technical view in its design optimism.
Finally, Figure 1.b simply adds the dimension ‘collective orientation’ vs. ‘individual orientation’ (cf.
[95]) to the above analysis. The technical view typically has an individual orientation whereas the social
view, a collective orientation. The sociotechnical view, as an intermediate position, has elements of both
the collective and individual orientation.
2.2 Practical Implications of the Three Views
We believe that the three views of the organizational role of an information system have important
repercussions on the ISD process, especially on IS implementation. Table 1 identifies a set of issues
related to ISD, with respect of which the three perspectives differ to a considerable degree.
Technical Sociotechnical Social
Priority of design
Causes of implementation
problems
Critical conditions foi
implementation success
Development strategy
Implementation strategy
Role of the change agent
Technical system
Poor technical quality
Human resistance
Technical quality
Analysis and design
Empirical-rational
Engineer
Technical and organizational/ Organizational/social
social systems equal partners system
Misfit between the technical Social inertia
and organizational/social
subsystems
Additionally, fit between the Additionally, social
technical and organizational/ desirability and fea-
social subsystems sibility of changes
Sociotechnical design of social Evolutionary deve-
and technical options lopment
Additionally, normative-edu- Additionally, power-
cative coercive
Facilitator Arbitrator
Table 1: Practical implications of the Three Views
First, the primary focus of the technical view lies in seeing an information system as a technical
artifact, leading to the primacy of the technical design of an IS and to the omission of any conscious
design at the organizational level. It may be recognized that the artifact to be developed may have
’ One should note, however, that Iacono and Kling’s discussion has influnced our analysis of the three views (technical, socio-
technical, and social), and consequently is not independent.
556 JUHANI IIVARI AND RUDY HIRSCHHEIM
different organizational/social implications. These issues are addressed, however, as afterthoughts. The
social view, in contrast, emphasizes that ISD is first of all concerned with the development of an
organizational system (its communication, control, coordination, cooperation, and work structures and
arrangements). ISD is primarily seen as organizational design and development, with technical design
concerning only the technical implementation of organizational changes. Our interpretation of the
sociotechnical school as an intermediate value on the continuum between the technical and social views
may become more understandable when it is viewed from the perspective of priority of IS design. The
sociotechnical approach suggests that an information system consists of mutually interdependent
organizational/social and technical subsystems, which should be treated as “equal partners” in IS design
without any priority of action [ 1021.
Second, in the case of organizational implementation problems, the views lead to different emphases.
The technical view sees implementation problems as primarily attributed to poor technical quality of the
system or to human resistance (cf. system-oriented and people-oriented theories of resistance in [97]). The
resistance may be considered ‘irrational’ behavior, or in a moderate form, understandable, rational
behaviour from the user’s own subjective viewpoint. The sociotechnical view emphasizes the misfit
between the technical and social/organizational subsystems as the primary reason for implementation
problems (cf. interaction theories of resistance in [97]), while the social view adds social (institutional)
inertia to the list of implementation problems. As a consequence, the technical view emphasizes the
technical quality of the system; the sociotechnical, the fit between the two subsystems; and the social
view, the desirability and feasibility of changes (cf. [29]) as critical prerequisite criteria for
implementation.
In the case of systems development strategies, the technical view relies on traditional analysis and
design approaches or prototyping, whereas the sociotechnical view has its own design approach, including
the design and merging of social and technical options [ 1021. The social view sees that ISD is constrained
by historical commitments [67], and consequently emphasizes the evolutionary nature of ISD’. In the case
of organizational implementation strategies, the technical view primarily adheres to empirical-rational
strategies [28], whereas the sociotechnical view sees an additional need for normative-educative ones,
while the social view adds the requirement of power-coercive strategies if social inertia is to be overcome.
As a consequence of the above, the role of a developer as change agent can be characterized in terms of an
engineer (scientist), facilitator (teacher) and arbitrator (cf. [122]).
3. THE VIEWS OF INFORMATION REQUIREMENTS
The view of information requirements is concerned with the basic notion of what constitutes or
determines an information requirement [66]. We distinguish three views: the objective, subjective and
inter-subjective views of users’ information requirements. The objective view emphasizes the importance
of impersonal features such as the organizational position and task of the user as a determinant of his/her
information requirements or the objective existence of the slice of reality to be modelled by the system,
while the subjective view stresses that the information requirements are first and foremost determined by
the personal characteristics of the user (his frame of reference, cognitive styles, etc.). The view of
information requirements as the need to attain intersubjectivity among the IS users is a more recent idea in
the IS community [ 15, 53, 541. This view differs from the preceding two in several respects as will
become clear in the following analysis of the theoretical underpinnings and practical implications of the
three views.
The three views of information requirements have some similarity with the agreement dimension in the
three-dimensional framework for Requirements Engineering (RE) proposed by Pohl [113]. The first
dimension, “specification”, describes the degree of understanding of requirements ranging from opaque to
complete, the second dimension, “representation”, deals with the the different representation forms
ranging from informal to formal, and the third dimension, “agreement”, describes the degree of agreement
reached on the specification, ranging from a personal view to a common view. He seems to interpret the
dimension more as agreement between different stakeholders (systems analysts, manager, user, etc.) while
’ We make a clear distinction between prototyping and evolutionary approaches. The former is based on an experimental use of
the prototype, whereas the latter is based on the use of the operational information system in real circumstances (cf. [68]).
Analysing Information Systems Development 55’7
we view intersubjectivity solely within the IS user community. Pohl sees the RF process as a trace in the
3-dimensional space towards more complete specification and increasing agreement using more formal
representation forms. He sees the personal views of requirements as an initial state, the goal of the RF
process being to increase common system specification. Even though we largely agree with this, we do not
wish to exclude subjective (personal) views and differences in the desired output of the RE process.
Pohl [ 1131, asseses that there is not much earlier research done on the agreement dimension in the REI
area. In this sense the present paper complements Pohl’s analysis, focusing on ISD approaches instead of
software engineering (SE) methods. The analysis also indicates interesting variations in the RE process
when compared with Pohl’s view. The analysis of alternative ISD approaches shows that the view of RE as
a trace towards common agreement may not generally be valid or may be misleading, because there arc
some ISD approaches that have emphasized personality differences as essential determinants ol-
information requirements.
3. I Theoretical Underpinnings of the Three Views
The following analysis is based on a very similar framework as the one used in the analysis of the three
views of the organizational role of an information system. Figure 2.a depicts the three views 01
information requirements with regard to ontology, realism vs. nominalism, and voluntarism vs.
determinism of organizational behaviour and in particular IS use. Similar to Figure la, realism is
associated with permanent structures and nominalism with emergent processes’. The diagonal intersecting
the second and fourth quadrants forms a dimension ‘functionalism vs. interpretivism’ [21]. Functionalism
assumes that organizational behaviour and IS use are determined by organizational structure, which is
associated with a realist ontology. Interpretivism, on the other hand, sees that organizational ‘reality’ is
created by organizational (inter)action. The objective view of information requirements has a cleat
functionalist position; it assumes that the organizational structure (the position and tasks of a user) defines
his or her information requirements, including his or her conception of the Universe of Discourse. The
intersubjective view on the other hand has a nominalist ontology and emphasizes voluntarism in
organizational behaviour and information requirements. Information systems are seen as integral parts of
organizational sense making [15], and ISD as the development of organizational communication and the
formalization of the professional language of the user community [54]. Information requirements are
largely a matter of social agreement. We suggest that intersubjectivity leads to seeing information
requirements primarily as emergent *. The subjective view lies between the two extremes, functionalism
and interpretivism. The former in the context of the subjective view corresponds to the conception of
information requirements in terms of different, “measurable” personal characteristics of users (e.g.
cognitive styles), whereas the interpretivist side emphasizes the uniqueness and freedom related to each
user’s requirement: his/her requirements are largely his/her personal choice and interpretation, depending
on how he/she prefers to view his/her organizational role, tasks, the universe of discourse, and so on.
Figure 2.b adds the dimension of collective-orientation vs. individual-orientation to the above analysis. It
points out that the objective and intersubjective views have a clear collective orientation, whereas the
subjective view is individually oriented.
’ The distinction between permanent structures and emergent processes is inspired by Truex and Klein’s [ 1291 distinction between
a priori structures and emergent processes. They interpret a priori structures as objectivied, relatively stable structures, whereas
emergent processes may be seen at most temporal agreements in the continuous process of negotiation and renegotiation.
* Goguen [51, 521, also emphasizes the emergent nature of requirements but sees them to emerge from the interactions between the
requirements engineers and the client organization rather than from the continued social interaction of the users in the client
organization
5.58 JUHANI IIVARI AND RUDY HIRSCHHEIM
Nominalism Realism COlktW
orientation *
Figure 2.0 Figure 2.b
lntitldual
orientation
Fig. 2: Theoretical Underpinnings of the Three Views of Information Requirements
3.2 Practical Implications of the Three Views
The three views have important practical implications from the viewpoint of IS design. In the case of
the objective interpretation, information requirements analysis can, at least in principle, be conducted as
an impersonal activity or reality analysis; the subjective view presupposes focusing on each user’s
personal characteristics and requirements, whereas the intersubjective view considers requirements
determination as a social rule analysis and reconstruction [54]. They also differ with regard to the question
of user participation. Putting aside the ethical, motivational, and committal arguments for user
participation, the objective view presupposes user participation at most in the role of an application
domain expert: the user may be required to explain the intricacies of the tasks supported by the system.
User participation is beneficial as far as the representatives know the intricacies of work to be supported
by the information system. In the case of the subjective view, the user may be an object of different perso-
nality tests, the results of which are believed to help him/her and the analysts to understand his/her re-
quirements, or he/she may be an individual decision maker, whose cognitive models and preferences
define his/her information requirements as were outlined above. Finally, the intersubjective view sees
users as integral actors in organizational communication whose participation is by definition a necessity in
order to achieve intersubjectivity ([54] p. 26).
