Casre -an Easy to Use Software Reliability Measurement Tooi

Many different software reliability model.s have been proposed since the publications of the first mc)dels in 1971 [3,2]. Sufficient work has been done to demonstrate that these methods can be used to monitor reliability growth over a useful range of software development projects. }Iowever, the use of software reliability measurement as a management tool is not as widespread as it might be. Part of the problem is that there is no known method for identifying the model most applicable to a development effort. prior to test [2]. !l’his can be mitigated, however, by executing several models at once and using statistical. methods to identify the model that is most likely to produce accurate results [1,2]. Recent work has also shown that more accurate predictions can be produced by combining the results of several models in a linear fashion [3] or by recalibrating a model based on its predictive bias [4]. Another major issue in software reliability measurement is the ease-of-use of currently available tools. Several tools are currently available [5], but many of them have properties making them difficult for non-specialists to use. For instance, the primary outputs of several tools are in the form of model. parameter values. Without knowledge of the mathematical details of the model s, the results cannot be interpreted. Although software developers and managers may have a general knowledge of failure rates and mean time to failure, there will be many who are not familiar with the details of software reliability models, and have no time to become acquainted with them. Because of this, many of the currently-available tools will be of limited use to these individuals.