Finding relevant research solutions for practical problems: the serp taxonomy architecture

Abstract

Background: Experience and research indicates that there exist a communication gap between research and industry in software engineering. Objective: We propose the Software Engineering Research and Practice (SERP) taxonomy architecture to support communication between practitioners and researchers. The taxonomy architecture provides a basis for classifying research from a problem perspective which in turn supports the breaking down of complex practical challenges to researchable units. Thus such taxonomy may support the mapping of challenges in industry to research solutions in the software engineering context. Method: In this paper we present SERP and exemplifies its usage based on two literature studies in the field of software engineering. Further, we discuss how a taxonomy based on this architecture could have helped us in two past research projects that were conducted in close collaboration with industry. Finally we validate SERP by applying it to the area of software testing, developing SERP-test, and interviewing two industry practitioners and two researchers. Results: The taxonomy architecture has been applied to two problems in software testing, and has been assessed through interviews with practitioners and researchers. The interviews provided suggestions of how to improve the taxonomy architecture, which have been incorporated. With two examples, we demonstrated how the taxonomy architecture could be used to find solutions for industrial problems, and to find the problems addressed by a particular solution. Conclusion: SERP may be useful in multiple ways: (1) Given that SERP taxonomies are populated with industrial problems and scientific solutions, we could rapidly identify candidate research solutions for industrial practice. (2) Researchers could benefit from the taxonomy in the reporting of their research to ease the mapping to industrial challenges.