Information Systems Development and Data Modeling:
life cycle coverage; approach to uncertainty; methodology's Conceptual and Philosophical Foundations. Rudy Hirschdefined environment; respective objectives; and data modeling heim, Heinz K. Klein, and Kalle Lyytinen. Cambridge, UK: approach on the other axis. The four methodologies examined Cambridge University Press; 1995: 289 pp. Price: $44.95.
are: information systems planning, prototyping, soft systems (ISBN: 0-521-37369-7.) methodology, and the professional work practices approach.
The chapter summary includes two figures on the common and different aspects of the methodologies. These provide a clear Kirschheim, Klein, and Lyytinen have done a valuable sercompare and contrast of the alternative development apvice by tying together a series of related research lines into a proaches. single thread and-perhaps as importantly-all of the research The authors report several conclusions and consider two thread histories into a single volume. Synthesis of research from implications on IS research and practice. First, recognizing that more than 800 references provides a single integrated look at there is conflict as to the nature of IS research that has focused the manner in which information systems and data modeling on analysts specific assumptions about users and themselves. methodologies have evolved. The authors should feel proud of
The authors argue the danger of this approach because-unlike this significant integration accomplishment.
scientific disciplines-the research foci are conscious beings An extensive presentation of systems development methodand researchers have not found satisfactory ways of examining ologies stresses the importance of understanding the breadth of human community-based behavior with mathematical and natumaterial covered. My guess is that most IS researchers would ral science-based methods. Second, academic practices promote be hard pressed to name the seven types of information system functionalism, typically excluding alternative perspectives. The development methodologies as: formal live-cycle approaches, literature analysis indicates the vast majority of research has structured approaches, prototyping and evolutionary apbeen focused on fact-based approaches as per the functionalist proaches, sociotechnical participatory approaches, sense-makparadigm. They suggest perhaps more research should be foing and problem formulation approaches, trade union led apcused on the alternative approaches as a means of addressing proaches, and emancipatory approaches. A similar treatment of some of the deficiencies associated with the functionalist apthe evolution of data modeling draws parallels between the proach. Indeed, the author's note this is a situation where the evolution of systems development and data modeling methodolacademic community is falling behind more progressive indusogies. The key point is that both have ''been shaped by rival try practices that have understood the need and are moving concerns.'' Competing schools of thought have developed difbeyond strict, functional-based approaches to systems developfering approaches to both information systems development and ment and data modeling. As the authors state at the text's condata modeling. These schools differ greatly as to what constitute clusion: ''If our philosophical journey-metaphorically speakdevelopment activities and how should they be accomplisheding-to reach beyond functional has any merit, it is because leading to differing methodology goals and means.
many practitioners have also embarked on a journey arriving at The text presents a comprehensive treatment of the philothe same place, but they do not know where they are or how sophical foundations of systems development and data modeling they got there.'' methodologies. The authors note how ''systems development
While the text contents are quite valuable, I also found the assumptions can be reflected in developer behaviors and/or be text particularly frustrating in two ways. First, the text contains enforced by methodologies but that these assumptions certainly differing amounts of materials covering many of the topics. For affect the systems that are developed.'' They define two classes example, presenting 12 pages on the seven information system of assumptions: those about the way developer's come to acdevelopment methodologies made the presentation feel awkquire knowledge required to develop systems; and those relating wardly detailed with followed by 4 pages devoted to data modelto the developer's view of the social and technical worlds in ing evolution. This pattern is repeated in later chapters with which the system exists. Analysis of these assumptions yields large sections on some topics and much smaller section on four development paradigms-describing functionalist, social relativist, radical structuralist, and neohumanist information others. There is approximately three times as much discussion systems development. These are compared on a number of of information system development methodologies as there is points for both information systems development and data modon data modeling. While the uneven presentation is confusing eling.
to readers, it is consistent with the text's goals of illustrating The third topic area links the theoretical concepts to the the problematic nature of historical information systems develapproach specific assumptions derived by topic area two. Four opment practices and their study. development paradigms are detailed to permit more detailed
In defense of the authors, reducing the detail of the informaexamination of the each's affect on systems development. Four tion system development coverage would have been difficult methodologies are systematically evaluated, populating a matrix because it is already quite compact. The authors decision to with according to their respective: key actors; narrative; plot; present so much information system development materials can and assumptions on one axis and the techniques: analysis scope; be justified in the interest of research coverage completeness.
Extending the data modeling discussion would have clouded the points, and because data modeling hasn't the history of