使用程序静态分析技术识别Web程序中的用例结构

摘要

In this paper, identifying use cases in Web program is investigated in detail by program static analysis techniques. User behavior in Web pages is analyzed in detail and modeled,and event flow graph which describes user behavior in the whole system is extracted from the model. Independent control flow graphs are extracted from Web pages by existing method which based on purpose success statement and with dominator as core. Global control flow graph, which is composed of event flow graph and independent control flow graphs, is used for global use case identification. Comparing with existing related work, the features of Web program, especially user behavior in front-end program, are taken into consideration in this paper. JSP program is used for experiment in this paper since JSP is widely used at present, and the results show that use cases can be identified from JSP program effectively by this method.%本文利用程序静态分析技术对Web程序中的用例识别进行了深入研究。对Web页面的用户行为进行了深入分析并建模，从行为模型中提取出描述整个系统中用户行为的事件流图，结合采用现有的从目的成功语句出发、以必经结点为基础的用例识别方法，从各个Web页面中提取的独立控制流图，形成全局控制流图，进行全局用例识别。与以前已有工作相比，本文考虑了Web程序的特点，尤其是前端程序如何实现用户行为，并在前端程序中进行精确的用例识别。本文采用目前使用较为广泛的JSP程序进行了实验，结果表明该方法能够有效识别JSP程序中的用例结构。