PDF Accessibility

[kathari00]

Hello everyone,

I’ve recently been evaluating the accessibility of PDFs generated from Reveal.js presentations using the PDF Accessibility Checker (PAC), and I’ve identified some crucial accessibility issues. These problems also persist when using DeckTape for the PDF conversion process. The report can be found here.

Here’s what I’ve noticed:

Missing Metadata

Important metadata, such as the document language, is not set in the exported PDF. This is a critical issue as it impacts how assistive technologies, like screen readers, interpret the content.

Lack of Structural Translation
Structural elements in the HTML (e.g., headings, lists, and paragraphs) are not properly translated into the PDF. As a result, the logical structure of the document is missing. This makes it very difficult (or even impossible) for screen readers to navigate the content, creating significant accessibility barriers for users with disabilities.

These issues make the generated PDFs non-compliant with accessibility standards, such as PDF/UA.

Has anyone managed to implement a workaround or enhancement for generating accessible PDFs with Reveal.js?
If no solutions currently exist, I’d like to start a discussion on how we can improve accessibility for PDFs generated from Reveal.js presentations.

Looking forward to your insights!

[bbitarello]

Same, I've switched to making all my slides in quarto to revealjs but conversions to PDF are awful. decktape works most of the time but messes up some fonts/math sometimes but those PDFs are entirely not accessible. Even headings are not recognized and any math or code is interpreted as figures that need alt-text. On that note, even if we use alt-tex for actual images in quarto, those then don't show up in the PDF format. :/