View from the Fringe of the Fringe

Formal analysis remains outside the mainstream of system design practice. Interactive methods and tools are regarded by some to be on the margin of useful research in this area. Although it may seem relatively academic to some, it is vital that this the so-called “theorem proving approach” continue to be as vigorously explored as approaches favoring highly automated reasoning. Design derivation, a term for design formalisms based on transformations and equivalence, represents just a small twig on the theorem-proving branch of formal system analysis. A perspective on current trends is presented from this remote outpost, including a review of the author’s work since the early 1980s. In memory of Dexter Jerry Johnson, August 12, 1918 – June 23, 2001 1 On Behalf of Interactive Reasoning “Formal methods for systems,” the application of automated symbolic reasoning to system design and analysis, remains well outside the mainstream of engineering practice, even though it is philosophically central to the science of computing (for instance, eight (by my count) of the forty-one ACM Turing Award recipients are formal methods luminaries explicitly recognized for this aspect of their work.) One could say, on the one hand, that formal methods research in software has significantly influenced practice. Type checking, structured and object-oriented programming languages, and other advances have been deeply assimilated. So, too, have equivalence checking and hardware description languages in hardware design practice. On the other hand, it is hard to tell whether, or to what extent, these tools have improved design methodology. Do more programmers use loop invariants? Does object oriented syntax improve the way engineers organize their thoughts, or does it merely impose structure, and if the latter, does that particular structure skew other aspects, such as communication and architecture? Let us focus on verification. In the past twenty years, the question of whether formal analysis is meaningful or practical has turned into a question of how and where they can be most profitably deployed. The digital hardware industry is ? Work described in this paper was supported by the National Science Foundation under grants MIP8707067, MIP8921842, MIP9208745, and MIP9610358.