A table reduction approach for software structure testing

Abstract

One major aspect of software structural testing emphasizes the coverage criterion. Not only procedure-oriented programs, but also member functions of object-oriented programs can be dealt with by structure testing. A program with higher degree of testing coverage indicates that the program might achieve better reliability. There is a conventional metric for testing coverage measurement. However, the metric is affected inadvertently by including undesirable entities. The measurement of coverage rate increases rapidly when the initial groups of test data are executed, which results in the overestimation of the software reliability. The goal of this research is to improve the conventional testing coverage metric, then to use it to precisely estimate the ratio of elements tested. In this paper, a set of rules is developed to identify dominant tested elements. Measuring the coverage rate of the dominant elements can rationalize the measurement results.