Automatic identification of drop-down menu widgets using mutation observers and visibility changes

Abstract

Drop-down Menus are JavaScript/CSS based Widget implementations which frequently represent the main navigation mechanism of a web application. These Widgets are used to display sub-level navigation links which are presented only as users position the mouse over the Widget. However many of the Widgets available on theWeb do not implement Accessible Rich Internet Applications (ARIA) requirements which address how to make these Widgets accessible. This paper presents an automatic approach for identifying these Drop-down Menu Widgets in web applications. The automatic identification of these Widgets works towards the implementation of ARIA automatic evaluation tools to assist web developers while developing accessible Widgets. The proposed approach was implemented in a tool called aria-check-menus which is open-source available. The tool was evaluated through a case study and the results showed evidences that the proposed approach is capable of automatically identifying Drop-down Menu Widgets in web applications, having correctly identified 146 Drop-down Menus Widgets in high traffic websites with 0.77 precision and 0.90 recall in average. Henceforth, the tool could be used in an automatic ARIA evaluation tool to support web developers in the implementation of accessible Drop-down menu Widgets.