Resilient process theory (RsPT)

Abstract

This short paper is intended to highlight the focus and direction of my research work to date. Its purpose is to illustrate the course of future studies that I intend to undertake during my last year of research, leading up to the submission of my Ph.D thesis in September 1997. My research has concentrated upon the modelling of object-orientated communicating processes using process algebras. Behavioural reuse and the modification of process behaviour within an environment have been central to my research so far. I have indentified certain shortfalls in the ability of current process algebras to model a system based on objects. Modelling inheritance between processes and maintaining the stability of communications between those processed has proved difficult given the facilities of existing process algebras. Process substitution and modification (via inheritance) can introduce deadlock into a previously stable system. I propose a variant of process algebras that use an asynchronous communications model (known as Resilient Process Theory-RsPT). My theory permits the modelling of a stable object-oriented communicating system which can allow the behaviour of processes (viewed as objects) to be reused and extended. Processes in my proposed formal model are less likely to fail if synchronising communications do not occur. A resilient process can continue to execute provided that inputs to the process are still available; outputs do not effect a process's ability to execute as these are place in a buffer of infinite size.