Web Application Model Generation through Reverse Engineering and UI Pattern Inferring

A great deal of effort in model-based testing is related to the creation of the model. In addition, the model itself, while a powerful tool of abstraction, can have conceptual errors, introduced by the tester. These problems can be reduced by generating those models automatically. This paper presents a dynamic reverse engineering approach that aims to extract part of the model of an existing web application through the identification of User Interface (UI) patterns. This reverse engineering approach explores automatically any web application, records information related to the interaction, analyses the gathered information, tokenizes it, and infers the existing UI patterns via syntactical analysing. After being complemented with additional information and validated, the model extracted is the input for the Pattern-Based Graphical User Interface Testing (PBGT) approach for testing existing web application under analysis.