Unique fixpoint induction for value-passing processes

Abstract

We investigate the use of unique fixpoint induction as a proof method for value-passing process languages with recursion. An intuitive generalisation of unique fixpoint induction based on loop invariants for symbolic graphs yields strong completeness results; we give an axiomatic characterisation of both late and early observational congruence for a class of fully parameterised processes. This new, generalised, rule is shown to be derivable from existing formulations of unique fixpoint induction for value-passing calculi, thereby providing original completeness results. An example of the use of this new rule is presented in detail.