Probing Questions, Participatory Democracy, Quality Assurance, and Customer Data

Asking the Right Questions “Probing for Requirements Knowledge to Stimulate Architectural Thinking,” by Preethu Anish and her colleagues, describes an inexpensive, down-to-earth requirements elicitation technique to ensure that software requirement speci cations (SRSs) contain architecturally relevant information.1 This research aims to equip business analysts with the appropriate probing questions (PQs) software architects typically ask. This will enable the analysts to elicit and specify architecturally relevant information themselves. Using this approach, they can produce richer speci cations containing the details software architects need to make decisions. This research is necessary because SRSs often lack architecturally relevant details. Consequently, architects often either make assumptions, which can lead to incorrect decisions, or conduct additional stakeholder interviews, potentially delaying a project. Anish and her colleagues’ approach can reduce the necessary communication between architectures and customers. According to one architect who used it, “[In the absence of such an approach] the effort [to unearth architectural details] would be quite high, and more than that, user acceptance testing will raise questions as to why things were not clari ed earlier in the requirements phase.” Through interviews with 40 experienced architects, Anish and her colleagues identi ed reusable PQs for ve areas of functionality and organized them into structured ows. They trained a machine-learning algorithm to classify architecturally signi cant functional requirements. On the basis of this classi cation, they could determine whether a PQ is appropriate for a project. The authors evaluated their approach on more than 8,000 requirements from 114 industrial requirements speci cations and validated its usefulness. This research also produced a repository of PQs business analysts can reuse in other projects Probing Questions, Participatory Democracy, Quality Assurance, and Customer Data