Obiective Subjective Intersubiective
Asking (x) X (Xl
Normative analysis X
Strategy set transformation X
Critical factors analysis (x) X
Process analysis X
Decision analysis X
Input-process-output analysis X
Table 2: Underlying Assumptions of Strategies of Information Requirements Determination
Analysing Information Systems Development 5%
Clearly different strategies of information requirements determination differ in their assumptions
concerning the views of information requirements, or that they, in practice, are not neutral with regard to
the different views. Table 2 analyzes some of the well-known strategies proposed by Davis [31] from this
perspective. Davis’ first strategy “asking” is, of course, a very general approach, which may be used to
elucidate all types of requirements. It is clear, however, that it tends to emphasize either the subjective
views of the analyst in the case of closed questions, or of the user in the case of open questions. Davis [3 11
also includes group-oriented methods (brainstorming, guided brainstorming and group consensus) within
the category of asking ([31] p. 13). It is still an open question to what extent these group methods are able
support the elucidation of intersubjective aspects of requirements. In the case of the remaining strategies
our conclusion is that they primarily reflect either the objective view or the subjective view. The
interesting aspect of Table 2 is the finding that except asking, none of the strategies directly reflects or
supports an intersubjective view+.
More recently, ethnography and ethnomethodology inspired methods of requirements analysis (e.g.
[ 13, 14, 19, 93, 115, 1241) partially address the problem of intersubjectivity. They see the users’ work as
socially organized and look closely at how the users actually organize their work, and the categories and
methods they use to render their actions intelligible to one another [5 1, 521. Despite this orientation, they
do not see increased intersubjectivity among the users as the goal of requirements analysis. Apparently,
they see shared understanding among users as unproblematic and focus more on the mutual learning
between the designer and customer rather than on increasing intersubjectivity among users. Their focus
also lies more in the detailed understanding of users’ work than users’ information requirements and
shared understanding of the Universe of Discourse. These methods are, however, in their infancy as
methods of requirements analysis (see [115] for a discussion of their difficulties, and [2] for possible
misconceptions).
Even more formally oriented methods of requirements analysis do not explicitly recognize in-
tersubjectivity among users as the goal of the analysis, although they indirectly provide instrumental
support for this view. “View integration” in the conceptual IM tradition (e.g. [ 121) includes some notions
of intersubjectivity, but it is usually not associated with the nominalist (constructivist) ontology underlying
the intersubjective view (see also [86]). Recent viewpoint-oriented requirements engineering methods
(e.g. [85, 92, 103, 1051) interpret “viewpoints” more generally as perspectives of the artifact to be
developed, adopted by different agents (participants or actors) without restricting “viewpoints” to “user
views”. They do not specifically recognize intersubjectivity defined over the user population of the system
as the goal of requirements engineering but rather only sufficient consistency [47]. The work on
integration of viewpoints [46, 961 and conflict resolution [38] may, nevertheless, be useful when
promoting the intersubjectivity of information requirements. Recent work on design rationales. when ap-
plied to information requirements (cf. [ 114]), may also be instrumental in advancing it.
4. THE RESEARCH APPROACH
The purpose of this section is to introduce the ISD approaches selected for review, and the method
used in the analysis.
4.1 The Approuches Selected fur the Review
The field of ISD has grown dramatically over the past three decades. In addition to the expansion in
the number and complexity of systems developed, the number of ISD methodologies has been growing
leading to an increasingly dense “methodology jungle” [8]. To our knowledge there has been little
systematic analysis of the major approaches of ISD (but see [64]) and still less any generally accepted
classification for them, even though the topic has been discussed in a number of articles 162, 7 I, 72, 104,
1341. Banville and Landy [IO], for example, characterize MIS (or IS research) as a fragmented adhocracy,
in which research is rather personal and weakly coordinated in the field as a whole.
’ They may, of course, support intersubjectivity indirectly by increasing consensus on organizational goals, processes, critical
factors. etc
560 JUHANI IIVARI AND RUDY HIRSCHHEIM
Part of the confusion stems from the notion of ‘methodology’ and ‘approach’. In this paper, we
explore these notions concluding that there may be value in moving beyond ISD methodologies to focus
on more general ISD approaches. We define LSD approach as a class of specific ISD methodologies
which share a set of common features. An LSD methodology or method refers to a codified set of goal-
oriented “procedures” which are intended to guide the work and cooperation of the various parties
(stakeholders) involved in the development of an IS application+. Typically, these procedures are
supported by a set of preferred techniques and tools, and guiding principles [65]. An ISD approach may
include zero, one, or more concrete methodologies as its instances. As an example of the first case (i.e.
zero methodology instances), consider the Interactionist approach (see below) which to our knowledge
has not been developed into any specific methodology, whereas the IM approach (see below) has
numerous methodology instances.
Our claim is that the concept of an approach makes it meaningful to compare various ISD approaches
which may be in quite different stages of their development in terms of the number of their methodology
instances. The concept ‘approach’ also leads us to focus on more general principles underlying specific
methodologies [69] manifesting them as similarities and differences. It is possible to analogize our
categorization of approach and methodology to Brooks’ [ 181 famous distinction between essences and
accidents of software products. An ISD approach focuses on the essences of its methodology instances
while concrete methodologies necessarily include also accidental features (e.g. related documentation
notations), which while important in practice, do not define its essence.
We selected four traditional ISD approaches - Information Modelling (IM), Decision Support Systems
(DSS), the Socio-Technical Design approach (STD) and the Scandinavian Infological approach - and four
more recent ISD approaches - the Interactionist approach, the SA-based approach, Soft Systems
Methodology and the Scandinavian Trade Unionist approach.
Approach Journals/conferences Textbooks Founders
Information Modelling X X
Decision Support Systems X X
Sociotechnical approach X Mumford
Infological approach X Langefors
Interactionist approach Kling
Speech Act-based approach Flores, Goldkuhl
Lyytinen, Winograd
Soft Systems Methodology X Checkland
Trade unionist approach Nygaard, Sandberg,
Ehn, Kyng
Table 3: The IS Development Approaches Selected for the Analysis
In addition to the distinction between the established and emerging approaches, the eight approaches
are selected to represent the variety of positions taken with regard to the assumptions. The selection of the
traditions is primarily guided by their institutionalization in the scientific community and secondarily on
the existence of identifiable founders and followers. Referring to Kuhn’s discussion about scientific
’ We use the tern “methdodology” and “method” interchangeably; however we note that the latter term is becoming increasingly
preferred, especially in Europe.
Anaiysing Information Systems Development 561
communities in his postscript to the second edition of his book ([87] pp. 176-178), the assessment of
institutionalization is based on three subcriteria: the existence of scientific journals, scientific conferences,
and textbooks’. The approaches are listed in Table 3.
The IM approach has its roots in the data modelling tradition of the database field. The term
‘Information Modelling’ is preferred here to point its application to ISD. It has been very influential both
in Europe and North America. In Europe, it has especially dominated research within IFIP (International
Federation for Information Processing) TC8 (Information Systems) Working Group 8.1 (e.g. [109-l 1 I]).
It has also heavily influenced specific methods such as Information Engineering [98-loo]. The DSS
approach has been visible for quite some time and started to emerge in the context of several experimental
projects in the early 1970’s [78]. The STD approach and the infological approach are more European
traditions. The former has its roots in the Tavistock Institute in the late 1940’s. Its application to ISD was
pioneered by Enid Mumford [ 1011, in the U.K. starting from the late 1960’s. Subsequently, it stimulated
interest in North America as well [16, 1121. The Infological approach initiated by Langefors [90] in the
mid 1960’s has been influential especially in Scandinavia.
The four more recent approaches - the Interactionist approach, the Speech Act (SA)-based approach,
Soft Systems Methodology (SSM), and the Scandinavian Trade Unionist approach - are called emerging,
since they do not have an institutional status comparable to the four traditions introduced above and their
relevance to ISD may not be fully recognized. The Interactionist approach refers to the body of research
conducted at the University of California, Irvine, by professor Rob Kling as his colleagues. The SA-based
approach to ISD borrows its basic concepts from the linguistic philosophy of Searle [119-1211. It was
pioneered by Fernando Flores and Terry Winograd (e.g. [49, 1321) in North America and Goldkuhl and
Lyytinen [.54,55] in Europe. SSM is a general systems approach developed by professor Peter Checkland
and his colleagues at the University of Lancaster [29, 30, 1311. Although SSM was originally a general
systems approach, without any specific orientation towards information systems, its developers are
increasingly perceiving it to be well suited to ISD. Within the IS community SSM has been incorporated
as an integral part in such approaches as MULTIVIEW [9] and FAOR [118]. The Trade Unionist
approach is a tradition that has evolved mainly in Scandinavia as a trade union response to the challenge
of Scandinavian co-determination arrangements and laws enacted in the mid-1970’s that ensured the
employees and unions the right to participate in the design of and decision-making concerning computer
systems [40]. This approach was initially based on a strong ‘class politics’ perspective to organizatiom
[84]. This Marxist ideology has, however, been weakening recently and the approach is in a transition
towards “cooperative design” [57]; it has raised considerable interest lately [88].
The eight approaches are by no means assumed to be exhaustive and/or mutually exclusive. Nor is that
our goal. Exhaustiveness is not considered any absolute goal since it is a moving target. Rather a careful
and convincing analysis of the selected approaches is our preference, and in order to keep this paper
within reasonable limits, the analysis is confined to these eight approaches. Moreover, the approaches are
not necessarily mutually exclusive because they have influenced each other to
extent during their evolution. some greater or lesser
4.2 The Method of Literature Analysis
Given that the primary purpose of this paper is to analyze two assumptions underlying the concept
‘information system’ in the eight approaches introduced above, we must acknowledge that this kind of
analysis faces many difficulties. First, as we noted above the approaches are not necessarily clearly
distinct and they have continued to evolve over time. One way to deal with this is to try to evaluate these
approaches as intellectual traditions in their original forms. We prefer, however, to view them in their
contemporary form to make the analysis more pertinent to current research. Second, we must take into
account that there may be considerable variation within each approach. One solution to this, applied by
[ 1041, is to treat them as “ideal types” in the Weberian sense. Ideal types have the benefit that they make it
’ Kuhn also mentions professional societies, informal and formal communication networks and citations as means of identifying
community structures. Analysis of the latter two is beyond the scope of this paper. Professional societies can be expected to
coincide largely with scientific journals and conferences
562 JUHANI I!VAR~ AND RUDY HIRSCHHEIM
possible to find clearer differences between the approaches, but they are normally purely theoretical con-
structions that do not exist in reality. Therefore, the analysis could remain at a rather general and abstract
level without any clear addressable target’.
