Abstract:

Most Software Reliability Growth Models (SRGMs) ignore either the learning phenomenon of the testing team or quality of debugging or both for the sake of simplicity. Some SRGMs consider imperfect debugging tightly integrated with the learning phenomenon. In practice, imperfect debugging and the learning phenomenon of the testing team are independent of each other since the former indicates quality of debugging process and the latter the quality of the testing process, usually carried out by different teams. Furthermore, sometimes the debugging team may find and correct some more faults in addition to the faults which caused the failures leading to what is known as efficient debugging. In this paper, a new mathematically derived SRGM is proposed, which also provides estimates of various quality metrics such as the number of faults in the software system, debugging and the learning indices. The proposed SRGM seems to describe the time to failure data of many software projects satisfactorily.