摘要
软件测试是软件开发中不可或缺的部分,也是软件工程化方法中的重要环节。目前各种软件测试技术日趋成熟,但相关的测试基本原理框架还有待开发。在前人经验的基础上,试图提出一个初步理论框架来定义软件测试的样本空间,引入反映软件某种情况(比如缺陷)的随机变量,概括白盒测试和黑盒测试的概率测度及数学期望描述。这样的构建不仅能够加深对软件缺陷存在的理论根源的理解,从而进一步提出更好的测试方法,还对发展软件测试的科学理论有所帮助。
Abstract Software testing is a dispensable and important part for software development and software engineering.Various techniques of software testing are nowadays refined,however,relevant testing foundations are still missing.This paper,based on experiences of pioneers,tried to provide an elementary theoretical framework in which a sample space of software testing is defined,and a random variable reflecting something （e.g.bugs） of software is introduced and the probability measure and conditional expectation about white-box testing and black-box testing are generalized.This construction aims to deepen our understanding about why software bugs exist,so that software testing can be improved.Moreover it may be beneficial for developing the scientific theory of software testing.
出处
《计算机科学》
CSCD
北大核心
2014年第3期12-16,35,共6页
Computer Science
基金
国家自然科学基金项目(61170071)
金陵科技学院科研基金(jit-b-201207)资助
关键词
软件测试
软件度量
程序语言
随机测试
Software testing
Software metrics
Programming language
Random testing
