Requirements Eng (2000) 5:67±73 ß 2000 Springer-Verlag London Limited

Requirements Engineering

Socio-Technical and Soft Approaches to Information Requirements Elicitation in the Post-Methodology Era Christopher J. Atkinson

Department of Information Systems and Computing, Brunel University, Uxbridge, UK

The contributors to this special issue focus on sociotechnical and soft approaches to information requirements elicitation and systems development. They represent a growing body of research and practice in this ®eld. This review presents an overview and analysis of the salient themes within the papers encompassing their common underlying framework, the methodologies and tools and techniques presented, the organisational situations in which they are deployed and the issues they seek to address. It will be argued in the review that the contributions to this special edition exemplify the `postmethodological era' and the `contingency approaches' from which it is formed. Keywords: Information requirements; Information systems development; Methodologies; Research; Sociotechnical; Soft systems

1. Introduction The authors in this special edition [1±6] universally argue that seeking to establish information requirements cannot be reduced to either a rational process or the production of technical speci®cation in some form of data model, nor can it ever be a one-off event ± especially in the current turbulent organisational and interorganisational environments in which much of current information systems development (ISD) takes place, replete with their politics and multiple stakeholders. This, they argue, has profound implications for Correspondence and offprint requests to: Department of Information Systems and Computing, Brunel University, Uxbridge UB8 3PH, UK. Email: [email protected]

ISD and systems usability within and between organisations, particularly in the area of information requirements elicitation (IRE). Having said this, none of the authors reject the notion that information systems or the processes underpinning their development are devoid of any technical component or for that matter any form of rationality. Indeed, the majority of them put forward methodologies, tools and techniques of their own that seek to address the human or organisational dimension through a combination of reason and participation while arguing, as we shall see, for their complementarity with existing ISD tools and techniques. The authors, however, do question the utility of what have become known as `hard' or structured, `data-driven' [7] approaches to ISD, such as SSADM [8], Jackson [7] and information engineering [8] with their reliance on the one-off identi®cation of information requirements. Paul [9] has termed this their `®xed point theorem'. What then is the nature and status of this soft/sociotechnical paradigm to ISD and IRE? This review seeks to summarise and analyse the contributions to the special edition and, in so doing, map out the paradigm they represent and constitute, arguing in the concluding discussion that they are manifestations of what has been termed the `post-methodological era' [10].

2. A Template for Analysis To address the question set out in the introduction it is ®rst necessary to analyse the salient themes running through the contributions. Checkland and Scholes [11] suggest a template of inquiry through which to review the authors' contributions (see Fig. 1). The template encompasses an underpinning framework ± here Galliers

C. J. Atkinson

68

Fig. 1. Frameworks, methodologies and organisational IS requirements problem solving. (After Checkland [13] and Galliers and Swan [1]).

and Swan's [1] ± through which ISD methodologies and approaches are classi®ed. Organisations consisting of integrations of human and machine activity, sitting within a wider environment, choose and then use the approaches to address their ISD/IRE and any associated problems. This, in turn, affords an opportunity for learning about and developing the framework(s), its methodologies, their use and organisational impact ± in so doing, learning about learning takes place.

3. Salient Themes Using the template in Fig. 1, this section explores the salient themes emerging from the contributions to this special edition to map out the inherent paradigm, summarised in Table 1.

3.1. The Framework From the summary in Table 1 there is, as would be expected, a common underpinning world-view to the contributors' work; namely the soft [12,13] or sociotechnical perspective [6]. Galliers and Swan [1] see it as based on a pluralism found in social networks and the individual's subjectivity they contain, as opposed to the unitary top-down organisational formalism common to most data-focused approaches to ISD and information requirements elicitation, such as Martin's information engineering [7]. This can be found in all the papers in different guises, for example Sutton's [5] `epistemological subjectivity' and Al Karaghouli et al. [3] seeking a reconciliation or acknowledgement of the `gaps' between the customer's and the analyst's perspectives, or the participation of Mumford's [6] ETHICS methodology.

