Abstract

A major barrier to more common use of mechanical theorem provers in verifying software designs is the significant distance between proof styles natural to humans and proof styles supported by mechanical provers. To make mechanical provers useful to software designers with some mathematical sophistication but without expertise in mechanical provers, the distance between hand proofs and their mechanized versions must be reduced. To achieve this, we are developing a mechanical prover called TAME on top of PVS. TAME is designed to process proof steps that resemble in style and size the typical steps in hand proofs. TAME's support of more natural proof steps should not only facilitate mechanized checking of hand proofs, but in addition should provide assurance that theorems proved mechanically are true for the reasons expected and also provide a basis for conceptual level feedback when a mechanized proof fails. While infeasible for all applications, designing a prover that can process a set of high-level, natural proof steps for restricted domains should be achievable. In developing TAME, we have had moderate success in defining specialized proof strategies to validate hand proofs of properties of Lynch-Vaandrager timed automata. This paper reports on our successes, the services provided by PVS that support these successes, and some desired enhancements to PVS that would permit us to improve and extend TAME.KeywordsState InvariantInduction StrategyInduction ProofProof StepAbstract Data TypeThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call