Enabling Flexibility for Scoping Decisions in Software Development : A Case Study

Sammanfattning: Context. Flexibility in requirements management allows organizations to react quickly to changes in market demand. Organizations rely primarily on different documents to track projects, which are inefficient at maintaining traceability between requirements and other development artifacts. Finding a model to structure change impact analysis and change propagation is needed. Objectives. First, this thesis studies the current practice for enabling flexibility of scoping decisions. Second it evaluates the identified flexibility enablers – requirements centralization, requirements abstraction, and requirements grouping – in an industry context. The current practice is then contrasted with these flexibility enablers. Methods. To understand current practice, a case study was performed in industry. Data collection relied on interviews and document and repository observations, and grounded theory was used to analyze it. A solution containing the flexibility enablers was developed and evaluated at the organization where feedback on the tool was received. Results. The practitioners followed a strategy to handle changes in requirements where the impact analysis and decision making process were the most critical. Several challenges were identified which hindered flexibility for scoping decision. As a consequence, the studied project suffered from delay and budget overrun. The three presented enablers supported flexibility in decision making. The flexibility enablers facilitated decision making by providing a single, linked, consistent, and versioned set of requirements of multiple abstraction levels. Traceability between requirements was maintained which allowed propagation of changes between requirements. The multiple abstraction levels allowed for targeted viewing of the requirements based on role. Grouping requirements with implementation proposals provided traceability between the requirements and the design, which further helped change propagation. Conclusions. Although the studied organization followed standard procedures such as Change Control Boards (CCB) and change impact analysis (CIA) to enabling flexible decision making, they did not adopt any of the methods provided in recent studies. Additionally, the organization still depended on documentation to capture requirements and counted on their tacit knowledge to analyze the impact of changes. This hindered flexibility for scoping decision. Traceability as a strong point of centralization was aligned with previous studies in enabling flexibility while the family feature was a concept with no previous empirical evidence. We also proved that abstraction enabled flexibility, which has not been discussed in any previous studies. Similarly, we found no empirical evidence related to requirement grouping based on implementation proposals in enabling flexibility in decision making. Although the three enablers supported flexibility for scoping decision, the practitioners still had to invest a certain amount of manual work to benefit from them, since the tool did not handle everything automatically.