A Loop-detecting Interpreter for Lazy, Higher-order Programs

Abstract

Interpreters that detect some forms of non-termination have a variety of applications, from abstract interpretation to partial evaluation. A simple and often used strategy is to test for a repeated state, but this cannot handle infinite values (such as first-class functions) or unevaluated states such as arise in lazy programs. In this paper we propose using Berry and Curien’s theory of sequential algorithms as a semantic foundation for loop detection: we obtain straightforwardly a loop detector for lazy higher-order functional programs, which is more effective than the simple strategy even for strict first order cases.