The Principle of Comprehensive Modelling Applied to Data and Behaviour

Abstract

In their paper "Trace Semantics are Fully Abstract" Nain and Vardi propose that denotational semantics of behavioural modelling formalisms should adhere to the "Principle of Comprehensive Modelling" which requires models are exhaustive of all possible circumstances. However their exploration of this principle is made only in the context of machines that use sets of single symbols to define their state and alphabet. We extend the application of the principle to the denotational semantics of machines that use data structures, rather than single symbols, for the definition of state and alphabet. We show that, in the context of modelling the interaction between data and behaviour, this extension enables a unification of what are normally considered distinct styles of behavioural composition.