Analytic methods in coverage testing of communications software
Jadranka Alilovic-Curgus · Conference of the Centre for Advanced Studies on Collaborative Research · 1992
Efficient test selection and evaluation algorithms are important in practical situations during the system verification phase, especially when the software involved is complex and manual and semi-automated methods of system verification can require weeks. To help automate this process, a simple procedure using a stratified selection approach for the control space of system behaviour is presented. This new method introduces a rigorous mathematical treatment of the discipline of testing, through the definition of testing distance, test coverage metrics, and a metric-based test selection algorithm. The test coverage metric is based on the concept of testing distance between execution sequences, which captures approximations of different patterns of system behaviour, and encompasses recursion and parallelism. The metric is incorporated into a test selection algorithm that allows generation of e-dense sets of test sequences that approximate some original test suite to some target accuracy of distance e in a convergent manner. This method is intended to facilitate the testing of communication software and other systems that involve many concurrent and highly recursive processes of largely unvisualisable character.