Increasing access to information for the print disabled through electronic documents in SGML
Bart Bauwens, Jan Engelen, Filip Evenepoel, Chris Tobin, Tom Wesley · 1994 · Proceedings of the First Annual ACM Conference on Assistive Technologies (Assets '94) · doi:10.1145/191028.191043
Summary
This paper describes work by the CAPS Consortium (Communication and Access to Information for People with Special Needs), an EU-funded project, and ICADD (International Committee on Accessible Document Design) to build accessibility into SGML (Standard Generalized Markup Language) encoded electronic documents. The authors identify two fundamental barriers to information access for print-disabled people: accessible formats (braille, large print, synthetic speech) are treated as peripheral activities divorced from normal publishing workflows, and their production is slow and manually intensive. CAPS proposed that accessibility should be an automatic by-product of the normal information creation process, achievable through standardized structured electronic documents. The paper provides a thorough overview of SGML — its emphasis on descriptive rather than procedural markup (tagging a paragraph as <para> rather than specifying formatting instructions), and its document type definition (DTD) concept where a formal specification defines the structure of a class of documents. The key insight is that if a DTD can be made accessible, then all document instances conforming to that DTD are automatically accessible without editing individual documents.
Key findings
CAPS developed "Associated Specifications" — SGML-conformant extensions added to DTDs (not to individual documents) that enable transformation to accessible formats. The mechanism uses four fixed SDA (SGML Document Access) attributes: SDAFORM maps complex DTD elements to a simple 22-element ICADD base tag set (headings h1-h6, paragraphs, inline formatting); SDARULE provides context-sensitive mappings (e.g., a title within a section maps to h1, within a chapter to h2); SDAPREF and SDASUFF replace start/end tags with explanatory text. For interactive applications like synthetic speech browsing, CAPS added SDAEXPL to provide human-readable explanations of element names. Crucially, ICADD succeeded in having these accessibility mechanisms incorporated into the new ISO 12083 standard for Electronic Manuscript Preparation and Markup — reportedly the first time disability issues were directly built into an ISO standard for commercial use. The CAPS project also developed the CAPSNEWS DTD as an interchange format for electronic newspapers for print-disabled readers, and built a Pilot Electronic Library accessible via adapted workstations and telephone-based voice response systems.
Relevance
This paper is a direct ancestor of modern web accessibility standards. The principle that document structure should be explicitly encoded through descriptive markup (not visual formatting) to enable accessible transformations is exactly the philosophy underlying HTML semantics, WAI-ARIA, and WCAG. The ICADD base tag set — mapping complex document elements to a simplified accessible vocabulary — presages how ARIA roles map custom web components to accessible semantics. The achievement of embedding accessibility into ISO 12083 demonstrated that accessibility could be built into publishing standards rather than bolted on afterward — a principle now central to EPUB accessibility, PDF/UA, and accessible publishing workflows. The paper's emphasis on making accessibility automatic within commercial production processes, rather than a separate remediation activity, remains the aspirational goal of modern accessible content creation. For practitioners, this work shows that the core arguments for structured, semantic markup as a foundation for accessibility have been understood for over 30 years.
Tags: document accessibility · SGML · structured documents · print disability · braille · large print · standards · accessible publishing · European accessibility
Standards referenced: ISO 8879 (SGML) · ISO 12083 · ADA