Ronald K. Stamper

New directions for systems analysis and design

Information systems analysis and design is stagnant. It exists largely as an adjunct to software engineering to facilitate the application of computers. Its fundamental ideas, which have scarcely advanced since the 1950s, are based a paradigm that makes us think with a technical bias. This paper proposes a new perspective that can make ISAD more productive, both practically and intellectually.

Information Systems as a Social Science: An Alternative to the FRISCO Formalism

The FRISCO Report has initiated an important debate on the discipline of information. Chapters 1, 2 and 6 contribute ideas for all shades of opinion, a Broad View contrasting with Chapters 3, 4 and 5, “the core of the report”. This paper presents a Social View as an alternative to this Core View, which it criticises in two ways. Firstly, it limits the scope of information systems by treating it as an adjunct to software engineering, having no place for key properties of information in organisations and society. Secondly, it does not satisfactorily link its framework of concepts to the empirical world, but relies on the mentalistic notions of ‘perceptions’ and ‘conceptions’ in the mind of some, unspecified interpreter. The strength of the Core View lies in its formal precision. But a Social View can lead to an alternative formal framework that can place a much wider range of information systems concepts on a firm empirical basis. Thus provides a foundation for information systems as a social science, rather than a branch of applied mathematics. The Core View includes the belief in an objective reality to which we have direct access via the supposedly transparent languages of words, numbers and diagrams, whereas the Social View, without rejecting that position, requires us to explain how we come to construct such knowledge. It also compels us to examine all the other functions we perform with information, especially our construction of social reality. Forced to examine these questions, we have to address many elusive problems, such as the creation of meanings, the role of intentional communication, the construction of time, and the systems of norms we call ‘organisation’. Finally, the paper re-works the Japan Wines case study used in the FRISCO Report. This shows that the Social View leads to a more detailed analysis of the business problems than the Core View, with as much formal precision, which lends itself to computer interpretation, while also yielding models that are easier for users to interpret.

From database to normbase

After the database concept, we are ready for the normbase concept. The object is to decouple organizational and technical knowledge that are now mixed inextricably together in the application programs we write today. The underlying principle is to find a way of specifying a social system as a system of norms. Our existing languages, developed to handle machine-like structures, do not enable us to embed our technical systems comfortably into the far more subtle human systems they should be serving. The present approach to the design of formal systems incorporates a philosophical position that obstructs ones thinking about the relationship between social systems and technical systems. The existing specification languages embody a view of the world as an objective reality. But social systems are constructed by their participants, so we propose a language that treats the world as essentially subjective. Our way of doing so introduces two fundamental postulates: (1) there is no reality without an agent, and (2) the agent only knows the world through actions. Based on these postulates, a formalism has been created that enables us to represent systems of social norms. In this system, meaning is regarded as a relationship between sign and behaviour (more strictly, invariants in the flux of behaviour). Semantic analysis is an essential prelude to norm analysis. A new prototype implementation of this normbase is under construction. The goal envisaged is explained in terms of an illustration of how the normbase would be used to develop a system. At every stage of design, the specification is turned immediately into a default version of a working system. The application programmer can then concentrate on tuning this system or developing another that can perform the business functions in the most efficient technical manner. Experiments have already shown the viability of the concepts and methods incorporated in this normbase.

Enterprise information systems: issues, challenges and viewpoints

This paper is based on a panel presentation to the first International Conference on Enterprise Information Systems (EIS). The panel intends to address important issues in EIS from the perspectives of research and industry. It has not only raised questions, concerns and topics related to the research, development and application of information technologies and systems, but also suggested approaches and solutions for the researchers and practitioners. This paper covers a wide range of topics, such as: the notion of enterprise information system; the scope, features, functions and components of EIS; strategic importance of an enterprise information system in industry in the twenty-first century; the role and contribution of the research scientists, engineers and practitioners; factors affecting successful deployment and development of EIS in both organisational and technological aspects; the state-of-the-art enabling technologies for EIS; and the future of EIS in research and industry. The paper consists of sections contributed by the panel members, each addressing several issues from a viewpoint.

An information systems profession to meet the challenge of the 2000s

Technical proficiency has not enabled our profession to deliver value for money in terms of improved organisational performance. Evidence of this failure is adduced. The recommended change is to approach organisations as the information systems we develop. To achieve this we need a better understanding of information, the key resource, better ways of modelling organisations in information terms, and new tools for analysis, design, and systems development. Semiotics, the theory of sign, provides a suitable framework in which to make “information” and related concepts more precise. A semiotics of organisations leads to modelling them as systems of social norms from which information requirements can be logically deduced. This approach yields methods and tools for analyzing and designing the social, pragmatic, and semantic aspects of information systems that receive little attention in our current methodologies.

Measurement or the semantics of numerical information

It is unwise, though common practice, to base decisions on numerical values without knowing how they can be justified as correctly describing aspects of reality. Scientists have great responsibility in this regard, especially in the social sciences where it is easy to provide an appearance of rigour through the use of numbers and mathematical apparatus. We must be prepared to justify any numbers we use by clarifying the relationship between a numerical variable and a “real thing”. The article defines fundamental, explicit, scientific measurements and how they may be used in a meaningful way. It then considers instruments, in particular pointer measurements, that simulate these fundamental measurements. It concludes with a look at the validity of the resulting information.

Position paper on How far harmonisation?: Information system modelling myopia

In this position statement on the question, How far harmonisation? I expresses the view that, in the context of our profession, our theories, as far a possible, must meet the stringent quality criteria which philosophers of science impose on other disciplines that deal with problems in domains of inter-subjective activity, such as the natural sciences. I also show that our present ‘tunes’ are played on almost one note, making any harmony rather thin and uninteresting. Sometimes I despair of our profession ever abandoning old, demonstrably weak ideas (conceptualist or naive objectivist ontologies, for example) or of ever raising its eyes above the technical layer of problems to face the other half of the issues before us. However, I justify my long-term optimism because I have not encountered any intrinsic difficulty in meeting the most exacting criteria of scientific quality in dealing with information systems. Only a general unwillingness to apply those criteria obstructs our progress.

Comments on “An Organizational Theory of Information” by A.W. Holt and F. Cardone

This paper is rather tough reading. Like FRISCO Chapter 4, it is dense, abstract and somewhat formal. The reader is advised to populate its abstractions with concrete instances. It would also be interesting to visit the corpus of raw material that inspired this approach to information systems. The authors have an abundance of varied practical experience to draw upon. One can often appreciate a theory better when its roots are visible, at least this guides one’s choice of concrete illustrations.

Comments on Semiotic approach for object abstraction by B. A. Calway

Object-oriented systems tend to impose a naive objectivist view of reality but, because an interpreter is always involved in relating sign to reality whenever we are using information or analysing systems, it is a gross over-simplification. A subjectivist approach is more accurate an it brings to 00-analysis and design the benefit of a canonical object structure.

Comments on Theories of meaning and different perspectives on information systems by P. Holm and K. Karlgren

Every semantic theory presupposes an ontology, or a theory about what exists. By uncovering the ontologies behind the various semantic theories used by people working on information systems, we can learn much about the limitations of their methods of analysis and specification.