Repository files navigation

Since its inception, arXiv’s primary distribution format has predominantly beenPDF generated from LaTeX submitted by authors. While there are no plans to moveaway from LaTeX as the preferred submission format, nor to abandon PDFs, werecognize the need to provide distribution formats that make scientific papersmore broadly usable and accessible.

• In particular, the use of mobile devices--for which PDF is unsuitable-—toaccess internet resources including arXiv content is on the rise, especiallyin developing countries.
• Adopting HTML5 also opens up the potential for authors to integrate dynamiccontent in their papers, such as embedded video or interactive elements.
• Providing an HTML5 distribution format provides a foundation for a broaderarray of enhancements and integrations by third-party developers andresearchers that can add value for arXiv authors and readers.
• Well-formed HTML5 documents (and in particular MathML for formulae) hasadvantages over PDF for accessibility, particularly for use with screenreaders 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 andreaders. The overarching objective of this project is to significantly improvethe usability and accessibility of arXiv papers. While providing HTML is not apanacea, it is a first step in the right direction.

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

Ultimately, we would like to incorporate HTML5 as a primary distribution formatfor 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