Analyzing web applications by visualizing their state-flow graphs

Abstract

In todays society, it is impossible to imagine life without web applications such as our webmail client or online word processors. These applications are becoming increasingly complex, which makes it hard to maintain a thorough understanding of the application by their developers. Therefore, a need has arisen for tools that support the comprehension of web applications. This thesis builds upon previous work in which a crawler has been created that crawls a web application and generates its state-flow graph. We have pro- posed several concepts for the visualization of this state-flow graph to make it more comprehensible, such as choosing a dynamic graph drawing algorithm and applying clustering. We also integrated a visualization of test results into the state-flow graph to facilitate the locating of errors. These proposals have been implemented in a proof of concept tool called Graphicrawl. Using this tool as context in a contextual interview, we then per- formed an evaluation study to verify the usefulness of the proposed concepts and gathered feedback to improve their effectiveness. The results indicate that Graphicrawl has the potential to support comprehension and several pointers for improvement have been identified.