Comparing inspection strategies for software requirement specifications

Abstract

This paper reports on a laboratory experiment on the use of decomposition strategies in software requirements inspections. The experiment follows on from the work of (Porter et al., 1995) who compared the use of scenarios with ad hoc and checklist techniques, finding (among other things) that the scenario technique was superior. This experiment compares the scenario technique with inspection strategies which are self-set by the inspection team prior to the inspection but after they have seen the documents to be inspected. The specification used was a system developed by a software company for a client in the commercial sector. It was found that the commercial scenarios developed for the experiment were not superior to self-set strategies. This suggests that the benefits to be derived from scenarios are derived through the decomposition process and that experienced people may be able to derive strategies that are at least as good, if not better, than a provided set of scenarios. An advantage we noticed with the provided scenarios was the manner in which this technique could be used to focus the reviewers' attention on particular defect types. This could be used to advantage in industry. The overall findings of this experiment supports and extends the earlier research on inspections.