A reduction specification for DTL

Abstract

This paper describes techniques which are being used to specify the behaviour of programs written in DTL. The intended Semantics of DTL are similar in many respects to those of Hoare's CSP. However, DTL forces a hierarchical structure on a network of communicating processes and the language specification described capitalises on this to simplify the definition.

The principle technique employed in the DTL Semantic specification involves the generation of a set of reduction rules. Each rule specifies a class of behaviours for some DTL construct. An important aspect of these reduction rules is that they can be viewed as defining the behaviour of a reduction machine which would execute the specification of a particular DTL program. Such a reduction machine could be implemented by software on conventional machines or with much greater efficiency on the kind of recursive architecture made possible by VLSI technology.