Observational Completeness on Abstract Interpretation

In the theory of abstract interpretation, a domain is complete when abstract computations are as precise as concrete computations. In addition to the standard notion of completeness, we introduce the concept of observational completeness. A domain is observationally complete for an observable π when abstract computations are as precise as concrete computations, if we only look at properties in π. We prove that continuity of state-transition functions ensures the existence of the least observationally complete domain and we provide a constructive characterization. We study the relationship between the least observationally complete domain and the complete shell. We provide sufficient conditions under which they coincide, and show several examples where they differ, included a detailed analysis of cellular automata.