User Individuality Management in Websites Based on WAI-ARIA Annotations and Ontologies
Xabier Valencia, Myriam Arrue, J. Eduardo Pérez, Julio Abascal · 2013 · Proceedings of the 10th International Cross-Disciplinary Conference on Web Accessibility (W4A) · doi:10.1145/2461121.2461128
Summary
This paper presents a comprehensive system for automatically adapting websites to individual users' accessibility needs using WAI-ARIA-based annotations and an ontology that models users, adaptation techniques, and their relationships. Unlike previous transcoding systems that targeted single user groups (typically blind users), this system addresses cognitive, physical, and sensory impairments through a repository of 99 identified adaptation techniques, of which 48 were fully implemented. The techniques are classified into three groups: content adaptations (modifying what is displayed), presentation adaptations (changing visual style and layout), and navigation adaptations (altering site structure and navigation mechanisms). The system uses an annotation language that extends WAI-ARIA with new roles (ContextInfo, Tutorial, FAQ, SiteMap, Captions, GeoMap) and properties (dimmable, hideable, stretchable, removable, priority) to semantically describe website elements. An annotator marks up a website's CSS elements with these roles and properties, and the ontology's reasoning rules then determine which adaptation techniques to apply based on a given user's characteristics.
Key findings
The system defines 13 user subgroups across three general categories: cognitive impairments (decline in attention, learning disabilities, language disabilities, reduced memory), physical impairments (limited movement, inability to use mouse), and sensory impairments (low vision, blindness, colour blindness, photosensitivity, eye strain, hearing loss, deafness). To manage the complexity of applying 48 techniques across these subgroups, the authors introduced three stereotype profiles: Narrow Style (for users with combined movement, vision, attention, and memory issues — applies 32 techniques including linearized content, simplified navigation, enlarged fonts, high contrast), Index Style (for users who cannot use a mouse — applies 4 techniques focused on keyboard navigation with a sidebar table of contents), and Sorted Style (for users with sight disabilities — applies 15 techniques including content reordering by significance, enlarged images and fonts, high contrast with negative polarity). The system was demonstrated on two real websites: Discapnet (a disability services portal) and Bidasoa Turismo (a tourism site). Notably, the authors found that Discapnet — despite being designed for users with disabilities — was well-tagged for blind users but presented barriers for users with cognitive impairments due to navigation inconsistency and information overload.
Relevance
This research demonstrates that accessibility is not one-size-fits-all — different disabilities require fundamentally different adaptations that can even conflict with each other (e.g., enlarging graphical buttons for low vision increases scrolling for users with limited movement). The ontology-based approach to managing these conflicts through stereotypes and constraint rules is a sophisticated attempt at personalized accessibility. For practitioners, the repository of 99 catalogued adaptation techniques (classified by disability subgroup and technique type) serves as a valuable reference for understanding the breadth of adaptations different users may need. The use of WAI-ARIA as the foundation for the annotation language is pragmatic, leveraging existing web standards. However, the manual annotation process was acknowledged as tedious and requiring knowledge of CSS, HTML, and the custom annotation language — a significant barrier to adoption. The finding that a disability-focused website still created barriers for some disability groups reinforces that accessibility must consider the full spectrum of user needs, not just visual impairments.
Tags: WAI-ARIA · personalization · web accessibility · adaptive interfaces · ontology · content adaptation · user modeling · cognitive accessibility · visual impairment · motor accessibility
Standards referenced: WAI-ARIA · WCAG 2.0 · ISO/IEC TR 29138