In today’s world, software is used almost everywhere, from mission-critical applications, like rocket technology, to cartoon games. The necessity of certifying the reliability of software products and services has gained much importance.
Chang and Jeng make a nice attempt at proposing software reliability models that can evaluate software reliability in a faster and easier way. They have extended the statistical usage testing principle using a prejudgment factor of software quality. They propose impartial reliability evaluation models that are fair for the end user (consumer), as well as for the producer (developer). The impartial evaluation of reliability is based on two proposed reliability models: impartial success-ratio reliability (ISRR) and impartial transition-probability reliability (ITPR). The authors claim that these models help provide estimates more quickly, with a smaller number of test scripts. An example, focused on the evaluation of a Web site‘s performance, is also provided. Through this demonstration, the authors attempt to prove that impartial reliability based on a Markov model (ITPR) is superior to the ISRR model.
The software reliability evaluation proposed in this paper requires a proper estimation of the prejudgment constant that decides the impartiality of reliability evaluation. A fair estimation of this prejudgment constant is very difficult in a true software development scenario. Thus, the effectiveness of the proposed models is highly dependent on a prejudgment constant. Also, it would have been better if the authors had demonstrated their results with a more complicated software example, rather than the Web site performance example.