Combination of Analysis and Test
Quality assurance is an integral part of modern software development. High-quality software is often a must, particularly in the numerous software systems that surround us everywhere today, e.g., in the automotive, rail, and aviation domains. In such systems, which are often very comprehensive and complex, defects can have serious consequences (e.g., financial loss, loss of trust, or risk to human lives). Different quality assurance activities are applied to models or to code. However, quality assurance is often an expensive and time-consuming affair, and ultimately the quality of the software is often still bad in the end. All this means that new approaches are needed for mastering these challenges. An emerging trend is the increasing integration of various quality assurance techniques instead of the use of single applications in order to make it possible to exploit additional synergy effects such as higher effectiveness, efficiency, or coverage.
In the MBAT project, Fraunhofer IESE is working on a concrete possibility for integrating different analysis and testing activities. Based on the “In2Test” approach, which combines especially inspections and testing, concepts are being generalized and used to enable different possibilities of combining analyses and tests: For example, test results can be used to employ analyses in a systematic and complementary way, or these techniques can be used in parallel to enable coverage of more types of defects. In particular, Fraunhofer IESE is working on a solution that allows using statistical analyses for a more systematic performance of tests. In addition, various prototypes are used to develop a tool chain that allows building models for analyses and tests, and commercial and open-source tools are incorporated that enable performing analyses and tests. A tool prototype for visualizing a wide variety of quality assurance data complements the tool landscape. This prototype provides support for focusing tests by means of selection rules.
In addition to developing models and tools, Fraunhofer IESE is involved not only in the evaluation of the sketched solution, but also in the evaluation of the solutions of other partners in the MBAT project. Quantitative data as well as qualitative knowledge is collected and packaged by Fraunhofer IESE and used for deriving recommendations for a wide variety of combinations of analyses and tests, which also provide added value in the development and quality assurance of software in other contexts.