Dynamic model extraction and statistical analysis of Web applications: Follow-up after 6 years

In 2002 we proposed a method to reverse engineer a Web application model. The proposed method deals with dynamic Web applications, consisting of server components, typically interacting with a persistent layer, which build the Web pages displayed on the browser dynamically. Dynamic analysis and page merging heuristics were used for model extraction. The proposed model was successfully adopted in the Web analysis and testing research community. However, the features of future Web applications (involving rich client components and asynchronous communication with the server) challenge its future applicability. In this paper, we analyze the key properties of the 2002 model and identify those modeling decisions that remain valid and can be used to guide the extraction of models for future Web applications