HTML as an Accessible Format for Papers

Source: Hacker News

HTML as an accessible format for papers

Accessibility barriers in research are not new, but they are urgent. The message we have heard from our community is that arXiv can have the most impact in the shortest time by offering HTML papers alongside the existing PDF.

arXiv has successfully launched papers in HTML format. We are gradually backfilling HTML for arXiv’s corpus of over 2 million papers over time. Not every paper can be successfully converted, so a small percentage of papers will not have an HTML version. We will work to improve conversion over time.

The link to the HTML format will appear on abstract pages below the existing PDF download link. Authors will have the opportunity to preview their paper’s HTML as a part of the submission process.

The beta rollout is just the beginning. We have a long way to go to improve HTML papers and will continue to solicit feedback from authors, readers, and the entire arXiv community to improve conversions from LaTeX.

Why “experimental” HTML?

Did you know that 90 % of submissions to arXiv are in TeX format, mostly LaTeX? That poses a unique accessibility challenge: to accurately convert from TeX—a very extensible language used in myriad unique ways by authors—to HTML, a language that is much more accessible to screen readers, text‑to‑speech software, screen magnifiers, and mobile devices. In addition to the technical challenges, the conversion must be both rapid and automated in order to maintain arXiv’s core service of free and fast dissemination.

Because of these challenges we know there will be some conversion and rendering issues. We have decided to launch in beta with “experimental” HTML because:

• Accessible papers are needed now. We have talked to the arXiv community, especially researchers with accessibility needs, and they overwhelmingly asked us not to wait.
• We need your help. The obvious work is done. Reports from the community will help us identify issues we can track back to specific LaTeX packages that are not converting correctly.

Error messages you may see in HTML papers

HTML papers on arXiv.org are a work in progress and will sometimes display errors. As we work to improve accessibility we share with you the causes of these errors and what authors can do to help minimize them.

Learn more about error messages you may see in HTML papers

Ways to help

1) Read HTML papers and report issues

We encourage the community to try out HTML papers in your field.

Report an issue

1. Go to the abstract page for a paper you are interested in reading.

2. Locate the PDF download link and click the new link for HTML.

3. Report issues by either

   • clicking the Open Issue button,
   • selecting text and clicking the Open Issue for Selection button, or
   • using Ctrl+? on your keyboard.

   If you are using a screen reader, use Alt+y to toggle accessible reporting buttons per paragraph.

Please do not create reports that the HTML paper doesn’t look exactly like the PDF paper.
Our primary goal is to make papers more accessible, so the focus during the beta phase will value function over form. HTML layouts that are incorrect or illegible are important to report. We do expect HTML papers to present differently than the same paper rendered in PDF: line breaks will occur in different places, there may be more white space, and intricate typographic layouts will be simplified. This is by design.

HTML is a different medium and brings its own advantages versus PDF. In addition to being much more compatible with assistive technologies, HTML adapts far better to the characteristics of the device you are reading on, including mobile devices.

2) Help improve the conversion from LaTeX

• Authors: Follow our guide to LaTeX Markup Best Practices for Successful HTML Papers.
• Developers: Contribute to the conversion tooling. Our collaborators at LaTeXML maintain a list of issues and welcome feedback and contributions.
• Publishers, societies, conference organizers: Review the .cls files your organization recommends to authors for unsupported packages. Providing .cls files that use supported packages is an easy way to support and sow accessibility in the scientific community.

Thank you to our collaborators

We want to share a special thank you to all the scientists with disabilities who have generously shared their insights, expertise, and guidance throughout this project.

We also thank two organizations without which HTML papers on arXiv would not be possible: the LaTeX Project and the LaTeXML team from NIST. We deeply appreciate each member of these teams for their knowledge, incredible work, and commitment to accessibility.