Test coverage dependent software reliability estimation by the HGD model

The hyper-geometric distribution model (HGDM) has been presented as a software reliability growth model with the capability to make estimations for various kinds of real observed test-and-debug data. With the HGDM, the discovery and rediscovery of faults during testing and debugging has been discussed. One of the parameters, the `ease of test´ function w(i) represents the total number of faults discovered and rediscovered at a test instance. In this paper, the authors firstly show that the ease of test function w(i) can be expressed as a function of the test coverage for the software under test. Test coverage represents a measure of how much of the software has been tested during a test run. Furthermore, the ease of test function w(i) can integrate a user-experience based parameter c that represents the strength of test cases to discover faults. This parameter c allows the integration of information on the goodness of test cases into the estimation process. The application of the HGDM to a set of real observed data clearly shows that the test coverage measure can be integrated directly into the ease of test function w(i) of the model