Senior lecturer in the Software Engineering Research Group (SERG)
My research area is software engineering in the industrial context with a special focus on decision support in the testing processes. Current strategies for large scale software development are adapted to meet the demands for frequent and customized releases of software systems. This have major impact on test management since they force much repetitive testing of similar software in both time and space. As software systems grow in size and complexity the combinatorial explosion of test possibilities makes it infeasible to make informed test scoping decisions without tool support. My research interests regard the challenges faced by developers, testers and test managers within this context.
Several factors complicate the utilization of gained knowledge regarding testing in industry practice. Test maturity in an organization may be generally low, leading to short sighted test planning with too short lead times for test. A lack of testing resources in combination with short lead times prevents both the use of systematic approaches and improvement work itself.
Even with a more mature test organization and dedicated resources for test improvements it is hard for a practitioner to identify relevant and applicable research. Many proposed testing techniques solve a very specific problem in a similarly specific context. For a tester operating in a complex industrial context it might not make sense to isolate a point of improvement to match those proposals. While practitioners tend to describe their challenges from a problem perspective at high abstraction levels, researchers tend to report and compare solution proposals and empirical results from a solution perspective and at a low abstraction level.
To address the challenges of miscommunication between practitioners and researchers in the field, we have constructed and evaluated a taxonomy, SERP-test (Software Engineering Research and Practice taxonomy of testing), aiming at this interface. SERP-test was constructed through a systematic and goal oriented approach which included literature reviews and interviews with practitioners and researchers. It was further evaluated through an online survey and by utilizing it in an industry-academia collaboration project.
To promote the use of the taxonomy and enable community-based evolution of the same we are currently developing a web-based tool, SERP-connect, for classifying, matching and exploring collections of research results and practical challenges based on SERP-test.
EDAF25 - Object-oriented Modelling and Design