AChecker: open, interactive, customizable, web accessibility checking
Greg Gay, Cindy Qi Li · 2010 · Proceedings of the 2010 International Cross Disciplinary Conference on Web Accessibility (W4A) · doi:10.1145/1805986.1806019
Summary
This paper introduces AChecker, an open-source web accessibility checker developed at the University of Toronto's Adaptive Technology Resource Centre. AChecker was designed to address two fundamental shortcomings of existing accessibility evaluation tools: the lack of transparency about what is actually being checked, and the inability to involve human judgment in the evaluation process. Different accessibility checkers routinely produce different results when analyzing the same content, yet most tools operate as black boxes where reviewers cannot verify what rules are being applied or how. AChecker makes its entire checking logic visible and editable, allowing users to inspect, modify, and create accessibility checks and guidelines. The tool supports nine international standards out of the box, including WCAG 1.0 and 2.0, Section 508, the Italian Stanca Act, and the German BITV. It also provides web services APIs for integration into authoring tools, with initial implementations in the TinyMCE HTML editor and the ATutor learning management system.
Key findings
AChecker introduces a three-tier classification system for accessibility issues that acknowledges the fundamental limits of automated testing. "Known Problems" are barriers the tool can identify with certainty, such as images missing alt attributes. "Likely Problems" are probable barriers identified through heuristics — for example, very short link text that is probably not meaningful enough, but might be acceptable in context (like "home"). "Potential Problems" are issues that current technology simply cannot verify automatically, such as whether an embedded video has captions or audio descriptions. This tiered approach creates a structured workflow: reviewers first fix known problems, then make yes/no decisions on likely problems, and finally manually verify potential problems. The tool's customizability allows users to create entirely new guideline sets — useful for organizations with specific technology environments or accessibility requirements beyond standard guidelines. As open-source software, AChecker ensures its checking logic can be studied, verified, and extended by anyone as accessibility knowledge evolves.
Relevance
AChecker made an important contribution to accessibility testing methodology by formalizing the distinction between what automated tools can and cannot detect — a concept now widely understood but not well-articulated in tools at the time. The three-tier Known/Likely/Potential classification influenced how the accessibility community thinks about the role of automated versus manual testing. For practitioners, the key lesson remains highly relevant: no automated tool can fully evaluate accessibility, and any complete evaluation must include human judgment on issues of meaning, context, and equivalence. The tool's open-source, customizable approach also anticipated the modern trend toward configurable rule sets in tools like axe-core. While AChecker itself has been superseded by newer tools, its design principles — transparency, interactivity, and extensibility — represent best practices for accessibility evaluation tools that continue to apply today.
Tags: automated testing · accessibility evaluation · open source · WCAG compliance · accessibility tools
Standards referenced: WCAG 1.0 · WCAG 2.0 · Section 508 · Stanca Act · BITV