MENDES, F. I. L.; http://lattes.cnpq.br/5941904266469546; MENDES, Francisco Igor de Lima.
Resumen:
The automatic generation of graphical user interface (GUI) tests plays a crucial role for detecting faults at web applications. In that regard, scriptless testing streamlines the process of generation and executing test cases through the identified GUI elements. However, identifying such actionable elements in an automatic and unique manner is still a challenge when dealing with complex web pages. Current approaches are tied with intricate and extensive adaptations to fit the Application Under Test (AUT). As an alternative and solution to this problem, we propose the adoption of a technique called Unique Actionable Element Search (UAES) that aims to uniquely discover actionable webelements by its functional role in an automatic fashion. Our empirical study assessed the effectiveness of UAES localization against a manual explicit markup approach. The study was conducted using four open-source projects where experienced testers identified the actionable elements as part of the explicit markup approach. The results show that our approach managed to discover 79.81% of the marked elements, while identifying new ones (8.1% of all discovered elements) that weren't highlighted by the testes.
