• Summary
• Related Documents
  • Active Bibliography
  • Co-citation
• Version History

Abstract:

There appears to be a general consensus within the Information Systems literature that formal speci cation of software systems is an inappropriate response to the perceived general failure of information systems to meet user requirements. Such views would seem to be based primarily on the di culty of constructing formal speci cations { and on the di culty of understanding such speci cations once constructed. Research into the applicability of formal methods has therefore tended to concentrate on the needs and the context of software developers specialising in critical and extremely complex software such as operating systems, transaction processing monitors, or nuclear reactor protection. More recently, however, formal methods have been applied successfully in more conventional and commercial areas, such as the development of a CASE tool, indicating that many of the perceived disadvantages of formal methods may bemyths, or outdated. This paper discusses the di ering research directions of the Information Systems and Software Engineering disciplines and suggests that signi cant bene ts may result from a synthesis of the two approaches. We further suggest that there is a serious danger that approaches which have been shown to have value in one of the two domains are often ignored automatically in the other as being \irrelevant". While each of the two areas ignores the contribution of the other, software systems will continue to be sub-optimal (in terms of relevance, as well as quality).

Citations