A Synthesis of Coverage Testing and Modelling in Software Reliability Estimation

9313861 Singpurwalla软件可靠性评估的技术有很多。目前已有十多种软件可靠性评估模型。这些模型大多主要使用程序测试中获得的故障时间和故障数据。通常这些数据被用来估计模型的参数，然后这些参数被用来估计在一个暴露期间程序失败的概率。这个概率在文献中被广泛用作软件可靠性的度量。与黑盒模型相比，白盒模型需要了解程序结构。除了故障时间和故障数据外，本研究还试图开发和评估利用在模块和系统或子系统测试期间获得的代码结构和覆盖信息的白盒模型。＊＊＊

9313861 Singpurwalla Techniques for estimating software reliability abound. Over a dozen models exist for estimating software reliability. Most of these models primarily use the failure time and fault data obtained during program testing. Usually these data are used to estimate the parameter of the model, and the parameters are then used to estimate the probability of program failure during an exposure period. This probability is widely used in the literature as a measure of software reliability. In contrast to such black-box models, white-box models require a knowledge of the program structure. This research seeks to develop and evaluate white-box models that make use of code structure and coverage information obtained during module and system or sub-system testing, in addition to the failure time and fault data. ***