In this paper we have selected a different approach. The idea is based on Kuhn [87]. He emphasizes
throughout his book the importance of textbooks as manifestations of existing paradigms and in the
dissemination of paradigms to the next generation. “They are the bases for a new tradition of normal
science” (p. 144) and “expound the body of accepted theory” (p. 10). Applying his suggestion, we have
chosen one or two books, preferably textbooks, to represent each established approach. The selection of
books is based on two main criteria. Firstly, a book should be a good representative of the tradition from
the viewpoint of ISD. Secondly, the goal has been to select books by leading authors who have effectively
influenced the development of the tradition they represent. The books representing IM [98-100, 1091
represent European and North American views. The book of Keen and Scott Morton [78], is a classic in
the DSS field, and the textbook of [ 1231 one of the best-known on the topic, with particular emphasis on
how to develop a DSS. The two books on the sociotechnical approach [ 102, 1121 represent European and
American views. Mumford’s [102], book summarizes the ETHICS methodology, and the book of Pava
El 121,
represents an important extension of the sociotechnical approach to nonroutine office work.
Finally, the two books [94, 1251 on the infological approach cover comprehensively its major
contributions. The book of Lundeberg et al. [94], introduces the ISAC methodology that largely
summarizes the Scandinavian infological tradition. The book of Sundgren [ 1251, represents his formalism
for conceptuaYinfologica1 modelling*.
To counter any potential criticism that may be attracted to the use of textbooks, we wish to direct
attention to the following four points. First, the selected books are highly research-oriented in the sense
that they include numerous summarizing references to existing research or are heavily based on the
authors’ earlier research. Taking into account that they are written by leading researchers in the field, one
can expect that the results of our analysis reflect the mainstream research in each tradition. Second, the
use of books should be assessed taking into account the alternative of using original research contributions
representing each tradition. In view of the breadth of some traditions, it is clear that the review of original
research contributions could not be exhaustive. When one furthermore recognizes the problems of
drawing strict borderlines for the traditions and internal variation within them, it becomes obvious that in
choosing an “appropriate” selection of the articles to be reviewed, one may reach almost any conclusion in
the case of some traditions. The important point in the use of textbooks is that the authors of the books
have made their selection of “the body of accepted theory” independently of the purposes of this paper.
Third, applying Argyris and Schon’s [4] distinction between “theory-in-use” and “espoused theory”, one
can argue that textbooks make it possible to analyze the “theories-in-use” better than by adhering to
original research papers. Fourth, the use of textbooks assures that the results are highly relevant from the
viewpoint of IS education since they are mediated to practice through the education of future IS
professionals.
In the case of the emerging approaches, where it is too early for textbooks, we had to rely on a more
comprehensive analysis of the existing literature. The analysis of the Interactionist approach is based on
the following articles: [67, 80-841. Although the SA-based approach has aroused considerable interest in
the IS community recently (e.g. [5-7, 33-37, 48, 49, 54, 55, 74, 76, 79, 91, 1321 our analysis is based
mainly on the works of Auram&i, Flores, Goldkuhl, Lehtinen, Lyytinen and Winograd [5,48,49, 54, 91,
95,132]. SSM has been published in a number of books [29,30, 1311. Within the IS community it has also
aroused considerable interest, being integral parts of such approaches as MULTIVIEW [9] and FAOR
[ 1181. In order to keep our discussion as close to what Checkland intended, the analysis of SSM is
primarily based on the original texts on SSM emphasizing the most recent ones. The analysis of the Trade
Unionist approach is based on the following material: [23-27, 39-44, 89, 106, 116, 1171
’ A second risk is that the ideal types may turn out to be too extreme, which may lead to unnecessary controversy and
misunderstandings.
* The major neglect of ISAC as a representative of the Scandinavian infological tradition concerns conceptual/infological
modelling of the Universe of Discourse (UoD). Sundgren [125], complements Lundeberg et al. [94], in this respect.
Analysing Information Systems Development
5. RESULTS OF THE ANALYSIS
5.1 The views on the organizational role of an information system
The assumptions of the eight ISD approaches, concerning the organizational role of information
systems, are summarized in Table 4.
Information Modelling: Martin [99], defines an information system as “a system of data and processes
that can be used to record and maintain information” (p. 467). An information system consists of one or
more applications. An application system refers to “the automated and related manual procedures
supporting a set of business processes” (p. 448). Even though ISD in information engineering (IE) is very
heavily based on the analysis of the organization to be supported, the view of the organizational role of an
information system is technical. Information Strategy Planning (ISP) results in a number of business areas
based on the functional and entity modelling of the organization. The applications to be developed are
identified in the Business Area Analysis (BAA) based on more detailed data modelling and process
modelling. Martin [99], does not describe exactly their identification, but proposes JRP (Joint
Requirements Planning), JAD (Joint Application Design), and prototyping as major means of the further
specification and design of individual applications [IOO]. The procedure for JAD, suggests that in-
formation systems are specified as input data, output data and related processing in the detailed procedure
steps of the processes identified in BAA.
Technical Sociotechnical Social
Information Modelling
Decision Support Systems
Sociotechnical approach
Infological approach
Interactionist approach
Speech Act-based approach
Soft Systems Methodology
Trade unionist approach
*** *
*** *
***
** ** **
***
***
** **
*** *
*** Strong orientation ** Some orientation * Weak orientation
Table 4: The Views of the Organizational Role of an Information System
In terms of Kling and Scacchi [84], the view of information systems in IE seems to reflect the discrete
entity perspective that was associated with the technical view. One should nevertheless note that IE
strongly emphasizes the interaction between the organization and information technology. While the
identification of information systems in IE is heavily based on organizational analysis and modelling,
Martin [99], underlines in the context of ISP that the purpose is not only to analyze the enterprise and to
identify information needs leading to the computerization of the existing procedures, but it should identify
the fundamental activities and the fundamental information needs likely leading to rethinking the
enterprise structure (p. 20-21). He points out that more advanced (information) technology justifies a
different corporate structure. “A computerized corporation ought to have different procedures from a
corporation with manual paperwork. A corporation with workstations on all knowledge workers’ desks,,
IS 21/7 B
564 JUHANI JIVARJ AND RUDY HIRSCHHEIM
connected by a corporatewide network to databases, ought to have different procedures from a corporation
with batch processing” (p. 62).
The framework for information system development advanced by Olle et al. 11091, resembles in many
respects IE. They characterize an information system as: “a means of recording and communicating
information to satisfy the requirements of all users”, engaged in a business activity “by providing the
information it needs or by automating some or all of it” (p. 229). Olle et al. [109], assume information
systems to be identified in the Information Systems Planning stage as change alternatives that are new
ways of doing things in an information system or a business activity in order to deal with some business
problems (p. 222)+.
Decision Support Systems: In the case of the DSS tradition, our conclusion is that it primarily reflects a
technical view, but includes aspects of a social view. Keen and Scott Morton [78], interpret a DSS in
rather technical terms as: “a conversational, interactive computer system with access through some form
of terminal to the analytical power, models, and data base in the machine” (p. 58). They also emphasize
the role of a DSS as a support tool under the control of managers (p. 2). On the other hand, they
emphasize the predominance of the organizational, political, and behavioural context of DSSs (pp. 50, 7 1,
171-172, 189), having a relatively rich discussion on these contextual factors in their book (pp. 61-73).
They also recognize the implementation problems of DSSs (pp. 71, 175176), devoting a chapter to DSS
implementation (pp. 189-211). They remark that “the failure of a system or a model is now rarely due to
technical inadequacies” (p. 50), and mention “political factors that may make these technically excellent
products organizationally infeasible” (p. 71). Despite these remarks, they regard a DSS as a discrete entity
that may be embedded in a rich organizational environment: “the ‘system’ is not an artifact but a set of
computer routines that must be meshed into the organizational, political and behavioural context” (p.
189). To summarize we interpret that the view of the organizational role of a DSS in Keen and Scott
Morton [78], is primarily technical, but includes social elements. The view cannot be regarded as
sociotechnical, however, since Keen and Scott Morton [78] do not see any clear need for the
organizational design of DSSs. Organizational perspectives referred to above are only mentioned in the
“entry” of the DSS design process (p. 174)‘. The book of Sprague and Carlson [123], does not essentially
change our conclusion. Even though they emphasize the organizational nature of DSSs more than Keen
and Scott Morton [78] do (p. 26), they in fact pay less attention to the social aspects of DSSs.
The Socio-Technical Design approach: Not surprisingly, the view of the organizational role of an
information system in the case of the STD approach is sociotechnical. Both books [102, 1121 include
explicit design of technical and social options for an information system.
The Znfological approach: The view of the organizational role of an information system in the
Infological approach comprises technical, sociotechnical and social aspects. ISD in ISAC [94] is
intimately related to “change analysis” which is the first phase of ISAC. Change alternatives identified and
selected in Change Analysis are assumed to normally combine different development measures such as
ISD, organizational development including individual development, and development of the direct
business activities (p. 118-l 19). Change analysis clearly places priority on the organizational/social design
of the system. However, Lundberg et al. [94], do not pay any particular attention to the social meanings
of information systems. Even though Lundeberg et al. [94] do not directly propose joint development of
technical and social options, they emphasize that the change alternatives should be subjected to both
social and economic evaluation. This implies some sociotechnical thinking. Information (sub)systems are
identified as a result of detailed “activity analysis”. Activity analysis is continued until identified informa-
tion processing activities can be classified as unformalizable, formalizable but unautomatable, and
automatable (pp. 148-149). If we exclude unformalizable subinformation systems, we can conclude that
information subsystems simply formalize and computerize the identified organizational activities. In this
sense information (sub)systems are seen as separate technical artifacts.
The Znteractionist approach: The view of the organizational role of an information system in the
Interactionist approach is clearly social. As discussed in section 2, Kling et al. propose the web model as a
’ Cf. the Infological approach below
’ One explanation for the exclusion of organizational design is the individual orientation of their book [77]. A second explanation
might be that Keen and Scott Morton [78], see the social aspects largely as emergent and consequently not designable. It is hard to
find any direct support for this second explanation from their book, but their emphasis on the evolutionary design is consistent with
it.