Intra-/inter-organisational networks

Soft/subjective pluralism Versus Objective/formal and Unitary approaches

Epistemological subjectivity Organisations have an Language and incommensurate plurality of communications are agendas and perspectives problematic sources of information requirements IT and humans Engineering sector

Socio-technical approach

Sutton

Mumford

ISD is problem solving

Ethnography with sociotechnical and SSM approaches

Participation

Organisational competencies, company structures and working practices

Multiple stakeholders and perspectives

New technology

E-companies

Global market

Inter-organisationally usable

Internal to the organisation

Retail sector

Con¯icting perspectives of user and developer

Coakes

Multi-stakeholder

Developer's problem of understanding user requirements

Human and technological dimensions

Inter-organisational networks in the health sector

Inter organisational collaboration between various parties

Al Karaghouli et al.

NHS, clinical professions

Business processes

Socio-technical approach

Organisational boundary spanning

Inter-organisational communication networks

Organisational context

Patel

Multiple stakeholders

Change management

A Framework for ISD

Galliers and Swan

Organisation and issues

Framework

Author

Table 1. The papers' frameworks, methodologies and organisational IS requirements problem solving Information tools/techniques

Development

Information systems

Information requirements

ETHICS, Multiview

Multidisciplinary

Multiview ± SSM modelling, DFDs, HCI, job design

ETHICS ± Socio-technical Objectives, options, designs, participatory decision making

Requirements ± incomplete, change with organisation

IS features are experiences, not abstract formalisms

Continuous approaches

Con¯icting and competing data

Object oriented

Prototyping and RAD

Requirements engineering

Con¯ict management

Pluralistic approaches

Fact ®nding, data modelling O-O

Ethnography participative

Graphical requirements Knowledge requirements systems representation techniques ± Venn ± ISD diagrams Reconciliation of customer (user) Soft systems methodology ± rich and developer views pictures

Role activity diagrams

Socially mediated information 7hard/soft distinction required

Advocates soft systems and IS requirements planning informal methodologies to ISD and requirements and knowledge Information support strategy management. No unitary approaches Knowledge management

Methodology

Information Requirements Elicitation in the Post-Methodology Era 69

C. J. Atkinson

70

3.2. The Organisation The underlying framework [1] is replicated in the view of the organisation that pervades the contributions. Organisations are seen not as mechanisms or even organisms. Rather, they are seen as socio-technical networks, heterogeneous domains of both mutual and competing interests, sometimes, as in Coakes and Coakes' work [4], in con¯ict. This is not to deny that out of the interactions of humans and technologies comes concerted purposeful action [12]. It clearly does ± goods are delivered, patients are treated. But in this paradigm organisations cannot be reduced to pure functionality. Organisations and interorganisational networks, as the contributors point out, are imbroglios of politics and multiple `circuits of power' [14]. Technology-based IS accounts for only a percentage of the data called on and used either as part of concerted purposeful action or in organisational politicking. In social networks there are no independent data entities [7] or objects [7], only the shifting partisan vocabularies of the individual or the group. They have no existence or meaning beyond the world-views and practices of those who subjectively hold them, or outside the transient discourses in which they appear. The fact that communication takes place at all is a result of the human capacity to deal with the ad hoc nature of discourse and meaning contexts. But even this is open to misinterpretation, which further clari®cation can alleviate but never totally eliminate. For these reasons eliciting information requirements in organisations can never be engineering. Rather, as Sutton [5] points out, it is a matter of providing an environment in which requirements can be expressed as part of actual working practices. For him, even a person's speaking about his/ her own information needs is one step removed from the wide range of data and information a person accesses from a multiplicity of sources, including computer-based IS, when dealing in situ with a situation in an organisation, working individually or collectively (though not all contributors would agree on this point). Change in organisations in this framework, as Mumford [6] emphasises, is seen as incessant, driven by globalisation, markets, governments, technological innovation, professional practices, organisational takeovers and rationalisations. That information requirements change and are contested, even in a project's lifetime, however, is not to be regretted or obviated as in the established approaches. Change in fact has to be accepted, even welcomed, as a measure of organisational vitality [5], its response to the global market, and the political and technological environments in which it operates. All this has a considerable impact on ISD and IRE. Required is a set of what Mumford [6] sees as

organisational and personal competencies, methodologies and approaches capable of accommodating this change and diversity, yet capable of ameliorating con¯ict. Paul [9] identi®es this as the `living systems' approach.

