Page Flow Oriented Variability Reverse Analysis And Customization Approach For Web Applications

A set of domain-specific web applications are usually similar variant software products. With regard to the similarity amongst their requirements and designs, developers usually reuse code frivolously when writing the codes, e.g., writing code in copy-paste-modify way. With the increase of the number and complexity of variant products, it has been a critical problem how to understand and analyze their commonality and variability. On the other hand, if developers can achieve the design models regarding commonality and variability by reverse engineering those legacy variant products, they can efficiently develop new applications by customizing variant points as well.To this end, this thesis propose a webpage flow oriented approach to reverse engineering and customizing variability of domain-specific applications. The approach extracts web page flow of various applications, i.e., a graph consisting of web page nodes and business operation nodes, which reflects the overall business logic of different web applications. The approach takes as input the source code and configuration files of several legacy variant web applications, and extract the webpage flow model of each variant product through reverse engineering analysis technique. Next, the approach adopts variability analysis technique to analyze the variability of different web page flows. As a result, the generated webpage flow graph with variability is able to assist developers to understand the commonality and variability of variant products in the aspect of web flow. Based on the resulted variability, the approach is even able to assist developers to derive and instantiate web flow of a new web application by customizing variant points. The approach is implemented as a tool with the functionality of reverse engineering technique and graphical variability presentation. In addition, this thesis also introduce the implementation technique to the proof-of-concept tool and validate the effectiveness of the approach as well as its tool by one case study.