Supporting the Combined Selection of Model-Based Testing Techniques

Abstract

The technical literature on model-based testing (MBT) offers us several techniques with different characteristics and goals. Contemporary software projects usually need to make use of different software testing techniques. However, a lack of empirical information regarding their scalability and effectiveness is observed. It makes their application difficult in real projects, increasing the technical difficulties to combine two or more MBT techniques for the same software project. In addition, current software testing selection approaches offer limited support for the combined selection of techniques. Therefore, this paper describes the conception and evaluation of an approach aimed at supporting the combined selection of MBT techniques for software projects. It consists of an evidence-based body of knowledge with 219 MBT techniques and their corresponding characteristics and a selection process that provides indicators on the level of adequacy (impact indicator) amongst MBT techniques and software projects characteristics. Results from the data analysis indicate it contributes to improve the effectiveness and efficiency of the selection process when compared to another selection approach available in the technical literature. Aiming at facilitating its use, a computerized infrastructure, evaluated into an industrial context and evolved to implement all the facilities needed to support such selection approach, is presented.