Is Wikipedia Usable for the Blind?
Marina Buzzi, Barbara Leporini · 2008 · Proceedings of the 2008 International Cross-Disciplinary Conference on Web Accessibility (W4A) · doi:10.1145/1368044.1368049
Summary
This paper evaluates the accessibility and usability of Wikipedia for blind screen reader users by analyzing three key interaction modes: browsing the home page, searching and viewing results, and editing content. The authors tested both the English and Italian Wikipedia using JAWS 8.0 with Internet Explorer 6.0.29 and Mozilla Firefox 2.0.0.12. Notably, the evaluation was conducted by two researchers with complementary perspectives — one author has been totally blind since childhood and is an expert JAWS user, while the other is sighted but tested using only basic JAWS commands. This dual approach covered both advanced and beginner screen reader interaction patterns. The study systematically catalogs usability barriers across all three interaction modes, finding that while Wikipedia content is technically accessible via screen reader, significant usability problems make interaction difficult, time-consuming, and frustrating. The editing interface presents the most severe challenges, as it was designed primarily for mouse-based visual interaction with formatting applied through clickable graphical toolbar icons that cannot be effectively operated by keyboard.
Key findings
The evaluation uncovered numerous specific usability barriers. On the home page: graphical links lacked alternative descriptions (JAWS read filenames like "ed/Pix" for pixel-spacer images), the Italian version had over 350 links on a single page, search fields were unlabeled, and content sections used too few heading levels — particularly in the Italian version. Browser differences significantly affected the experience: with Firefox, JAWS read button labels as "Go" and "Search" (the title attributes), but with IE it read the more descriptive tooltips ("Go to a page with this exact name if one exists"). Search result pages lacked clear success/failure messages — users had to infer results from contextual text. The editing interface was the most problematic: formatting required selecting from graphical toolbar icons inaccessible by keyboard, inserting special characters involved navigating over 500 graphical links, there was no way to switch between editing and selecting modes via keyboard alone, and focus management was broken when applying formatting commands. The authors propose concrete solutions including: use heading levels and hidden labels consistently, place search fields at the beginning of the content flow, provide keyboard shortcuts for all formatting (Ctrl+B for bold, etc.), offer a text-based editing alternative using wiki or HTML syntax, use a separate dedicated editing page, group special characters in a combo-box with descriptive text labels, and maintain focus position after formatting operations. They also identify WAI-ARIA as a promising solution for building an accessible editing interface.
Relevance
Although this study examined Wikipedia circa 2008, its findings illuminate usability challenges that persist across collaborative web platforms and content management systems. The distinction between technical accessibility (content can be reached by a screen reader) and practical usability (tasks can be completed efficiently) remains one of the most important concepts in accessibility practice. Many of the specific barriers identified — unlabeled form fields, missing alt text on functional images, broken focus management, mouse-dependent toolbar interactions, and excessive link counts — continue to plague modern web applications. The editing interface findings are particularly relevant to any web application that provides rich text editing, as WYSIWYG editors remain one of the most challenging interface patterns for screen reader accessibility. The paper's suggestion to offer both visual toolbar and text-based markup editing modes anticipates the dual-mode approach now common in platforms like GitHub and many CMS systems. The dual-perspective evaluation methodology — pairing a blind expert user with a sighted researcher using basic screen reader commands — is a valuable model for accessibility evaluations that captures both power-user workarounds and beginner-level barriers.
Tags: screen readers · blind users · wiki accessibility · usability · Wikipedia · JAWS · collaborative environments · keyboard accessibility · WAI-ARIA · e-learning
Standards referenced: WCAG · WAI-ARIA