Quality Improvement of Requirements Specification via Automatically Created Object Oriented Models

Abstract

In industry most software requirements specifications are written in natural language. Software analysts prefer natural language over formal languages because it increases the communication between all stakeholders. However, the downside is that natural language specifications are often imprecise and ambiguous. This paper explores whether natural language processing can help to reduce ambiguity and to increase precision in natural language specifications. To examine the possibilities of this approach, a CASE tool was implemented. The developed approach uses syntactical knowledge to build an object oriented analysis model. Since the automatically extracted diagram is smaller than the original specification and clearly visualized, it can enhance cognition. Therefore, a human can more easily find ambiguities tracing the defects back to the original specification. A case study illustrates and evaluates the developed approach.