Reuse of Scenario Specifications Using an Automated Relational Learner: A Lightweight Approach

Abstract

With the growing popularity of object-oriented technology, specification by use cases has grown. Since use cases provide a bridge from the informal descriptions gathered from user interviews to the more formal software model, use cases are being considered a prominent part of software analysis. Yet, tools to assist the development of use cases are rare, especially, in practice. Our study aims to provide use case reuse, while relying only on the semantics of UML and the diagrams typically used by analysis. We present an automated technique that provides for the assisted reuse of use cases. A software analyst defines an initial description of a use case. An automated graph-based relational learner then retrieves a set of similar use cases from a database. The results of a preliminary application of our approach suggest that it can assist in the classification as well as retrieval of use cases.