Abstract

In this paper, we present an overview of our recent research on programming language expressivity. The goal of this work is to make it possible for programs to clearly capture all of the important aspects of a system's behavior, including not only its functionality, but also issues such as its failure handling strategy, its communication strategy, its coordination strategy, its memory reference locality, etc. Our current work is based on the belief that programming languages based on any SINGLE abstraction framework---procedures, constraints, whatever---are ultimately inadequate for many complex systems. The reason is that the different aspects of a system's behavior that must be programmed, each tend to have their own "natural form", so while one abstraction framework might do a good job of capturing one aspect, it will do a less good job capturing others. This conclusion has led us to develop a concept we call Aspect-Oriented Programming (AOP). In AOP, the different aspects of a system's behavior are each programmed in their most natural form, and then these separate programs are woven together to produce executable code. Our work on AOP is being carried out in the context of both general-purpose and domainspecific languages, we believe that it has contributions to make to both areas.

Citations