Model Based Testing in Software Product Lines

Overview and Motivation Due to the rising demand for individualised software products and software-intensive systems (e.g., mobile phone or automotive software), organizations are faced with the challenge to provide a diversity of software systems at low costs, in short time, and with high quality. Software product line engineering is the approach for tackling this challenge and has proven its effectiveness in numerous industrial success stories, including Siemens, ABB, Boeing, Hewlett-Packard, Philips, and Bosch [Pohl et al. 2005]. Software product line engineering differs from the development of single systems in two key aspects: ! The development process is split into two interacting processes: During the domain engineering process, reusable domain artefacts are created. In the application engineering process, the domain artefacts are reused to build individual applications, where the initial effort for creating the domain artefacts pays off by means of lesser development effort as well as shorter time-tomarket. ! The variability of the product line is explicitly modelled: The commonality and the variability of the product line is explicitly defined during domain engineering and is specified in all domain artefacts. The variability in the domain artefacts is then bound in application engineering by selecting application specific variants. Thereby, specific product line applications are derived from the domain artefacts. Testing plays an important role during software product line engineering, because the quality of the reusable domain artefacts affects all applications that are derived from these artefacts. Because of the above differences between software product line engineering and the development of single systems, software product line testing faces specific challenges that cannot be tackled by “traditional” testing techniques. Therefore, testing techniques from the development of single systems