Analysing Information Systems Development 565
contrast to the discrete entity model of an information system [80, 83, 841 and later the institutional view
as a contrast to the tool view [67, 821. According to Kling and Scacchi [84], “the basic unit of analysis of
the discrete-entity model is a computing resource” (p. 9): “A computing resource is best conceptualized as
a particular piece of equipment, application, or technique which provides specifiable information
processing capabilities” (p. 6). They continue that “computer-based technologies are tools, and are
socially neutral” (p. 6), the use of which is isolated “from the actual work practices and organization of
labor within which automated data systems are typically developed and used” (p. 3). Web models on the
contrary “make explicit the salient connections between a focal technology and its social and political
contexts” (p. 3). In addition to their information processing capabilities, “computer-based technologies are
also social objects which may be highly charged with meaning” (p. 7-8). Kling and Iacono have proposed
the distinction between the tool and institutional views in their recent papers [67, 821. They suggest that
when “analysts emphasize the information-processing capabilities of a computer-based technology, they
are foregrounding its “toolness” or instrumental value for particular social units”, whereas when “analysts
emphasize the social and political choices that organizational actors have made over time, they are fore-
grounding its institutional character” (p. 104). They contend that computer-based information systems
(CBIS), developed from complex, interdependent social and technical choices are better conceptualized as
institutions than as tools, because “the image of a CBIS as a tool is associated with tremendous personal
freedom”, “there is an underlying assumption that computer-based technologies have no inherent politics:
they are consistent with any social order” and “attention is focused on a future of technological
perfection”, whereas “institutional analyses emphasize the social use of CBIS and social control over
computing arrangements”, “politics play an important role” and the focus lies on “the developmental
trajectories of CBIS. Institutions develop a character based on the interest they have served in the past,
their organizing ideologies, and the world views which bind their participants together” (pp. 104-105).
Further, the institutional view emphasizes the social ‘inertia’ related to computer-based information
systems ([82] p. 11).
The Speech Act-based approach: The view of the SA-based approach is clearly social. Goldkuhl and
Lyytinen [54], suggest the notion of information systems as “social systems only technically implemented”
([7] p. 127). More specifically, Flores and Ludlow [49], claim that organizations are essentially
institutional arrangements to create, take care of and initiate commitments and Lehtinen and Lyytinen
[91 J, propose that an information system consists of speech acts that create, set up, control, and maintain
an organization’s transacting contracts, and report on their status. Flores and Ludlow [48], also point out
that “technology is not design of physical things. It is design of practices and possibilities to be realized
through artifacts” (p. 153).
Soft Systems Methodology: The view of the organizational role of an information system in SSM is not
very well articulated, but we conclude that it includes both technical and social aspects. Checkland and
Scholes [30] conceive an information system to entail data manipulation, which machines can do, and
meaning attribution which is uniquely a human act (p. 55). They furthermore suggest an approach for ISD
that is essentially based on defining a human activity system to stimulate debate on the relevance of the
human activity systems. Once a ‘truly relevant’ system has been agreed upon, one can proceed to the
analysis of information flows. For each activity identified in the relevant human activity, one can ask what
information would have to be available to enable someone to do this activity and what information would
be generated by doing it (pp. 56-57). The priority of the human activity system analysis reflects a social
view of an information system. The relatively mechanistic procedure of identifying required information
suggests that an information system is perceived to be an isolated technical artifact.
The Trade Unionist approach: The view of the organizational role of an IS in the Trade Unionist
approach has both a technical and social orientation. Earlier projects such as DEMOS, when analyzing
computer use in planning and control, emphasized that “in this area the computer serves as a controlling
instrument, in contrast with a device such as a lathe which could be categorized more aptly as a working
machine” ([27], p. 250). The second generation of the Trade Unionist approach, ‘the collective resource
approach’ [23, 391 strongly advocates a tool view as an ideal of computer artifacts, a tool being under
complete control of the users. This was associated with the technical view in the second section. Although
the authors do not make it clear, the emphasis on the tool view is partly explained by the specific
application area of the UTOPIA project, computer use in text and image processing in the graphic
industry. Even though the second generation adheres to the tool perspective, it at the same time
566 JUHANI IWARI AND RUDY HIRSCHHEIM
emphasizes that “the systems design process must be regarded as part of a larger organizational develop-
ment process” ([40], p. 33). Ehn [39] also points out that “in designing artifacts we do not merely design
the artifacts themselves: deliberately or not, we also design conditions for their human use” (p. 1) or “a
changed or reformed practice” (p. 171). He also underlines the social nature of the artifacts in several con-
texts (pp. 39, 100-101, 124, 208). It is noteworthy, however, that the design in Ehn [39] is assumed to
take place in terms of the computer artifacts and their use situations (cf. [57]) without providing any de-
liberate approach for designing the work practice per se independently of the computer artifacts+. The
likely reason for this paradox in the “work-oriented design” is that the essential part of work knowledge is
assumed to be tacit [39, 40, 891 and therefore largely beyond the scope of research. As tacit, it is not
describable, it is difficult to communicate, reflect upon and deliberately design without “design by doing”
using mock-ups and prototypes, etc ‘. It seems that the third generation, ‘the cooperative design’ approach
[56] has not changed this tool-oriented view of computer artifacts.
5.2 Views of information requirements
Table 5 depicts the views of information requirements in the eight ISD approaches.
Objective Subjective Intersubjective
Information Modelling *** ** *
Decision Support Systems ** ***
Sociotechnical approach *** *
Infological approach *** *
Interactionist approach
Speech Act-based approach ** ***
Soft Systems Methodology ** **
Trade unionist approach ** * *
*** Strong orientation ** Some orientation * Weak orientation
Table 5: Views of Information Requirements
Information Modelling: The view of user information requirements in the IM approach is
predominantly objectivist, but includes some subjectivist and intersubjectivist features. The objectivist
stance in IE [98-1001 is clearly visible in the whole approach. The identification of business areas in Infor-
mation Strategy Planning (ISP) is essentially based on the cross checking of business functions and
entities (data subjects). Both functional modelling and entity modelling in ISP ,take place in a very
impersonal way as a process of identifying fundamental business functions and entities. Business Area
Analysis continues the entity analysis leading to a normalized (usually 3NF) data model and a more
’ In fact, one can argue the title of Ehn’s [39], thesis “Work-oriented design of computer artifacts” is a misnomer. A more
appropriate title would be “Computer artifact-oriented design of work”.
* Button and Harper [22], contend that if the design of computer artifacts is to be based on understanding of the actual lived-work
of those who must use the computer system, this has to be grounded in analytical explications of work that reveal the practices
through which the users orderly handle the contingencies of their work situations.
Analysing Information Systems Development 567
detailed process model. The most articulated procedure for specifying individual application systems,
JAD, suggests that specification is very heavily based on further analysis of the processes at the level of
procedures [ 1001. The subjective view in IE is most clearly visible in the use of Critical Success Factors in
the identification of information requirements especially for executive information systems [loo]. The
intersubjective view is visible only weakly. One can nevertheless identify some flavor of intersubjectivity
in the process of integration of different user views ([99], pp. 253-254)+.
Olle, et al. [ 1091, are not very explicit in the case of information requirements, but its view is mainly
objectivist. Information requirements are implicitly based on impersonal business activity and entity
modelling. The authors also briefly introduce critical success factors whick again display objectivists
leanings.
Derision Support Systems: The view of information requirements in the DSS tradition is pre-
dominantly subjectivist. Keen and Scott Morton [78] emphasize in several contexts that a DSS must be
tailored to match the decision maker (pp. 6, 11, 50, 58). They also introduce individual differences as an
important model of decision-making (pp. 73-77), remarking that the: “implications of individual diffe-
rences for DSS are obvious. The system should mesh with the cognitive structure of its users” (p. 73).
They also utilize the perspective effectively in the DSS design process, in particular in the comparison of
the descriptive and normative models in the predesign phase of DSS development (pp. 174-176). The
descriptive and normative models lead us to question to what extent the definition of information
requirements in Keen and Scott Morton [78], reflects the objective view. Unfortunately, the authors are
not very explicit in defining these concepts. Our interpretation is, however, that the descriptive models
usually reflect the subjective view (cf. [78] p. 175), whereas the normative models may be based on a
more objective analysis of the decision-making activity* . Keen and Scott Morton [78], emphasize, how-
ever, that the normative models may be infeasible, and advocate matching the initial DSS to decision
makers’ current processes and later evolution of the system toward the normative model by facilitating
learning and willingness of decision-makers to explore alternative models (pp. 174-176). Earlier in their
book, Keen and Scott Morton [78], introduce a framework for information systems based on Anthony’s
hierarchy of planning and control activities and Simon’s distinction between programmed (structured) and
nonprogrammed (unstructured) decisions (pp. 79-98). Even though they discuss the differences in
information requirements at different levels of decision-making (pp. 82-85), it seems to us that the
framework is primarily presented to locate and distinguish DSS as a category of information systems.
rather than as an effective method to be used in information requirements analysis*
Sprague and Carlson [ 1231, also recognize subjective factors as part of information requirements, but
their conclusions concerning how to meet these requirements seem to differ from those of Keen and Scott
Morton [7X]. They emphasize that a DSS should “provide decision makers with a set of capabilities to
apply in a sequence and form that fits each person’s cognitive style” (p. 27), but later on conclude that “if a
DSS is to support varying styles, skills, and knowledge, it should not attempt to enforce or capture a
particular pattern. Rather, the DSS should help decision makers use and develop their own styles, skills
and knowledge. If this requirement can be met, the cost effectiveness of DSS should improve because
several decision makers could make effective use of the same DSS” (pp. 98-99). The authors do not
clearly state how the requirements and capabilities of this kind of a generalized DSS are derived: Are they
generalized from the subjective requirements of individual decision makers, possibly aiming at increasing
the intcrsubjectivity of’ related decision makers on the requirements, or are they based on a more
“objective” analysis of the decision-making activity?
It should be noted that Sprague and Carlson [ 1231 do not suggest any specific methods for the
“ob,jective” analysis of information requirements. They introduce Anthony’s classical hierarchy, Simon’s
’ Even though Martin [99] emphasizes that data modelling in BAA is a synthesis procedure (p. 247), he discusses the (user) YEW
integration only superficially.
