A brand new look for our publications!

Jacky Song | May 19, 2025

After a lot of work, we've made a big change - we've moved our technical overview and 2024 end-of-year report to the MyST framework! Let's dive into why we made this change, and what it means for the future.

At Project Elara, we had a constant problem with our online publications. Being an organization focused on open research, sharing our work is a big part of what we do. Previously, our existing solutions for sharing our research (papers, reports, etc.) were limited to three options:

• Markdown, which is plaintext (and thus human-readable without any special tools) and very easy-to-write, but cannot generate HTML and PDFs on its own
• LaTeX, which is plaintext and can automatically generate PDFs, but is a major headache to set up and doesn't solve the issue of generating HTML
• App-specific formats like TeXmacs, which is (reasonably) easy-to-write and can generate (with some effort) both HTML and PDFs, but is not plain text (or at least not human-readable)

We've pivoted back-and-forth between these approaches, but we never had anything suitable, which led to extremely hacky solutions like manually editing HTML code generated by TeXmacs (which generates HTML that resembled XHTML more than modern HTML5!). Luckily, not anymore!

MyST is a framework designed with open science in mind - creating interactive scientific articles and papers with rich metadata and supporting a variety of formats. We've already ported our two most important publications (the technical overview and 2024 end-of-year report) to MyST, and the results are stunning. Crucially, MyST solves multiple issues at once:

• It is an actively-developed framework that is a sub-project of the Jupyter Project, instead of an in-house solution we need to maintain, so its output is professional out-of-the-box and can be relied upon (one less thing to worry about!)
• It allows writing articles and papers in an extended markdown syntax, while delegating the build step of creating a modern HTML site and PDF to a simple (set of) commands, meaning we have the benefits of human-readable markdown without losing any of the output quality of LaTeX

I am very excited that we are using this new framework, and I hope it opens new possibilities to us and the Project!