How can one of the founding fathers of formal methods seemingly denounce the ?eld of research after over a quarter of a century of support? This is a question that has been posed recently by some formal methods skeptics. However, Prof.
Author: Jonathan P. Bowen
1 In a number of recent presentations – most notably at FME’96 –oneofthe foremost scientists in the ?eld of formal methods, C.A.R. Hoare,has highlighted the fact that formal methods are not the only technique for producing reliable software. This seems to have caused some controversy,not least amongst formal methods practitioners. How can one of the founding fathers of formal methods seemingly denounce the ?eld of research after over a quarter of a century of support? This is a question that has been posed recently by some formal methods skeptics. However, Prof. Hoare has not abandoned formal methods. He is reiterating, 2 albeitmoreradically,his1987view thatmorethanonetoolandnotationwillbe requiredinthepractical,industrialdevelopmentoflarge-scalecomplexcomputer systems; and not all of these tools and notations will be, or even need be, formal in nature. Formalmethods arenotasolution,butratheroneofaselectionoftechniques that have proven to be useful in the development of reliable complex systems, and to result in hardware and software systems that can be produced on-time and within a budget, while satisfying the stated requirements. After almostthree decades,the time has come to view formalmethods in the context of overall industrial-scale system development, and their relationship to othertechniquesandmethods.Weshouldnolongerconsidertheissueofwhether we are “pro-formal” or “anti-formal”, but rather the degree of formality (if any) that we need to support in system development. This is a goal of ZUM’98, the 11th International Conference of Z Users, held for the ?rst time within continental Europe in the city of Berlin, Germany.