Advice on structuring compilers and proving them correct

F. Lockwood Morris · 1973

The purpose of this paper is to advise an approach (and to support that advice by discussion of an example) towards achieving a goal first announced by John McCarthy: that compilers for higher-level programming languages should be made completely trustworthy by proving their correctness. The author believes that the compiler-correctness problem can be made much less general and better-structured than the unrestricted program-correctness problem; to do so will of course entail restricting what a compiler may be.

Read the paper · More papers on PaperTik