Proceedings of the European Computing Conference: Volume 2

Nikos Mastorakis · Valeri Mladenov · Vassiliki T. Kontargyri

mars 2010 · Lecture Notes in Electrical Engineering Livre 28 · Springer Science & Business Media

À propos de cet e-book

Software testing has been considered so important that organizations can assign teams only for testing activities. Testing is an important activity to ensure software quality. Tests are usually run several times to certify that code ma- tenance did not accidentally insert defects into working parts of the software. In such situations, test teams shall be able to estimate the required effort to execute test cases in its schedules and to request more resources or negotiate deadlines when necessary. When regarding model-based testing approaches that we use in our documented Integrated and Optimized Software Testing methodology (IOSTP) [1–4], a high number of test cases can be automatically generated. As team resources are limited, it may be not practical to execute all generated test cases. Their complexity usually determines the effort required to execute them and it can be used for planning test resources and test suites. Several software development estimation models have been proposed over the years. However, these models do not estimate the effort for executing a given test suite, since their estimations are based on software development complexity instead of its test planning, test case design and test execution complexity. According to our reading of the literature (e. g. [5–14]), ‘‘best practices’’ in model-based effort estimation include: Local calibration (or LC); i. e. using local data to set two special tuning parameters; Stratification; i. e. given a database of past projects, and a current project to be estimated, restrict local calibration to just those records from similar projects.