An entropy-based complexity measure for web applications using structural information

Web applications tend to get more change requests, evolve faster, and have shorter life-cycles than general applications. However, there are few systematic approaches dealing with their development and management, which results in a degrading structural quality and high maintenance costs. Complexity is a measure that is closely related to maintainability, understandability, and testing efforts. Most of the existing complexity measures for the web are count-based, and they provide a biased view of WAs. Therefore, it is necessary to define another complementary complexity notion. In this paper, entropy-based complexity measures, WCOXIN and WCOXOUT, are proposed for web applications, with a model which is defined using pages, relations, and parameter information. Entropy, which is connected to a system's average information quantity, can be used as a metric of software aging and structural degradation. Several experiments are conducted to show the effectiveness of the proposed complexity measure, using WANA, a tool especially developed for the experiments. The experimental results reveal that the proposed WCOXIN and WCOXOUT measures effectively reflect the structural changes of web applications, which are not considered by count-based complexity measures.