* For instance, they may be based on some “theory” of the decision-making activity in questton (e.g. financial portfolio theory rn
the case portfolio management). It is, of course, by no means necessary that these “theories” omit individual differences (e.g. in risk
aversion), but they typcially are impersonal.
* In fact. Keen and Scott Morton [78] only briefly refer to the framework in their later description of the DSS design process (pp.
17% 187). they suggest it as a means to identify which decisions may benefit from decision support (p. 173).
568 JUHANI IIVARI AND RUDY HIRSCHHEIM
distinction between the structured and unstructured decisions, Simon’s intelligence-design-choice model
for a decision-making process as well as the Hackathorn’s and Keen’s classification of decision
interdependencies into independent, sequential interdependent and pooled interdependent (pp. 26-27, 94-
95). However, most of these frameworks are not effectively used in the later information requirements
analysis and definition of DSS capabilities, based on their ROMC framework (pp. 99-119). In fact, only
the intelligence-design-choice model is clearly referred to, but it is not clear whether the authors suggest it
as a method to be used in information requirements analysis or whether it is an illustration framework for
the components of the ROMC model. Taking into account the relative lack of methods for the “objective”
analysis of information requirements, we conclude that the requirements and capabilities are derived as a
generalization of subjective requirements of individual decision makers. This generalization is not
specifically introduced in terms of intersubjectivity+ .
The Socio-Technical Design approach: Since Pava [112], hardly addresses the definition of in-
formation requirements, the following analysis of the STD tradition is based on ETHICS [ 1021. We
conclude that the view is mainly objectivist, since key information needs in ETHICS are derived starting
from the mission and key objectives of the part of the organization which lies within the design boundary,
proceeding then to key tasks and finally to their information needs ([102] pp. 72-74). It is surprising, how-
ever, that in later design of organizational and technical options these key information needs are not taken
up or at least not explicitly integrated in the design of these options (pp. 91-97). The design of
organizational options refines the key tasks into various operating, problem prevention/solution, co-
ordination, development and control activities, which are allocated to different organizational units (large
work groups, their sub-groups, small teams and individuals), whereas the technical options are mainly de-
fined in terms of technical components, the descriptions of which implicitly include some understanding
of the conceptuaYinfologica1 features of the system. There is no explanation for these features, but overall
the process - as a seemingly impersonal activity analysis - supports an objectivist view. On the other hand,
the user participation emphasized in ETHICS brings a subjective element into information requirements
analysis* .
The Znfological approach: In the case of the Infological approach the view of the information
requirements is clearly objectivist. Information requirements in ISAC are derived using activity analysis in
change analysis and activity studies. After decomposing activities into more detailed ones, it is possible to
identify information processing activities which are potential candidates for information subsystems (cf.
[94] pp. 148-152). Further information analysis continues the refinement of each information subsystem
using precedence, component and process analyses, which specify the inputs and output and the derivation
rules of the subsystems more precisely (pp.198-203). These analyses take place in a very impersonal
manner. But, as in the case of ETHICS, user participation indirectly supports a subjective view. The view
of information requirements in [ 1251 can be interpreted to have a clear subjectivist flavor, however. He
assumes that “the ultimate purpose of a data base is to serve as a source of knowledge to more or less well-
defined group of end users. The end users need information, that is knowledge, in order to solve problems
or make decisions” (p. 2). Sundgren [125], interprets information as new knowledge, emphasizing that
information is always somebody’s information and exists only in the mind of a human being as a part of a
mental frame of reference (p. 10).
The Interactionist approach: The Interactionist approach does not address information requirements to
the extent to allow any conclusions to be drawn.
The Speech Act-based approach: The view of information requirements in the SA-based ap-preach is
most distinctively intersubjective, but also includes objectivist aspects. Goldkuhl and Lyytinen [54] claim
that ISD means to make common the rules of language by achieving formal language intersubjectivity (p.
20). The linguistic orientation naturally emphasizes common understanding among the participants in the
communication [91]. On the other hand, the concrete examples of the application of the SA-based
’ The authors remark, however, that group and organizational support systems “need to provide personal as well as group
instances of ROMCs. The former are defined by and for individuals, while the latter are standard for all individuals in the group”
(p. 142). It is not clear, however, whether the consensus is concerned with the intersubjectivity of language and communication
among the user community or just more generic DSS capabilities.
* One can imagine that the consensus form of participation in Mumford [ 1021, also supports intersubjectivity of information
requirements. Mumford [ 1021, does not yet make this explicit. Recall also that ETHICS does not specifically include mechanisms
for the conceptual/infological design of information systems (cf. [95]).
Analysing Information Systems Development 569
approach to modelling information systems [5 , 7, 55, 911, suggest that the information requirements can
be derived from a rather objective analysis of agents, instrumental acts and related speech acts.
Soft Systems Methodology: Information requirements in SSM has both intersubjective and objective
features. As described in section 5.1, SSM assumes information requirements to be dependent on the
human activity system selected. Once a ‘truly relevant’ human activity system has been agreed upon, one
can proceed to the analysis of information flows. For each activity identified in the relevant human
activity, one can ask ‘what information would have to be available to enable someone to do this activity
and what information would be generated by doing it’ ([30] pp. 56-57). A distinguishing aspect of SSM is
that it assumes several systems models, human activity systems, potentially relevant for analysis and
understanding a given organization. These models are subjected to a debate and the selection of the truly
relevant human activity system is assumed to be such that different stakeholder groups can find
‘accommodation’, i.e. that they are prepared to ‘go along with’. This process of achieving
‘accommodation’ has aspects of intersubjectivity. Once the truly relevant human activity system has been
selected, however, the detailed information requirements analysis proceeds in a fairly objectivist way.
The Trade Unionist approach: The Trade Unionist approach is not very explicit in its view of
information requirements. This is partly explained by its history. Its first generation projects were more
interested in the conditions of the trade union influence on the decisions concerning computer systems
rather than on novel design approaches. The second generation, the “collective resource approach” is
entirely based on one project, UTOPIA, in which the aim was to develop computer-based tools for text
and image processing. Because of the nature of the application, the system relied on direct iconic
modelling (of page make ups) rather than on symbolic models (see [l] for the distinction between the
iconic and symbolic models). It seems that in this case information requirements were not the major
concern, but rather the implications of the system on work. The design approach in UTOPIA was heavily
based on the use of mock-ups. The third generation, “cooperative design” has extended the “design-by-
doing” approach to entail prototyping [26]. Bedker and Gr@nb%k [26], also include an example of proto-
typing a more conventional, symbolic information systems, but they do not describe in detail how the
initial information requirements for the prototype were derived.
Prototyping may in principle support objectivist, subjectivist and intersubjectivist views of information
requirements. Nevertheless, our general impression is that the Trade Unionist approach places its major
emphasis on objective, although essentially tacit, requirements of work.
6. SUMMARY AND CONCLUSIONS
6.1 Summary
This paper has analysed two sets of assumptions in eight ISD approaches: the organizational role of
information system and the view of information requirements. In the former case, the paper distinguished
three alternatives - technical, sociotechnical and social views of the organizational role of IS - and in the
latter case again three alternatives - objective, subjective and intersubjective views. The findings of the
paper indicate that IM and DSS among the established traditions have a fairly technical-mechanistic view
of the organizational role of information systems, while the view of the STD approach is sociotechnical.
The Infological approach seems to reflect all the three views. The four emerging approaches, the Trade
Unionist approach as the only exception, emphasize the social nature of an information system, forming in
this respect a striking contrast to the established approaches (cf. Table 4).
In the case of information requirements, the established traditions were dominated by the objective
and/or subjective views; the intersubjective nature of information requirements was largely omitted.
Among the emerging approaches, the SA-based approach and SSM to a lesser extent differed in their
emphasis on intersubjectivity (cf. Table 5).
6.2 Conclusions
It could be concluded that the newer, emerging approaches arose because the established traditions
failed to deliver on their stated promises (cf. [63, 641). Indeed, the emerging approaches have adopted
very different sets of assumptions in an attempt to overcome the limitations of the “orthodox” approaches.
510 JUHANI IIVARI AND RUDY HIRSCHHEIM
Whilst many in the field may know of the existence of these “emerging approaches”, precisely how they
differ - fundamentally and conceptually - is not widely known. This then is one of the key contribution of
the paper - to show how they differ.
The results of our analysis can be expected to be of interest from three angles: 1) from the viewpoint of
IS research and theory, 2) from the viewpoint of IS practice, and 3) from the viewpoint of IS education.
Before proceeding into them, let us summarize the significance of the paper from the latter two points. The
paper pointed out that the alternative views on the two assumptional dimensions have implications that can
be expected to also have practical relevance (see Tables 1 and 2). These implications are currently
mediated to practice by ISD methodologies, techniques and tools reflecting the approaches. Because of the
findings concerning the established traditions were primarily based on widely used textbooks, they are of
importance from the viewpoint of IS education, since the underlying assumptions are transmitted to future
IS professionals through education. Because of the above two-way influence on practice, we contend that
the prevailing assumptions need to be critically reflected upon.
The paper makes several contributions to IS research and theory. Firstly, the interpretations of the two
assumptional dimensions in Figures 1 and 2 provide bridges from the classical philosophical dimensions
of nominalism vs. realism and voluntarism vs. determinism (see [21]) to more concrete IS specific
concepts. Secondly, the distinctions between the two sets of views extend earlier frameworks and allow a
more refined analysis and understanding of ISD approaches and methodologies. For example, many
articles in the book of Jirotka and Goguen, [75], on requirements engineering apply a broad interpretation
of the sociotechnical systems, covering both the sociotechnical and social views above. Even though the
interpretations of the “technical” and “social” vary in the sociotechnical tradition, especially when
sociotechnical ISD approaches are included [107], the sociotechnical tradition views the distinction
between the two subsystems as unproblematic. Woolgar [ 1331, however, recognizes that this may be falla-
cious. In the case of information requirements, we wish to note that the intersubjectivity of information re-
quirements is not widely recognized. Intersubjectivity based on social agreement has fundamental
implications in the sense that it questions the possibility of complete requirements. Thirdly, the
frameworks summarized in Figures 1 and 2 direct attention to the potentially emergent nature of the social
meaning of information systems and information requirements, shaped in the social interaction of users, in
their work to be supported by an information system. Emergence naturally leads to the problem of IS
evolution, partly as the continued social construction of the meaning of the system by users and other
stakeholder groups and partly as the continued need to develop the system to better satisfy the changed
requirements. Fourthly, as pointed out above, the analysis helps to understand the underlying assumptions
of the eight approaches. Even though the analysis of the established traditions was mainly based on text-
books, we contend that the findings can be expected to reliably reflect the assumptions made by the
mainstream research in each of those traditions. However, the results are based on a qualitative
interpretation of the texts documenting the selected ISD approaches. Therefore, the results summarized in
Tables 4 and 5 should be interpreted not as absolute facts but summaries which indicate the orientations of
the approaches. Fifthly, we believe that the emerging approaches indicate the future trends of ISD
approaches and methodologies towards a richer understanding of the social nature of information systems
and a more intersubjective view of information requirements.
