软件可靠性评估研究

软件错误引起的后果差别很大，有的几乎可以忽略不计，有的可能是灾难性的。目前，软件可靠性模型很少考虑到这一点，我们在实际工作中发现量化这一差别是很有必要的。为了增加人们对错误重要性差别的意识，讨论了在考虑错误级别时，如何应用Jelinski-Moranda模型和Goel-Okumoto,Nelson模型以及Nelson和Bron-Lipow模型。这些讨论和研究说明软件可靠性不仅决定于软件中的错误数，而且决定于软件中的错误级别。

Research on Software Reliability Estimation

The cost of errors varies widely from errors with negligible costs to catastrophic ones. Many researchers are aware that not all failures have similarly severe consequences, but there is reluctant to quantify these differences. In order to increase the awareness of error severity in this paper we make discussion about how to apply some famous reliability models to assess software reliability when we take these differences into account. What we have done shows that not only the number of errors but also the consequences of errors influence the software reliability.