Coinductive interpreters for process calculi

Abstract

This paper suggests functional programming languages with coinductive types as suitable devices for prototyping process calculi. The proposed approach is independent of any particular process calculus and makes explicit the different ingredients present in the design of any such calculi. In particular structural aspects of the underlying behaviour model (e.g. the dichotomies such as active vs reactive, deterministic vs nondeterministic) become clearly separated from the interaction structure which defines the synchronisation discipline. The approach is illustrated by the detailed development in Charity of an interpreter for a family of process languages.