Defects Detection Technique of Use Case Views during Requirements Engineering

Abstract

In the past decade, object-oriented software engineering (OOSE) has gained popularity from many software developers, especially OOSE with a unified modeling language (UML). Use case views are often used in most systems during an analysis phase. These views show the system functionality related to the system stakeholders. Hence, use case views seem to be a corner stone for a software system. The defects occurring in use case views will affect the later designs. If these defects can be found early, it would save time and cost in software development. Therefore, this research proposes a technique for detecting defects in use case views during an analysis phase or requirements engineering process. Correct users' requirements were created as requirements specifications in a traditional form-based style for testing comparison. The algorithms for generating decision table from form-based requirements and UML use case specification as well as the algorithm for use case view validation were invented. Two simple case studies were investigated and used as the preliminary evaluation. The nineteen fourth-year students were selected as the subjects for the preliminary investigation in order to compare between manual fault detection and our automated proposed system. They were asked to perform four tasks for each case study. The results show that the efficiency of manual fault detection is less than the proposed technique at 47% and 44% or overall average as 45.5% for the two case studies. Currently, we apply this proposed technique to more complex industrial setting and familiar software systems to software engineers.