Can Thomas Kuhn's paradigms help us understand software engineering?

Abstract

Recent articles in EJIS have discussed whether or not Information Systems is a ‘discipline’. In The Structure of Scientific Revolutions, Kuhn states that a scientific discipline can be identified by reference to its underlying belief system, the ‘paradigm’ or ‘disciplinary matrix’, to which all workers in that field must commit. An important element of Kuhn's model is the notion of ‘scientific communities’. We consider here the belief system underlying Software Engineering (SE). We examine the extent to which a belief system analogous to the disciplinary matrix of a Kuhnian science can be identified in SE. Our preliminary fieldwork has comprised an examination of books used by SE students and practitioners, and in-depth interviews with a number of practitioners. The results of this study suggest that the current status of the theory of SE parallels Kuhn's ‘pre-paradigm’ stage of scientific development. At this early stage, theorists and practitioners are divided into schools. These schools are based on differences in the beliefs and models forming their disciplinary matrices. We conclude that the application by analogy of Kuhn's view of scientific activity to SE is justifiable. Our findings can assist both SE theorists and practitioners in improving the understanding of how and why software development projects succeed or fail. Our findings also provide a framework within which to place the beliefs, models and values which underlie SE. Such a framework can contribute to the discussion as to whether the software development-related aspects of Information Systems can be considered to be a discipline, and if so how that discipline is structured.

Keywords:

• kuhn
• paradigm
• disciplinary matrix
• software engineering
• information systems development

Acknowledgements

The investigations reported here were funded by an Engineering and Physical Sciences Research Council (the then Science and Engineering Research Council) research studentship. Recent work by the authors has been supported by the University of Hertfordshire. We record our gratitude to Russel Winder for his help, advice and support during the work reported here. We also thank Rupert Read for comments before and during the ‘Interpretive’ Approaches to Information Systems and Computing Research workshop in July 2002 at Brunel University, West London, UK, and members of the Department of Computer Science at the University of Hertfordshire, in particular Chrystopher Nehaniv, for comments on drafts of this paper. We also thank the anonymous referees of EJIS for their helpful comments.

Additional information

Notes on contributors

Paul Wernick

About the authors

Paul Wernick is a Senior Lecturer in Computer Science at the University of Hertfordshire. After qualifying as a Chartered Accountant, he gained experience of commercial software development and use, as a computer audit specialist, analyst/programmer and user support consultant. His doctoral thesis examined from a Kuhnian perspective the belief systems underlying the theory and practice of software development. He also researched distributed systems security at the Computer Laboratory, University of Cambridge. Dr Wernick's current research interests include software evolution processes and their simulation, and the philosophical foundations of information systems and their development processes.

Tracy Hall

Tracy Hall leads the Systems and Software Research Group in the Department of Computer Science at the University of Hertfordshire. Tracy specialises in the empirical investigation of technical and non-technical issues within software engineering. She is active in the Empirical Software Engineering community and is regularly invited to talk both in the UK and abroad. Tracy is an accomplished researcher having been awarded various research grants and has published over 20 high-quality journal papers during the last 10 years. Tracy holds a PhD in Software Metrics from City University.