Sampling in Software Development
Per Runeson · Wiley StatsRef: Statistics Reference Online · 2014
Abstract In software development, sampling inspection is applied during the development process, as the production process for software is trivial. Insoftware inspection, which is a manual scrutiny of software code or specifications, sampling is applied to estimate the number of remaining faults and to guide which parts of the artifact to inspect. Capture–recapture techniques are used in the fault content estimation. Insoftware testing, which is the dynamic execution of code, sampling is implicitly or explicitly applied to select test cases. Since the number of states is extremely high in software products, a complete coverage of all cases is practically impossible. In addition to scoping the tests to a reasonable magnitude of size, sampling is used to estimate the number of remaining faults and to predict the operational failure behavior of the software, e.g. through software reliability growth models. The theory is quite well established in the field, while the industrial practice is underdeveloped.