Defects in automotive use cases

Abstract

This paper presents an empirical quality assessment of use cases with the purpose to provide defect data from industry. In the assessment, twelve criteria, based on earlier research, were applied to 43 use cases from Volvo Car Corporation developed according to current practice. The collected defect data were statistically analyzed to determine significant intensity differences and to establish a partial order between the defects types based on their intensity. In addition, a qualitative assessment was made to assess the effect of remaining defects on the established order. The study shows that the defect types with the highest defect intensity are Missing element and Incorrect linguistics. Further, the established defect order is the basis for a comparison between the criteria used and earlier research on guidelines and checklists for use case authoring, resulting in improvement propositions.