3.3. The Organisational Contexts Within this special edition diverse organisations and industries are cited as arenas in which the soft and sociotechnical approaches to ISD have been applied. The retail sector, [3], engineering [5], healthcare [2], transient e-commerce or `e-built to last' businesses [6], manufacturing and production specialists [4] are all represented. Like organisations these contexts, industries and sectors are reported as themselves being in ¯ux, impacting upon the organisations in which ISD and IRE are taking place. This incessant `churn' has profound rami®cations for the organisation, inter-organisational networks, business processes, current and future working patterns and practices, strategic expenditure on IT and the design, development and functional adaptivity of information systems. The methodologies and tools presented by the authors here seek to address these.

3.4. The Methodologies This framework is not without its methodologies or its tools and techniques, as the contributions indicate. Soft systems approaches such as Avison and Wood-Harper's Multiview [15] or the socio-technical ETHICS of Mumford [6] are appealed too. Mumford emphasises participatory decision making around human and technological options appraisal and design. Mulitiview uses SSM [12] tools, such as primary task models, data modelling, the human±computer interface (HCI) and job design (using ETHICS). More recently there is the reviewer's own soft information systems and technologies methodology (SISTeM) [16,17] that integrates soft systems with ISD and is based on the concept of the human/machine activity system. They each have the capacity to encompass the human and organisational aspects of ISD and IRE as well as the technological. Of particular importance is their ability to accommodate multiple stakeholders and politics, through incorporation and shared decision making, aimed at a robust accommodation between various viewpoints and agendas. In addition they also seek to facilitate effective information system design by incorporating the tools and techniques of the more structured approaches such as data ¯ow diagrams, entity diagrams and object-oriented design. One other feature is that these approaches are

Information Requirements Elicitation in the Post-Methodology Era

¯exible, meldable to the circumstances of their use by those who use them. We shall return to this theme in the concluding comments.

4. The Tools and Techniques The authors in this special edition propose a number of tools and techniques, some new to ISD and IRE. They also argue for working with established data-driven tools and techniques, but not the structured approaches they are associated with. Al Karaghouli et al. [3] use mathematical set theory to represent the `knowledge and understanding gaps' about information requirements that exist between the various `customers' and the ISD developer. The existence and nature of these different positions is revealed through the creation of an SSM-rich picture [12,13] of the problem situation. These are used, through participation and exploration, to ameliorate the `gaps' between the disparate parties as part of wider IRE process based on extensive research.

71

A further formal tool for addressing the organisational con¯icting viewpoints is found in the work of Coakes and Coakes [4], where ethnographic techniques are deployed in mapping out various and opposing information requirements. These enhance socio-technical and soft approaches ± relieving them, Coakes and Coakes [4] propose, of their over-dependency on interpretivism. They then go on show how such an analysis can produce a data model that re¯ects these con¯icting and competing perspectives which will subsequently feed into an object-oriented design. For them `improves interactivity and con¯ict management' as well as ISD and IRE. Another of the authors who links their work to established ISD approaches is Sutton [5]. He provides a socio-technical analysis of the linguistic dif®culties inherent in eliciting information requirements based solely on what people say is the information they use or need, as opposed to the actual information and technologies they spontaneously use in their working practices. These dif®culties, inherent in the very nature of language use, he argues, can only be addressed

Fig. 2. The soft/socio-technical domain of the framework (after Galliers and Swan [1]) for the classi®cation of information systems development approaches.

C. J. Atkinson

72

through an iterative and potentially never-ending prototyping approach to ISD and IRE in which requirements are intrinsically present. Patel [9] employs another modelling approach, role activity diagrams (RADs), to support information requirements elicitation. RADs, he argues, moves the emphasis away from data modelling to representing the organisational and human roles embedded within business process work ¯ows ± in his example ± within and between healthcare organisations. RADs need to be produced of the current situation, the `as is' model, and of the anticipated future, the `to be' model. This combination enables not only human interactions to be captured but also IS entities to be identi®ed. This has the advantage, as with the other methodologies cited here, of enabling the various parties to explore the development of information systems and requirements integral with business process recon®guration and associated role evolution, initiating an integration of both social and technological development. RADs also contribute to addressing the issue of multiple stakeholders, for not only can they be represented in the models, but they can also be engaged jointly in their production. Several tools and techniques, as well as methodologies, have been cited in this special edition: some are new uses of existing techniques; others are extensions, or integrating with data-driven approaches such as O-O [7], prototyping [7], DFDs [7] and beyond to business process recon®guration. All of them exemplify the underpinning socio-technical/soft framework. For they take into account, and in some cases seek to address, the fact that in any organisational situation there are multiple stakeholders ± internal and external ± with multiple con¯icting agendas and that these will along with internal business processes and external environments change, often rapidly. Figure 2, using Galliers' taxonomy [1], delineates the domain formed by the contributions to this special edition. In so doing it maps out the soft/socio-technical paradigm, the tools and techniques within it and their relation with datamodelling approaches.

