What Do We Know about Alignment of Requirements Engineering and Software Testing?

Abstract

Context: The alignment of different software engineering activities for coordinated functioning and optimized product development is of great importance, particularly in industrial-scale development. The link between intermediate activities has been researched extensively, but the link between requirements engineering (RE) and software testing (ST) is a relatively less explored area. Objective: The objective of this study is to aggregate, structure, and classify all existing research regarding alignment of RE and ST published by the end of 2015. Method: We conducted a systematic mapping study (SMS) and aggregated all studies relevant to our scope. The primary studies are analyzed in terms of publication trend, focus area, i.e., how alignment is supported, the application domain and benefits and challenges, methodological data, and scientific rigor and industrial relevance. Results: There is a growing interest towards the topic. Several different techniques have been identified to improve RE and ST alignment. Test generation from requirements specification has received most attention. Alignment of RE and ST is particularly important for large safety-critical domains. While many challenges have been reported, the supporting evidence for benefits is scarce. Frameworks/methods/techniques is the most frequent contribution type. Solution proposal and evaluation research were the most frequently applied research type. Case study research was the most frequently applied research method, however, almost half of the studies did not clearly report any research method. Conclusion: Despite the numerous approaches that are proposed, it is not clear what approach is suitable in what context and why. To support industry in RE and ST alignment, guidelines and tool support are needed. The supporting evidence for claimed benefits is very limited. Overall, the research area is in its early stages and an increase in both the number and rigor of empirical studies are required.