Bility : automated accessibility testing for mobile applications

Abstract

In this thesis, I designed and implemented a testing framework, called Bility, to assess the accessibility of mobile applications. Implemented in Java and Kotlin, this framework automatically navigates through an application, finding both dynamic and static accessibility issues. I developed techniques for unique view detection, automatically building a representation of the dynamic behavior of an application, and making navigation decisions based on application state and history in order to find these accessibility issues. I compared Bility's ability to detect issues against existing tools such as the Google Accessibility Scanner. I found that Bility performs significantly better and is able to detect additional static issues, as well as discover dynamic accessibility issues that current tools do not detect.