Requirements Tracing Strategies for Change Impact Analysis and Re-Testing An Initial Tracing Activity Model and Industry Feasibility Study

Software requirements tracing is a systematic approach to keep track of development artifacts that need to be re-tested when requirements change. An important question is how much investment into tracing seems reasonable in an environment. This paper proposes an initial tracing activity model for trace generation and usage to help determine tracing efforts and support test managers in tailoring the expected tracing effort with the parameters: tracing strategies, test case creation effort, and change request severity. Using the tracing activity model we report a re-testing industry feasibility study that compares effort, risk, and delay of 3 tracing strategies: ad hoc trace reuse (ATR), no trace reuse (NTR), and systematic trace-based retesting (TBRT). Main results of the feasibility study were: 1. The tracing activity model was useful to model and compare overall efforts of tracing alternatives. 2. TBRT took a) around 77% effort of the NTR strategy and b) 40% effort of the ATR strategy effort with similar re-testing risk and less delay.