Investigating the temporal behavior of defect detection in software inspection and inspection-based testing

Abstract

A major goal of analytical quality assurance (QA) activities, e.g., inspection and testing, is detecting defects in software artifacts to increase product quality and decrease rework effort and cost. Inspection aims at identifying defects early and traditional testing focuses on test case generation and execution late in the development process. Combining inspection and test-case generation to inspection-based testing (UBT-i) can help identifying defects early, increasing testability by systematically capturing requirements and quality attributes, and generating most valuable test cases based on inspection results. This paper reports on a controlled experiment to investigate the temporal behavior of UBR inspection and inspection-based testing regarding defect detection performance, i.e., effectiveness, efficiency, and false positives. Main findings of the study are that there are no significant advantages of UBR and UBT-i regarding defect detection performance and the temporal behavior of defect detection delivered contradictory results in two sessions of the study. © 2010 Springer-Verlag.