This paper is concerned with one of the basic problems in abstract interpretation, namely, for a given abstraction and a given set of concrete transformers (that express the concrete semantics of a program), how does one create the associated abstract transformers? We develop a new methodology for addressing this problem, based on a syntactically restricted language for expressing concrete transformers. We use this methodology to produce best abstract transformers for abstractions of many important data structures.

(Click here to access the paper: PDF; (c) Springer-Verlag.)