对逻辑覆盖软件测试准则的公理化评估

Axiomatic Assessment of Logic Coverage Software Testing Criteria

由于形式规格说明采用一种精确、一致、容易被机器自动处理的符号系统来描述软件需求,因而形式规格说明为测试用例的自动生成和软件功能的验证提供了基础.在基于形式规格说明的测试过程中逻辑覆盖测试准则是一组常用的测试准则,如何选择和使用其中的每个测试准则是应用这组测试准则时面临的主要问题.因此分析和比较这组测试准则中每个测试准则的性质将为测试工程师选择测试准则提供指导和帮助.对测试充分性准则的公理化评估是一种比较测试准则的方式,这种方式将对理想的测试准则的直觉需求定义为一觳椴馐宰荚蚴欠衤愀米楣砝捶治龊捅冉舷嘤Φ牟馐宰荚?描述了一组理想的逻辑覆盖测试准则应该具有的性质和用来确定一个测试充分性准则是否完全的生成算法.这组性质被形式化地定义为一组公理.利用这种形式化的定义,用定理的形式精确地给出了这些性质之间的关系.最后通过这组公理系统来评估现有的逻辑覆盖测试准则.评估的结果为测试人员在实际过程中选择逻辑覆盖测试准则提供了指导.

Since formal specifications precisely describe the software requirements in a form that can be automatically manipulated, they can be used as the base for automatic test generation and software verification. Logic coverage criteria are the common criteria used in specification-based testing. The main problem of applying these criteria that test engineers face is how to appropriately select each criterion. Thus, the comparison and analysis of these criteria will give a guide to applying each criterion. Axiomatic assessment of test adequacy criteria is an approach to comparing test criteria. This approach defines the intuitive requirements of ideal test adequacy criteria as some axioms, then compares the test adequacy criteria by checking if they satisfy these axioms. This paper proposes some positive properties as the intuitive requirements of ideal logic coverage criteria, and gives a generating algorithm that is used to determine whether a logic coverage criterion is complete. These properties are formally defined as an axioms system. With these formal definitions, the relations among the logic coverage criteria are described as some theorems. Finally, the common logic coverage criteria are assessed against the axioms system. From the assessing result, testers can get some conclusions that help them apply these criteria in practice.