Reasoning about Inconsistency in RE - Separating the Wheat from the Chaff

Inconsistency is a major challenge in requirements engineering. Traditionally, software requirements specifications (SRSs) are expected to be consistent, with the underlying assumption that this consistency is always achievable. However, with the growing complexity of software systems it has become clear that this assumption is not always realistic. This has led to new paradigms for inconsistency management, acknowledging that it is not only inevitable, but also even desirable at times, to tolerate inconsistency, even temporarily. However, for these paradigms to be widely accepted in industry, practicing software engineers must thoroughly understand the nature of inconsistency in SRSs and the strategies for its handling. This paper proposes a research agenda for preparing practicing software engineers to accept and successfully implement inconsistency management paradigms. As a first step in this direction, the paper describes an ongoing study in which we design an intervention into the perceptions of inconsistency for practicing software engineers. The intervention builds on teaching to them the Zave–Jackson requirements validation formula as an aid for analyzing the types of inconsistency they face, and conducting an empirical study of the effect of this intervention on their inconsistency management.