When claiming that ISD approaches are evolving towards a richer understanding of the social nature of
information systems and a more intersubjective view of information requirements, we do not suggest that
the alternative views are categorically wrong, only that they are biased. We accept that different
assumptions concerning the organizational role may be justified in the case of different information
systems; for instance a technical view may be totally warranted in the case of a personal information
system, but the important point is that most information systems are organizational in the sense of
supporting, implementing and/or constituting organizational activities. The more complex these
information systems are in their relationships with the host organization, the more relevant the social view
becomes. In the case of information requirements, we do not consider the three views mutually exclusive;
they can be applied co-existently, different information systems emphasizing the views differently. For
example, the subjective view may dominate a personal decision support system, and an objective view
may be highly relevant in a case of a routine operational-level information system, supporting some well-
defined organizational activity or a well-understood application area. Consider the Air Traffic Control
system case discussed in [13] and [115]. It appears that the core information requirements of the Air
Analysing Information Systems Development 571
Traffic Control system are objectively determined (e.g. information about airplanes, flights, etc). Their
ethnographic analysis of this case indicated, however, the significance of subtle, often implicit,
cooperation as central to the functioning of the system, the significance of intersubjective ‘at a glance’
understanding of controllers’ displays as essential for the controllers’ work, and the danger of personal
tailorability of the system.
The evolution of ISD towards towards a richer understanding of the social nature of information
systems and more intersubjective view of information requirements is partly explained by technological
developments and the evolution of application areas to cover knowledge work (e.g. CSCW). Information
systems are increasingly organizational and interorganizational communication systems. Such
communication systems, by definition, necessitate an increasing attention to the intersubjectivity of
communication. The established traditions largely do not provide such attention. The four newer
approaches, which are stronger in addressing these issues, are however, incomplete as ISD methodologies.
This poses a considerable challenge to the further development of ISD approaches and methodologies.
We hope this paper paves the way for such future work.
REFERENCES
[l J R.L. Ackoff and M.W. Sasieni. Fundamentul.~ ofOperurions Reseurch. John Wiley & Sons, New York (1968).
[2] R J. Anderson. Representations and requirements: the value of ethnography in system design. Humcm-Cornpurer Intemcfion,
9(2): 151-182 (1994).
[3] ANSVX3/SPARC. Interim report from the study group on data base management systems. FDT Bulletin of rhe ACM, 7(2)
(1975).
[4] C Argyris and D. Schon. Orgunizutionul Laming: A theory of ucfwn perspecrive. Addison-Wesley, Reading, Massac-
husetts (1978)
[SJ E. Auram&i, R. Hirschheim and K. Lyytinen. Modelling offices through discourse analysis: The SAMPO approach. Tile
Computer Journul, 35(4): 342-352 (1992).
[6] E. Auramiiki, R. Hirschheim and K. Lyytinen. Modelling offices through discourse analysis: A comparison and evaluation of
SAMPO with OSSAD and ICN. The Cornpurer Journul, 35(5): 492-500 (1992).
[7] E. Auram&i, E. Lehtinen and K. Lyytinen. A speech-act-based office modeling approach. ACM Trunscccf1on.s on Qmce
Injormatirm Systems, 6(2): 126-l 52 (1988).
[X] D.E. Avison and G. Fitzgerald. Inform&on Systems Development: Methodologies, Techniques und Tools Blackwell,
Oxford (1988).
[9] D.E Avison and A.T. Wood-Harper. Multiview: An Exploration in Informution Systems Development. Blackwell Scientific
Publications, Oxford (1990).
[IO] C. Banville and M. Landry. Can the field of MIS be disciplined?, Communications ofthe ACM, 32(l): 48-60 ( 1989).
1 i 11 R. Baskerville. Practitioner autonomy and the bias of methods and tools. In Nissen, H.-E., Klein, H.K. and Hirschheim, R.
(eds.), Information Sysrems Reseurch: Contemporary Approaches & Emergent Trends, pp. 613-397, North-Holland,
Amsterdam (1991).
1121 C. Batini, M. Lenzeri and S.B. Navathe. A comparative analysis of methodologies for database schema Integration ACM
Computing Surveys, 18(4): 323-364 (1986).
[ I-71 R. Bentley, T. Rodeen, P. Sawyer, I. Sommerville, J. Hughes, D. Shapiro and D. Randall. Ethnographically-mformed system
design for air-traffic control. Proceedings offhe Conference on CSCW’92, pp. 123-129, Toronto (1992).
[ 141 H.R. Beyer and K. Holtzblatt. Apprenticing with the customer. Communicutions ofthe ACM, 38(5): 4S.52 (1995).
[lS] R.J Boland, Jr. Conuol, causality and information systems requirements. Accounting, Orgunizutions and Sociefy, 4(4):
259-272 (1979).
[ 161 R.B. Bostrom and J.S. Heinen. MIS problems and failures. A socio-technical perspective: Part I: The causes. MIS Quartely,
l(3): 17-32 (1977).
[ 171 R.B. Bostrom and J.S. Heinen. MIS problems and failures. A socio-technical perspective: Part II: The application of socio-
technical theory. MIS Quartely, l(4): 11-28 (1977).
[18] F. Brooks, Jr. No silver-bullet, essence and accidents in software engineering. Computer, April: pp. IO-19 (1987).
[ 191 F. Brun-Cottan and P. Wall, Using video to represent the user. Communicrrfions @the ACM, 38(5): 6 1-7 1 ( 1995).
[20] T. Burns and G.H. Stalker. The Munugemenf oflnnovution. Tavistock Publications, London (1961).
[2 I] G. Burrell and G. Morgan. Sociological Purudigms and Orgunizurwnal Analysis Heineman, London (1979)
[22] G. Button and R. Harper. The relevance of ‘work-practice’ for design. Computer Supporred Cooperutive Work (CSCW),
4(4): 263-280 (1996).
IS 2117-C
572 JUHANI IIVARI AND RUDY HIRSCHHEIM
[23] S. Bodker, P. Ehn, J. Kammersgaard, M. Kyng and Y. Sundblad. A utopian experience, on design of powerful computer-
based tools for skilled graphic workers. ln Bjerkenes, G., Ehn, P. and Kyng, M. (eds.), Computers ond Democrucy, pp. 25 I-
278, Avebury, Aldershot (1987).
[24] S. Bodker, J. Greenbaum and M. Kyng. Setting the stage for design as action. In Greenbaum, J. and Kyng, M. (eds.), Design
at Work: Cooperutive Design ofComputer Systems, Lawrence Erlbaum Associates, pp. 139-154, Hillsdale, NJ (1991).
[25] S. Bodker and K. Gronbmk. Design in action: From prototyping by demonstration to cooperative prototyping. In Greenbaum,
J. and Kyng, M. (eds.), Design ut Work: Cooperutive Design of Computer Systems, pp. 197-218, Lawrence Erlbaum
Associates, Hillsdale, NJ (1991).
[26] S. Bodker and K. Gronbaek. Cooperative prototyping studies - users and designers envision a dental case record system. Jn
Bowers, J.M. and Benford, S.D. (eds.), Studies in Computer Supported Cooperative Work, pp.315332, Elsevier Science
Publishers B.V. North-Holland, Amsterdam (1991 b).
[27] J. Carlsson, P. Ehn, B. Erlander, M.-L. Perby and A. Sandberg. Planning and control from the perspective of labour: A short
presentation of the DEMOS project. Accounting, Grgunizutions and Society, 3(3/4): 249-260 (1978).
[28] R. Chin and K.D. Benne. General strategies for effecting changes in human systems. In Bennis, W.G., Benne, K.D. and
Chin, R. (eds.), The Plunning of Change (2nd ed.), Holt, Rinehart & Winston, London (1969).
[29] P. Checkland. Systems Thinking, Systems Practice. John Wiley & Sons, Chichester (1981).
[30] P. Checkland and J. Scholes. Soft Systems Methodology in Action. John Wiley & Sons, Chichester (1990).
[31] G.B. Davis. Strategies for information requirements determination. IBM Systems Journal, 21(l): 4-30 (1982).
[32] G.B. Davis and M.H. Olson. Munugement Information Systems: Conceptual Foundutions. Structure und Development.
McGraw-Hill, New York(1985).
[33] SK. Dewitz and R.M. Lee. Legal procedures as formal conversations: Contracting on a performative network. In DeGross,
J.I., Henderson, J.C. and Kosynski, B.R. (eds.), Proceedings of the Tenth Intemutionul Conference on Information Systems,
pp. 53-65, Boston, Massachusetts (1989).
[34] J.L.G. Dietz. Subject-oriented modelling of open active systems. ln FaJkenberg, E.D., Rolland, C. and El-Sayed E.N. (eds.),
Information Systems Concepts: Improving the Understanding, pp. 227-238, IFlP Transactions A-4, North-Holland,
Amsterdam (1992).
[35] J.L.G. Dietz and G.A.M. Widdershoven. Speech acts or communication action. In Bannon, L., Robinson, M. and Schmidt,
K. (eds.), Proceedings of the Second European Conference on Computer-Supported Cooperutive Work, pp. 235-248,
Kluwer Academic Publishers, Dordrecht (1991).
1361 J.E. Dobson, M.J. Martin, C.W. Olphert and SE. Powrie. Determining requirements for CSCW: the ORDIT approach. ln
Stamper, R.K., Kerola, P. and Lyytinen, K. (eds.). Colluborutive Work, Social Communicutions und Information Systems,
pp. 333-353, Elsevier Science Publishers B.V. North-Holland, Amsterdam (1991).
