Automata theory presents roughly three types of automata: finite automata, pushdown automata and Turing machines. The automata are treated as language acceptors, and the expressiveness of the automata models are considered modulo language equivalence. This notion of equivalence is arguably too coarse to satisfactorily deal with a notion of interaction that is fundamental to contemporary computing. In this paper we therefore reconsider the automaton models from automata theory modulo branching bisimilarity, a well-known behavioral equivalence from process theory that has proved to be able to satisfactorily deal with interaction. We investigate to what extent some standard results from automata theory are still valid if branching bisimilarity is adopted as the preferred equivalence.
Jos C. M. Baeten, Pieter J. L. Cuijpers, Bas Lutti