Contextual natural language processing and DAML for understanding software requirements specifications

Abstract

In software engineering a system requirements document written in a natural language (NL) needs to be translated into one of the formal specification languages for system execution. When this translation is to be automated, resolution of the ambiguity in the document and explicit definition of implicit domain knowledge are necessary. In our approach, Contextual Natural Language Processing is used to overcome the ambiguity and the domain knowledge is expressed in DARPA Agent Markup Language (DAML). The result is a formal representation of the informal requirements in NL for prototyping and even for implementation.