[37] S. Donaldson Dewitz. Contracting on a performative network: Using information technology as a legal intermediary. Jn
Stamper, R.K., Kerola, P. and Lyytinen, K. (eds.), Collaborative Work, Social Communications and Information Systems,
pp. 271-293, Elsevier Science Publishers B.V. North-Holland, Amsterdam (1991).
[38] S. Easterbrook. Resolving requirements conflicts with computer-supported negotiation. In Jirotka, M. and Goguen, J.A.
(eds.), Requirements Engineering, Sociul and Technical Issues, pp. 41-65, Academic Press, London (1994).
[39] P. Ehn. Work-oriented design of computer urtifacts Arbetslivscentrum, Stockholm (1988).
[40] P. Ehn and M. Kyng. The collective resource approach to systems design. In Bjerkenes, G., Ehn, P. and Kyng, M. (eds.),
Computers und Democrucy, pp. 17-57, Avebury, Aldershot (1987).
[41] P. Ehn and M. Kyng. Cardboard computers: Mocking-it-up and hands-on the future. In Greenbaum, J. and Kyng, M. (eds.),
Design ut Work: Cooperutive Design of Computer Systems, pp. 169-195, Lawrence Erlbaum Associates, Hillsdale, NJ
(1991).
[42] P. Ehn, M. Kyng, Y. Sundblad et al. The Utopia Project. In Briefs, U., Ciborra, C. and Schneider, L. (eds.), Systems Design
For, With, und By the Users, pp. 439-449, North-Holland, Amsterdam (1983).
[43] P. Ehn and A. Sandberg. Local union influence on technology and work organization. ln Briefs, U., Ciborra, C. and
Schneider, L. (eds.), Systems Design For, With, and By the Users, pp. 427-437, North-Holland, Amsterdam (1983).
[44] P. Ehn and D. Sjogren. From system descriptions to scripts for action. In Greenbaum, J. and Kyng, M. (eds.), Design ut
Work: Cooperative Design of Computer Systems, pp. 241-268, Lawrence Erlbaum Associates, Hillsdale, NJ (1991).
[45] D. Episkopou. The Theory und Pructice of Inform&ion Systems Methodologies: A Grounded Theory of Methodology
Evolution. Unpublished PhD Thesis. University of East Anglia, UK (1987).
[46] A.C.W. Finkelstein, D. Gabbay, A. Hunter, J. Kramer and B. Nuseibeh. Inconsistency handling in multiperspective
specifications. IEEE Trunsuctions on Software Engineering, 20(8): 569-578 (1994).
[47] A. Finkelstein and I. Sommerville. Editorial: The Viewpoints FAQ.Sofimure Engineering Journul, 11(l): 2-4 (1996).
[48] F. Flores, M. Graves, B. Hartfield and T. Winograd. Computer systems and the design of organizational interaction, ACM
Trunsuctions on office Information Systems, 6(2): 153-172 (1988).
1491 F. Flares and J.J. Ludlow. Doing and speaking in the office. In Fick, G. and Sprague, R.H. Jr. (eds.), Decision Support
Systems: Issues und Challenges, pp. 95-l 18, Pergamon Press, Elmsford, NY (1980).
[50] A. Giddens. The Constitution of Society. Polity Press, Cambridge (1984).
Analysing Information Systems Development 573
[511 J.A. Goguen. The dry and the wet. In Falkenberg, E.D., Rolland, C. and El-Sayed, E.N., Information Sysrem.~ Concepts:
Improving the Undmtunding , pp. 1 - 17, Elsevier North-Holland, Amsterdam ( 1992).
[52] J.A. Goguen. Requirements engineering as the reconciliation of social and technical issues. In Jirotka, M. and Goguen, J.A.
(eds.), Requirements Engineering. Sociul and Technical Issues, pp. 165.199, Academic Press, London (1994).
[53] G. Goldkuhl. Miinniskor, Information och Kommunikution - en Humuninfologisk Referensmm fZr Informutionsystem.
Institutionen for administrative databehandling, University of Stockholm, Stockholm (1980).
[54] G. Goldkuhl and K. Lyytinen. A language action view of information systems. In Ross, C. and Ginzberg, M. (eds.),
Proceedings ofthe Third Conference on Inform&on Systems , pp. 13-29, Ann Arbor, Michigan (1982).
[55] G. Goldkuhl and K. Lyytinen. Information system specification as rule reconstruction. In Bemelmans, Th.M.A. (ed.), Beyond
Productivity, Infi,rmution System7 Development for Orgunizutionul Eflectiveness, pp. 79-94, North-Holland, Amsterdam
(1984).
[56] I. Greenbaum and M. Kyng (eds.). Design ut Work: Cooperutive Desrgn of Computer Systems. Lawrence Erlbaum
Associates, Hillsdale, NJ (1991).
[57] J. Greenbaum and M. Kyng. Introduction: Situated design. In Greenbaum, J. and Kyng, M. (eds.), Design ur Work:
Cooperutive Design of Computer Systems, pp. l-24, Lawrence Erlbaum Associates, Hillsdale, NJ (1991).
[58] J.J. van Griethuysen (ed.). Concepts und Terminology for the Conceptuul Schemu and the Informution Base. IS0
‘TC97/SC5/WG3 Report (1982).
[59] M.R. Gustafsson, T. Karlsson and J. Bubenko, Jr. A declarative approach to conceptual information modeling. In Olle, T.W.,
Sol, H.G. and Verrijn-Stuart, A.A. (eds.), Information Systems Design Methodologies: A Compurutive Review, pp. 93-142,
North-Holland, Amsterdam (1982).
[60] M. Hammer and J Champy. Reengineering the Corporafion Nicholas Brealey Publishing, London (1993).
[6 I] P.G. Herbst. So&-Technicul Design, Stutegies in Multidisciplinury Research Tavistock Publications, London (1974).
[62] R. Hirschheim and H.K.KIein. Four paradigms of information systems development. Communicutions of the ACM, 32( 10):
1199-1216 (1989).
[63] R. Hirschheim and H.K. Klein. Paradigmatic influences on information systems development methodologies: evolution and
conceptual advances. In M. Yovits (ed.), Advunces in Computers, Academic Press, New York, 34: 293-392 (1992).
[64] R. Hirschheim, H.K. Klein and K. Lyytinen. Infr,rmation Systems Development und Dutu Modeling: Conceptuul and
Philosophciul Foundutions Cambridge University Press, Cambridge (1995)
[65] R. Hirschheim, H.K. Klein and K. Lyytinen. Exploring the intellectual structures of information systems development: A
social action theoretic analysis. Accounting, Munugemen t und Informution Technologies, 6(l) (1996).
[66] R. Hirschheim and G. Schlfer. Requirements analysis: A new look at an old topic. Journal of Applied Systems Anulysis, 15:
101-l 18 (1988).
[67] S. Iacono and R. Kling. Computer systems as institutions: social dimensions of computing in organizations. In DeGross, J.
and Olson, M. (eds.), The Ninth Internutionul Conference on Information Systems , pp. 101-l 10, Minneapolis (1988).
[68] I. Iivari. Taxonomy of the experimental and evolutionary approaches to systemeering. In Hawgood, J. (ed.). Evolutionary
irzformution Systems, pp. 101-l 19, North-Holland, Amsterdam (1982).
[69] 1. livari. Dimensions of information systems design: A framework for a long-range research program. Itzformution Systems,
ll(2): 185-197 (1986).
[70] J. Iivari. Levels of asbtraction as a conceptual framework for an informahon system. In Falkenberg, E.D. and Lindgreen, P.
[eds.), Inform&on System Concepts: An In-depth Anuljsis, pp. 323-352, Elsevier, North-Holland, Amsterdam (1989).
[71] J. Iivari. A paradigmatic analysis of contemporary schools of IS development. European Journul of Inform&on Systems,
l(4): 249-272 (1991).
[72] J. livari and R. Hirschheim. A paradigmatic analysis of four emerging information systems development approaches.
Proceedings ofthe Third Australian Conference on Information Systems , pp. 115-137, Wollongong, Australia (1992).
[73] J. Iivari and P. Kerola. A sociocybemetic framework for the feature analysis of information systems design methodologies. In
Olle, T.W., Sol, H.G. and Tully. C.J. (eds.), Information Systems Design Methodologies: A Feuture Analysts. pp. 87-139,
North-Holland, Amsterdam (1983).
1741 M.A. Janson and CC. Woo. Investigating information and knowledge gathering methods: A speech act lexicon perspective.
In Falkenberg, E.D., Rolland. C. and El-Sayed E.N. (eds.), Inform&ion Sysfemc Concepts: Improving the Understunding,
pp. 239-257, IFIP Transactions A-4, North-Holland, Amsterdam (1992).
[75] M. Jirotka and J.A. Goguen (eds.). Requirements Engineering, Social und Technicul Issues Academic Press, London (1994)
[76] P. Johannesson. Representation and communication a speech act based approach to information systems design.
Infiwmution Systems, 20(4): 291-303 (1995).
[77] P.G.W. Keen. Adaptive design for decision support systems. Datu Buse, 12(1/2): 15-25 (1980).
[78] P.G.W. Keen and M. Scott Morton. Decision Support Systems: An Orgunizutionul Perspective. Addision-Wesley, Reading,
Massachusetts (1978).
[79] F. Kensing and T. Winograd. The language/action approach to design of computer-support for cooperative work: A
preliminary study in work mapping. In Stamper, R.K., Kerola, P. and Lyytinen, K. (eds.), Colluborutive Work, Social
Communicutions and Inform&on System, pp. 3 1 l-331, Elsevier Science Publishers B.V. North-Holland, Amsterdam
(1991).
574 JUHANI IIVARI AND RUDY HIRSCHHEIM
[80] R. Kling. Defining the boundaries of computing across complex organizations. In Boland, R. and Hirschheim, R. (eds.).
Criticul Issues in Information Systems Research, pp. 307-362, John Wiley & Sons, Chichester (1987).
[81] R. Kling and S. Iacono. The control of information systems developments after implementation. Communicutions of the
ACM, 27(12): 1218-1226 (1984).
