What lies behind requirements? A quality assessment of statement grounds in requirements elicitation

摘要

In requirements engineering (RE), an early yet critical activity consists in eliciting the requirements from various stakeholders, who usually have different assumptions, knowledge, and intentions. The goal during elicitation is to understand what stakeholders expect from a given software, expectations which then feed the analysis, prioritization, validation, and ultimately specification activities of the RE process. Elicitation is an interactive activity. It relies on verbal communication of statements of stakeholders about their requirements, their ideas, their assumptions, the constraints they know apply in the environment of the future software, and so forth. Statements, we claim, build either on a past experience of the stakeholder or are the result of reasoning from indirect experience, i.e., they have different grounds. In this paper, we introduce the concept of "Statement Ground" during RE, contrast it with the classical perspective on requirements elicitation, and position the concept in existing RE literature. We conduct an empirical assessment of the relative qualities of statements that have different grounds. Our work results in a better understanding of the statements produced by stakeholders during requirements elicitation, of their qualities, and of the interplay between those qualities and the concept of statement ground. It also results in the definition of a series of research questions which focus on the implications of our findings on the overall requirements engineering activity.