A quantitative analysis for different testing criteria through program decomposition

Abstract

We present a quantitative comparison among different testing criteria. Testing criteria are methodologies that serve as the guidance for software testing. First, we survey various software testing criteria. We find that the quantitative difference among different criteria is a good guideline for selecting appropriate testing methodologies. Also, All-Statements and All-Branches testing criteria are two famous testing criteria that are widely applied in software testing. This paper infers a quantitative analysis to measure the difference between the All-Statements criterion and the All-Branches criterion. The quantitative analysis provides a theoretical basis for measuring testing efforts between different testing methodologies. A testing metric based on our analysis is then proposed to compare All-Statements and All-Branches criteria. An algorithm for this metric is also presented.