[82] R. Kling and S. Iacono. The institutional character of computerized information systems. Office: Technology und People,
S(1): 7-28 (1989).
[83] R. Kling and W. Scacchi. Computing as social action: The social dynamics of computing in complex organizations. In
Yovits, M.C. (ed.), Advunces in Computers, 19: 249-327 (1980).
[84] R. Kling and W. Scacchi. The web of computing: Computer technology as social organization. In Yovits, M.C. (red.),
Advances in Computers, 21: l-90 (1982).
[85] G. Kotonya and I. Sommerville. Requirements engineering with viewpoints. Softwure Engineering Journal, 11(l): 5-18
(1996).
[86] J. Krogstie, 0.1. Lindland and G. Sindre. Towards a deeper understanding of quality in requirements engineering. In Iivari, J.,
Lyytinen, K. and Rossi, M (eds.), Advanced Information Systems Engineering, pp. 82-95, Springer-Verlag, Berlin (1995).
[87] T.S. Kuhn. The Structure of Scientific Revolutions (first edition 1962 and second enlarged edition 1970). The University of
Chicago Press, Chicago (1970).
[88] S. Kuhn and M.J. Muller. Participatory design. Communicutions ofthe ACM, 36(4): 24-28 (1993).
[89] M. Kyng. Designing for cooperation: cooperating in design. Communicutions ofthe ACM, 34( 12): 65-73 (1991).
[90] B. Langefors. Theoretical Anulysis oflnjiormution Systems. Studenthtteratur, Lund, Sweden (1966).
[91] E. Lehtinen and K. Lyytinen. Action based model of information systems. Information Systems, ll(3): 299-317 (1986).
[92] J.C.S.P. Leite and P.A. Freeman. Requirements validation through viewpoint resolution. IEEE Trunsuctions on Softwure
Engineering, SE-17(12): 1253-1269 (1991).
[93] P. Luff, C. Heath and D. Greatbatch. Work, interaction and technology: The naturalistic analysis of human conduct and
requirements analysis. In Jirotka, M. and Goguen, J.A. (eds.), Requirements Engineering, Social und Technical Issues,
pp. 259-288, Academic Press, London (1994).
[94] M. Lundeberg, G. Goldkuhl and A. Nilsson. Informution Systems Development: A Systemuric Approach. Prentice-Hall,
Englewood Cliffs, NJ (1981).
[95] K. Lyytinen. A taxonomic perspective of information systems development: theoretical constructs and recommendations. In
Boland, R. and Hirschheim, R. (eds.), Criticul Issues in Inform&on Systems Reseurch, pp. 3-41, John Wiley & Sons,
Chichester (1987).
[96] N.A.M. Maiden, P. Assenova, P. Constantopoulos, M. Jarke, P. Johannesson. H.W. Nissen, G. Spanoudakis and A.G.
Sutcliffe, A.G.. Computational mechanism for distributed requirements engineering. Proceedings of SEKE’95, The 7th
Internutionul Ccmference on Sojiwure Engineering and Knowledge Engineering , pp. 8-15, Rockville, MD (1995).
[97] L.M. Markus. Power, politics and MIS implementation. Communicutions of the ACM, 26(6): 430-444 (1983).
[98] J. Martin. Inform&on Engineering. Book I, Introduction Prentice-Hall, Englewood Cliffs, NJ (1989).
[99] J. Martin. Information Engineering, Book II, PIunning and Anulysis . Prentice-Hall, Englewood Cliffs, NJ (1990a).
[lOO] J. Martin. Inform&on Engineering, Book III, Design and Construction Prentice-Hall, Englewood Cliffs, NJ (199Cb).
[loll E. Mumford. Systems Designfor People. NCC, Manchester (1971).
[102] E. Mumford. Designing human systems for new technology. The ETHICS method. Manchester Business School,
Manchester (1983).
[103] H.W. N&en, M.A. Jeusfeld, M. Jarke, G.V. Zemanek and H. Huber. Managing multiple requirements perpsectives with
metamodels. IEEE SofhYure, 13(2): 37-48 (1996).
[104] M.I. Nurminen. People or computers: Three wuys of looking at information systems. Studenthteratur/Chartwell-Bratt, Lund,
Sweden (1988).
[IO51 B. Nuseibeh, J. Kramer and A. Finkelstein. A framework for expressing the relationships between multiple views in
requirements specifications. IEEE Trunsuctions on Suftwure Engineering, SE-%& 10): 760-773 (1994).
[106] K. Nygaard and O.T. Bergo. The trade unions - New users of research. Personnel Review, 4(2): 5-10 (1974).
[107] A. Olerup. Socio-technical design of computer-assisted work: A discussion of the ETHICS and Tavistock approaches.
Scandinavian Journul of Inform&ion Systems , 1: 43-71 (1989).
[108] W. Odikowski and D. Robey. Information technology and structuring organizations. Inform&on Systems Research, 2(2):
143-169 (1991).
[log] T.W. Olle, J. Hagelstein, I. Macdonald, C. Rolland, H. Sol, F. Van Assche and A. Verrijin-Stuart. Informution Systems
Methodologies - A Frameworkfor Understanding. Addison-Wesley, Wokingham, U.K. (1988).
[ 1 lo] T.W. Olle, H.G. Sol and A.A. Verrijn-Stuart (eds.). Information Systems Design Methodologies: A Compurutive Review.
North-Holland, Amsterdam (1982).
[ 11 l] T.W. Olle, H.G. Sol and A.A. Verrijn-Stuart (eds.). Informution Systems Design Methodobgies: Improving the Pructice.
North-Holland, Amsterdam (1986).
[ 1121 C. Pava. Managing New mce Techncifugy: An Orgunizutionul Strutegy The Free Press, New York (1983).
AnaIysing Information Systems Development 575
[ll31 K. Pohl. The three dimensions of requirements engineering: A framework and its applications. ~~for~ti~,n ~y,~re~, 19(3):
243-258 (1994).
[I I41 C. Potts. K. Takashi and A.I. Ant6n. Inquiry-based requirements analysis. IEEE S@vure, ll(2): 21.32 (1993)
[I 151 D. Randell, J. Hughes and D. Shapiro. Steps toward a partnership: Ethnography and system design. In M. Jirotka and J.A.
Goguen (eds.), Requirements Engineering, Sociul und Technic& Issues, pp. 241-258, Academic Press, London (1994).
[I 161 A. Sandberg. Trade union-oriented research for democratization of planning in work life - problems and potentials. _/ournu[
of‘Occuputionu1 Behavior, 4(l): 59-71 (1983).
[l 171 A. Sandberg. Socio-technical design, trade union strategies and action research. In Mumford, E., Hirschheim, R.A.,
Fitzgerald, G. and Wood-Harper, A.T. (eds.), Research Mahods in Inform&on Sysrem.~, pp. 79.92. North-Holland,
Amsterdam (1985).
11181 G. Schlfer et al. Functional Analysis of Of&e Requirements: A Mulfiperspecfive Approuch. John Wiley & Sons. Chichester
(1988).
[I 191 J.R. Searle. Speech Acfs, An Essuy in the Phifosohy ofL.unguqe Cambridge University Press, Cambridge (1969).
[120] J.R. Searle. Expression and Meaning, Studies in fhe Theory of Speech Acts. Cambridge University Press. Cambridge
(1979).
[I211 J.R. Se&e and D. Vanderveken. Foundutions oflk~cufionury LOOK Cambridge University Press, Cambndge (1985).
[l22] J. Similii. ADP systems implementation success - theoretical perspectives and a model in terms of success In the user role. In
S&iksjtii, M. (eds.), Report of the Seventh Scundinuviun Reseruch Seminur on Sysremeerin,q: Purt II, pp. 62-85, Helsinki
School of Economics, B-75 (1984).
[ 1231 R.H. Sprague and E.D. Carlson. Building I?ecfive Decision Support Systems. Prentice-Hall, Englewood Cliffs, NJ (1982).
f124] L.A. Suchman and R.H. Trigg. Understanding practice; Video as a medium for reflection and design. ln Greenbaum, J. and
Kyng, M. (eds.), Design ut Work, Cooperutive Design of Computer Sysfem~, pp. 65-89, Lawrence Erlbaum, Hillsdale. NJ
(1991).
[125] B. Sundgren. Theory r,fDufu Buses. PetrocellXharter, New York (1975)
[I261 W.M. Taggart, Jr. and M.O. Tharp. A survey of information requirements analysis techniques. Compuring Surveys. 9(4):
273-290 (1977).
[ 1271 N. Tichy. A social network perspective for organizational development. In Cummings, T.G. (ed.), Sysrem.? Theory ,for
Orgunizutionul Devekpmenf, pp. 115-162, John Wiley & Sons, Chichester (1980).
[ 1281 E.L. Trist. The sociotechnical perspective, the evolution of sociotechnical systems as a conceptual framework and action
research program. In Van de Ven A.H. and Joyce, W.F. (eds.), Perspectives on Orgunizution Design und Behuvror. pp. 19-
75, John Wiley & Sons, New York (1981).
[ 1291 D.P. Truex, III and H.K. Klein. A rejection of structures as a basis for information systems development. ln Stamper, R.K.,
Kerola, P., Lee, R. and Lyytinen, K. (eds.), Cofluborutive Work. Soctul Communicufions und I@-mufion Systrms. pp. 2 13-
235, Elsevier (North-Holland), Amsterdam (1991).
[130] G. Walsham. lnterprering Infwmution System.~ in Orgunizutions John Wiley & Sons, Chichester (1993).
[ 13 1] B. Wilson. Systems: Concepts, Methodologies, und Applicufions John Wiley & Sons, Chichester (1984).
[132] T. Winograd and F. Flares. Understunding Computers und Cognition: A New Foundurion fi)r Deslfin Ablex. Norwood, NJ
(1986).
[ 1331 S. Woolgar. Rethinking requirements analysis: Some implications of recent research into producer-consumer relationships
on IT development. In Jirotka, M. and Goguen, J.A. (eds.), Reyuirementr Engineering, So&l and Technic& Issues, pp. 201-
216, Academic Press, London (1994).
[ 1341 A.T. Wood-Harper and G. Fitzgerald. A taxonomy of current approaches to systems analysis. The Computer Journul. 25( I):
12-16 (1982).