5. Concluding Discussion It was proposed in the introduction to this review that the contributions in this special edition represented what Avison and Fitzgerald [10] have called the `postmethodological' era. This they claim has arisen out of the limitations of the ISD approaches in the preceding `structured methodologies' era, which they see as being: in¯exible, over-rationalistic, data-driven, top-down, analyst-led, technique-®xated, process-focused, with a record of poor deliverability. They say of this new era,

`we expect that standard methodologies will decline in importance in the development of information systems within organisations, with the post-methodology era providing a greater focus on contingent and alternative approaches'. What distinguishes this new era is not that there will be an absence of methodologies or tools, but that what emerges will be approaches with a capacity to deal with contingency and diversity. Avison and Fitzgerald [10] see `contingency' approaches to ISD and IRE as having the following characteristics. Firstly, within them, the social dimension is to be embraced, not just as a palliative to the dominance of technical aspects in ISD but as being necessary to enable and achieve an integration of human and technological practices. Secondly, they are adaptable to and adapted by the circumstances of their use. Thirdly, they actively seek to merge with other disciplines, methodologies and realworld practices to address, where appropriate, the contingencies of the problem situation. Fourth and ®nally they encompass competing paradigms and perspectives. The answer to the question posed in the introduction as to the nature and status of the contributions in this special edition is therefore not dif®cult to discern. The previous discussions reveal their distinguishing features as being capable of accommodating multiple stakeholders, con¯icting interests, rapid change, sociotechnical/soft tools and techniques, a ¯exibility of approach, the integration of human and organisational with the technical and integratable with existing (datadriven) tools and techniques. From this description and the preceding analysis it can be argued that they exemplify the contingency approaches of Avison and Fitzgerald. In so doing they make manifest a sociotechnical/soft post-methodology era to information requirements, elicitation and systems development.

References 1. Galliers RD, Swan JA. There's more to information systems development than structured approaches: information requirements analysis as a socially mediated process. Req Eng 2000;5(2):74±82 2. Patel N. Healthcare modelling through role activity diagrams: for process based information systems development. Req Eng 2000;5(2):83±92 3. Al-Karaghouli W, AlShawi S, Fitzgerald G. (2000) Negotiating understanding information systems requirements: the use of set diagrams. Req Eng 2000;5(2):92±102 4. Coakes JM, Coakes EW. Speci®cations in contest: stakeholders, systems and modelling of con¯ict. Requirements Eng 2000;5(2):103±14 5. Sutton D. Linguistic problems and knowledge elicitation. Req Eng 2000;5(2)115±25 6. Mumford E. A socio-technical approach to systems design. Req Eng 2000;5(2):126±34 7. Avison DE, Fitzgerald G. Information systems development:

Information Requirements Elicitation in the Post-Methodology Era

8. 9. 10. 11. 12.

methodologies, techniques and tools. Blackwell Scienti®c, Oxford, 1995 Goodland M, Slater C. SSADM version 4: a practical approach. McGraw-Hill, London, 1995 Paul RJ. Why users cannot `get what they want'. Reprinted in ACM SIGIOS Bull 1993;14(2):8±12 Avison DE, Fitzgerald G. Information systems development. In: Currie WL, Galliers R (eds). Rethinking management information systems. Oxford University Press, Oxford, 1999 Checkland P, Scholes J. Soft systems methodology in action. Wiley, Chichester, 1990 Checkland P. Systems thinking systems practice. Wiley, Chichester, 1981

73

13. Checkland P, Holwell S. Information, systems and information system: making sense of the ®eld. Wiley, Chichester, 1998 14. Introna L. Management, information and power. Macmillan, London, 1997 15. Avison DE, Wood-Harper AT. Multiview: an exploration in information systems development. Alfred Waller, Henley on Thames, 1994 16. Atkinson CJ. Soft information systems and technologies methodology (SISTEM): a casestudy of the electronic patient recad, Req Eng 1997;2:pp 1±22 17. Atkinson CJ. The `soft information systems and technologies methodology' (SISTeM): an actor network contingency approach to integrated development. EJIS 2000;9:104±23