Retracing Some Paths in Process Algebra

The very existence of the concur conference bears witness to the fact that “concurrency theory” has developed into a subject unto itself, with substantially different emphases and techniques to those prominent elsewhere in the semantics of computation. Whatever the past merits of this separate development, it seems timely to look for some convergence and unification. In addressing these issues, I have found it instructive to trace some of the received ideas in concurrency back to their origins in the early 1970’s. In particular, I want to focus on a seminal paper by Robin Milner [Mil75]1, which led in a fairly direct line to his enormously influential work on ccs [Mil80, Mil89]. I will take (to the extreme) the liberty of of applying hindsight, and show how some different paths could have been taken, which, it can be argued, lead to a more unified approach to the semantics of computation, and moreover one which may be better suited to modelling today’s concurrent, object-oriented languages, and the type systems and logics required to support such languages.