Semi-Autonomous, Site-Wide A11Y Testing Using an Intelligent Agent

摘要

Most countries across the globe have now ratified and/or signed the United Nations Convention on the Rights of Persons with Disabilities, the culmination of an effort started in the 1980's to protect the rights and dignity of people with disabilities. W3C's Web Content and Accessibility Guidelines documents, WCAG 2.0 (2008), ISO/IEC 40500 (2012), and WCAG 2.1 (2018), establish conformance requirements, which most nations have adopted as their de jure standard for determining website compliance with the UN convention. Although several tools support static WCAG web page analysis, present tools lack the automation required to automatically evaluate an entire website. In addition, current techniques are capable of discovering less than 30% of WCAG Level A and Level AA conformance issues. This paper presents an intelligent testing agent, Agent A11Y, capable of semi-autonomously exploring a website and more thoroughly evaluating its compliance with WCAG guidelines. To expand the system's automated WCAG guideline coverage, Agent A11Y employs novel dynamic testing strategies in combination with existing static web page analysis solutions. Additional tooling focuses manual testing activities on WCAG requirements that are difficult to fully automate. Supporting the generation of effective and actionable reports, Agent A11Y utilizes site exploration information to improve the aggregation and categorization of site-wide testing results. In conjunction with a discussion of the aforementioned Agent A11Y exploration, testing, and reporting methods, the paper presents experimental results and analysis to illustrate the comparative usefulness of the described techniques for conducting site-wide accessibility assessments.