Automatic transactions identification in use cases

Abstract

Since the early 90's of the previous century, use cases have became informal industry standard for presenting functional requirements. The rapid popularity growth stimulated many different approaches for their presentation and writing styles. Unfortunately, this variability makes automatic processing of use cases very difficult. This problem might be mitigated by the use of transaction concept, which is defined as an atomic part of the use case scenario. In this paper we present approach to the automatic transaction discovery in the textual use cases, through the NLP analysis. The proposed solution was implemented as a prototype tool UCTD and preliminarily verified in a case study. © 2008 Springer-Verlag Berlin Heidelberg.