PDF/A-3u as an archival format for Accessible mathematics

TL;DR

This paper explores three standards-compliant methods for embedding LaTeX source code of mathematical expressions into PDF documents to enhance accessibility, enabling better support for assistive technologies and screen readers.

Contribution

It introduces three novel, standards-compatible techniques for embedding LaTeX and MathML sources in PDFs, improving accessibility for mathematical content.

Findings

All methods are compatible with ISO standards for PDF.

Embedding sources as attachments or /ActualText enhances accessibility.

The approaches enable simple copy-paste extraction of source code.

Abstract

Including LaTeX source of mathematical expressions, within the PDF document of a text-book or research paper, has definite benefits regarding `Accessibility' considerations. Here we describe three ways in which this can be done, fully compatibly with international standards ISO 32000, ISO 19005-3, and the forthcoming ISO 32000-2 (PDF 2.0). Two methods use embedded files, also known as `attachments', holding information in either LaTeX or MathML formats, but use different PDF structures to relate these attachments to regions of the document window. One uses structure, so is applicable to a fully `Tagged PDF' context, while the other uses /AF tagging of the relevant content. The third method requires no tagging at all, instead including the source coding as the /ActualText replacement of a so-called `fake space'. Information provided this way is extracted via simple Select/Copy/Paste actions, and is available to existing screen-reading software and assistive technologies.