ADA AND NIL: A COMPARISON OF TWO MODELS OF STRUCTURING SOFTWARE SYSTEMS.

Abstract

This paper compares two language design approaches for development of large software systems. One is an approach based on static inter-module binding, block structure, and abstract data types, which has been extended to support concurrency, and is exemplified by Ada. The second approach is based upon distributed processes, communication by message passing over ports, no shared data, and dynamic binding, as exemplified by IBM's NIL. We argue that the process model better meets the needs of large system development, and yields a simpler language with a smaller number of fundamental concepts.