Abstract: Software knowledge plays an important role in software testing and software reliability model. Based on the theoretical analysis and derivation on the Weibull distribution of defect density, this paper proposes that the knowledge is closely related to the morphological of the Weibull distribution, and proof that the software knowledge impacts the changes of the morphological of the distribution mainly from the changes of the scale parameter c, while c can be expressed as a quantitative expression of software knowledge amount. In this paper, software testing engineering experiment is carried out to verify the proposed conclusion, which shows that more knowledge testers have, the smaller the scale factor c of Weibull distribution becomes. Furthermore, according to the degree of the software knowledge, the trend of the problems found in testing can be guessed, so as to predict the reliability of the software.?

Key words: software knowledge, software test, reliability model, Weibull distribution, Fault mode

摘要： 软件知识在软件测试和软件可靠性模型中起着重要的作用．从缺陷密度的威布尔分布出发，分析测试者的知识水平与故障分布的定量关系，发现尺度参数c与知识量成反比关系．据此可以依据知识量与缺陷发现的分布曲线预测测试发现问题的趋势及软件的可靠性，从而确定测试结束时机．通过开展软件测试工程实验进行了所提出结论的验证，这表明，测试者的知识越多，Weibull分布的比例因子c变小．此外，根据软件知识的程度，在测试中发现的问题的趋势估计，以预测软件的可靠性．

关键词: 软件知识, 软件测试, 可靠性模型, 威布尔分布, 故障模式