Inconsistency Identification and Resolution in Goal-Driven Requirements Analysis.

Abstract

This thesis addresses the need for inconsistency identification and resolution techniques to assist practitioners employing goal-based requirements analysis. Several approaches for inconsistency identification exist, however, the techniques introduced in this thesis are targeted specifically for analyzing goals in conjunction with the Goal-Based Requirements Analysis Method (GBRAM). Reducing inconsistency in a requirements specification greatly impacts its quality, as well as the overall success of a software development effort. Addressing and reducing the number of inconsistencies before development begins, minimizes the need to continually revisit what needs to be developed (the requirements), and consequently reduces development time and costs since inconsistencies are resolved early on (prior to system design and implementation). This thesis introduces a series of techniques and associated heuristics for identifying and resolving inconsistency in requirements specifications. These techniques and heuristics are demonstrated in the analysis of two case studies, the second of which is further described in Deriving Goals from a Use Case Based Requirements Specification for an Electronic Commerce System [ADS00]. The main contributions of this thesis include techniques that augment the GBRAM which provide specific guidance to the identification of inconsistency, and a structured goal syntax to aid analysts in forming goals for future analysis, and heuristics to aids in inconsistency identification and resolution