Detecção Automática de Incompatibilidades Cross-Browser utilizando Redes Neurais Artificiais

Abstract

Cross-Browser Incompatibilities (XBIs) represent inconsistencies in Web Application when introduced in different browsers. The growing number of implementation of browsers (Internet Explorer, Microsoft Edge, Mozilla Firefox, Google Chrome) and the constant evolution of the specifications of Web technologies provided differences in the way that the browsers behave and render the web pages. The web applications must behave consistently among browsers. Therefore, the web developers should overcome the differences that happen during the rendering in different environments by detecting and avoiding XBIs during the development process. Many web developers depend on manual inspection of web pages in several environments to detect the XBIs, independently of the cost and time that the manual tests represent to the process of development. The tools for the automatic detection of the XBIs accelerate the inspection process in the web pages, but the current tools have little precision, and their evaluations report a large percentage of false positives. This search aims to evaluate the use of Artificial Neural Networks for reducing the numbers of false positives in the automatic detection of the XBIs through the CSS (Cascading Style Sheets) and the relative comparison of the element in the web page.