Since its inception, arXiv’s primary distribution format has predominantly been PDF generated from LaTeX submitted by authors. While there are no plans to move away from LaTeX as the preferred submission format, nor to abandon PDFs, we recognize the need to provide distribution formats that make scientific papers more broadly usable and accessible.

• In particular, the use of mobile devices--for which PDF is unsuitable-—to access internet resources including arXiv content is on the rise, especially in developing countries.
• Adopting HTML5 also opens up the potential for authors to integrate dynamic content in their papers, such as embedded video or interactive elements.
• Providing an HTML5 distribution format provides a foundation for a broader array of enhancements and integrations by third-party developers and researchers that can add value for arXiv authors and readers.
• Well-formed HTML5 documents (and in particular MathML for formulae) has advantages over PDF for accessibility, particularly for use with screen readers and other assistive technology.

• Michael Kohlhase (Friedrich-Alexander Universität Erlangen-Nürnberg)
• Ben Firshman (arXiv-Vanity)
• Deyan Ginev (Friedrich-Alexander Universität Erlangen-Nürnberg)
• Erick Peirson (arXiv)
• Martin Lessmeister (arXiv)

Our top priority is to provide a high-quality service to all arXiv authors and readers. The overarching objective of this project is to significantly improve the usability and accessibility of arXiv papers. While providing HTML is not a panacea, it is a first step in the right direction.

• O1: Develop a cloud-native service that provides HTML renderings from LaTeX source submitted to arXiv, leveraging LaTeXML.
• O2: Demonstrate the feasibility and value of the service by providing it on an experimental basis to arXiv authors, with links to HTML on the public abstract page. This will involve providing detailed guidance and feedback to authors about how to write LaTeX that generates high-quality accessible HTML.
• O3: Provided that O1 and O2 are achieved, render all arXiv papers submitted as LaTeX to HTML.
• O4: Provide HTML documents as API resources for third-party developers and researchers. A crucial component of this project is promoting experimentation by researchers and third-party developers, and making the results of their work visible to arXiv readers and at the same time providing added-value services to arXiv readers.

Ultimately, we would like to incorporate HTML5 as a primary distribution format for arXiv papers, alongside PDF.

First, install Docker and pull the Engrafo image:

$ docker pull arxivvanity/engrafo

Then run the development environment:

$ docker-compose up

$ script/test