The semantic web, web accessibility, and device independence
Lisa Seeman · 2004 · Proceedings of the 2004 International Cross-Disciplinary Workshop on Web Accessibility (W4A) · doi:10.1145/990657.990669
Summary
Written by Lisa Seeman of UB Access, this paper argues that Semantic Web technologies — particularly RDF (Resource Description Framework) annotations — can dramatically improve web accessibility beyond what WCAG guidelines alone can achieve, especially for people with learning and cognitive disabilities. Seeman contends that while WCAG addresses technical accessibility (keyboard access, alt text, screen reader compatibility), it falls short for users who need content simplification, symbolic representations, or fundamentally different renderings of the same information. The paper proposes that authors add a semantic layer to their content using RDF annotations that describe the meaning, role, and relative importance of content sections — mapping terms to concepts, identifying content roles (main menu, legal footer, contact link), and marking sections as more or less important for different user profiles. A proxy server or user agent can then interpret these annotations to generate alternative renderings tailored to individual needs. The paper introduces SWAP (Semantic Web Accessibility Platform), a concrete implementation by UB Access that creates "SWAPviews" — alternative renderings of websites generated by interpreting semantic annotations through a proxy server. SWAP provides basic WCAG-compliant accessibility features plus enhanced capabilities like navigation views for non-sighted users and simplified renderings for users with cognitive disabilities.
Key findings
The paper identifies several powerful applications of semantic annotations for accessibility. Content roles (main menu, sitemap link, contact link) can be annotated so that users can assign consistent access keys across all sites based on role rather than per-site arbitrary key assignments. Complex terms and jargon can be linked to a lexicon, enabling automatic simplification for users with reading difficulties. Non-literal or ambiguous text can be annotated with clarifying information, allowing user agents to render unambiguous versions. Information can be marked by importance relative to user profiles, enabling hiding of less-relevant content for users who benefit from reduced density. Documents can be rendered as SVG-based schematic diagrams with selectable sections, benefiting dyslexic users. Users can add their own annotations marking important sections for later review. The paper also expands the concept of device independence beyond screen size to include knowledge processing — arguing that a "device" for someone with a learning disability might need to transform content at a conceptual level, not just reformat it visually. Seeman notes that these accessibility-driven semantic annotations also benefit device independence, internationalization, and localization, creating additional incentives for authors to invest in the work.
Relevance
This paper is remarkably prescient. Lisa Seeman went on to co-edit the W3C Cognitive Accessibility guidance and lead work on the WAI-Adapt (previously Personalization) specification, which directly implements many ideas presented here — content annotations for simplification, symbol support, and user-profile-based rendering. The concept of annotating content roles anticipated ARIA landmarks and the HTML5 structural elements (main, nav, aside). The vision of proxy-based alternative renderings foreshadowed browser reader modes and accessibility overlay approaches (though with far more rigor). For practitioners today, the paper highlights that cognitive accessibility requires fundamentally different strategies than sensory or motor accessibility — it is not enough to ensure content is perceivable and operable; it must also be understandable at the user's level. The Semantic Web as originally envisioned never fully materialized, but the specific accessibility applications Seeman described are being realized through WAI-Adapt, schema.org annotations, and AI-powered content simplification.
Tags: semantic web · RDF · cognitive accessibility · learning disabilities · personalization · knowledge representation · assistive technology
Standards referenced: WCAG 1.0 · WCAG 2.0