Among the early process algebras, Hoare’s communicating sequential processes (CSP) [1] had an explicit syntax of process names, which could be thought of as “agents” that executed the processes. The communication mechanism involved handshake synchronization among these names [2]. Milner’s calculus of communicating systems (CCS) [3] was modeled more on the lines of the λ-calculus used in programming languages, where names were used to refer to local variables, and thus even passed around, as in the later π-calculus (Milner’s Turing Award lecture [4] has an accessible account).

The authors of this paper seem to be suggesting that Hoare’s viewpoint might have had something going for it, because one can talk about the knowledge held by an agent and reformulate processes as games, discovering assumptions about information available to processes within the semantics. The authors apply this idea to security. They refer to some of the early agent-based work where knowledge theory plays a significant role (older [5,6] and more recent works [7] also exist). The authors’ connection also needs to be explored in the opposite direction, since bisimulation is a strong equivalence, to see whether any of the other process equivalences [8] can be given a knowledge theoretic reading.

This paper will be of interest to researchers in concurrency theory, programming language semantics, and knowledge representation.