Abstract

In this paper, we show that one can “deep-embed” the Java bytecode language, a fairly complicated language with a rich semantics, into the first order logic of ACL2 by modeling a realistic JVM. We show that with proper support from a semi-automatic theorem prover in that logic, one can reason about the correctness of Java programs. This reasoning can be done in a direct and intuitive way without incurring the extra burden that has often been associated with hand proofs, or proofs that make use of less automated proof assistance. We present proofs for two simple Java programs as a showcase.

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