Empirical assessment of a defect detection technique for use cases
Beatriz Bernárdez · 2004
In this paper, an empirical assessment of a defects detection technique for use cases is presented. The technique consists of two main elements: a use case quality model and a set of metrics-based heuristics for predicting defects in use cases. The quality model is derived of an exhaustive analysis of existing quality models and of several requirements documents developed by Computing Engineering students at the University of Seville. We aim to build a detailed taxonomy of defects that were usually detected by the heuristics. The assessment of the heuristics is based on empirical data obtained from requirements documents that were verified by experts in requirements verification using a checklist based in our quality model. This empirical work has not only allowed us to corroborate that our refined quality model improves the capacity of the heuristics in predicting defects in use cases, but also has allowed to review and adjust some of the metrics thresholds values.