A tagging-based approach for eliciting requirements in established domains
Richard N. Taylor, Leyna C. Cotran · 2013
Many traditional requirements engineering and elicitation techniques assume that the requirements being developed are high-level requirements for a new, green-field system. However, this is not always the case. In established domains, such as the aerospace domain, systems are often developed as new members of existing product lines. In these cases, a significant amount of architectural design knowledge already exists within the product line, and the need is for detailed engineering requirements rather than high-level requirements. To address this situation, we present the Co-Evolving Traceable Requirements and Architecture Network (COTRAN), a novel collaborative requirements engineering technique that elicits detailed engineering requirements for new product line members through tagging architecture artifacts. Two experience studies have evaluated the COTRAN technique. The first was an industry experience in the aerospace domain. The objective of this industry experience was to apply the COTRAN technique to develop detailed engineering requirements in a relevant target field, on an actual live project, and observe the effects of verifiability and traceability of the requirements for the project. The industry experience demonstrated that COTRAN enabled mapping of requirements to the project's artifacts, demonstrating traceability from requirements to architecture. Also, the industry study demonstrated measurable improvements in verifiability of these requirements, through defining success criteria of the requirements that were guided by tagging architecture artifacts. The second was a classroom experience at a large northern California university. The COTRAN technique was incorporated in the curriculum of a graduate course in Requirements Engineering for the university's software engineering Master's program. This classroom experience demonstrated that COTRAN enabled mapping of resulting requirements with the students' architecture artifacts, demonstrating traceability from requirements to architecture and design decisions. The classroom experience also resulted in the added benefit of demonstrating that COTRAN can be taught in the classroom, and carried out by the students, providing insight in teaching COTRAN. Contributions of this research are a novel technique that supports the elicitation of requirements in the presence of rich domain and design knowledge, including reference architectures, and a demonstration that tagging can be a traceability mechanism to connect requirements and architecture.