The History of Software Engineering

In August 1996 about a dozen historians met with about a dozen computer scientists to discuss the history of software engineering. The term software engineering has been deliberately chosen as being provocative at the 1968 NATO Conference on Software Engineering. This notion was meant to imply that software manufacture should be based on the types of theoretical foundations and practical disciplines that are established in the traditional branches of engineering. This need was motivated by the so-called software crisis. Ever since, the debate whether such a crisis exists has continued within the software engineering community. It is a crucial question, because if the answer is yes, software engineering may not be called an engineering discipline yet. If the answer were no, the question would be, what is it that constitutes this discipline.

It turned out at the seminar that there may or may not be a software crisis, but there is definitely what might be called an identity crisis. A strong indicator for this phenomenon is the fact that after more than 30 years computer scientists are investigating the history of other established branches of engineering to find out (or to define?) what should be done to turn software engineering into a sound engineering discipline. In this endeavor, historians were regarded to be some kind of universal problem solvers who were called in whenever a general answer to some fundamental question was needed.

Of course, this could not work, because history is not a methodical vehicle to clarify ones own identity or to derive normative principles and guidelines for a discipline.

Furthermore, there is only little historic knowledge in the field of software engineering as compared to the ``History of Programming Languages'', for instance, or the history of electronic calculating devices. Thus, a Dagstuhl seminar on the ``History of Software Engineering'' can only act as a starting point, providing (a) a first overview of what has been accomplished so far and (b) identify crucial issues to be studied in the future.

With respect to my own expectations about the possible outcome, I have to admit that I was too optimistic when I took the first initiative for this seminar. I underestimated the personal and disciplinary identity problems and I was expecting more historical studies already having been carried out in the field. As a consequence, this seminar report does not provide the reader with a coherent and concise account of what constitutes the history of software engineering. It contains, however, some of the many facets, issues and approaches to the field which have made this Dagstuhl seminar a very stimulating event and which may serve as a starting point for further investigations. More than before, I am convinced that studying the history of software engineering is a necessary and rewarding activity especially also for young scholars in the field. I want to thank the people from Schloß Dagstuhl for providing such a stimulating environment. Thanks also to Bill Aspray and David Parnas who organized the seminar together with me, and finally many thanks to Andreas Brennecke who took the burden to prepare this seminar report.

Paderborn, July 1997 Reinhard Keil-Slawik