Introduction

This document is a technical overview of the MiniLatex project. The project consists of three elements:

1. A specification and grammar for a subset of LaTeX called MiniLatex

2. A parser-renderer package that can be used to parse MiniLaTeX and render it to HTML. The code is open-source: see Github or package.elm-lang.org

3. A content management system which hosts MiniLatex and provides user authentication, a searchable store of MiniLatex documents, etc. See knode.io. Documents written using knode.io can be made public, in which case users can search for them and read them withought signing in. See, for example, knode.io/427

For a demonstration of this technology, see MiniLatex Live.

MiniLatex is written in Elm, a statically typed functional language created by Evan Czaplicki. Elm is a language whose target use is the development of web apps. Its type system gives unparalled assurance that code, once deployed, "just works," as well as the ability to refactor without fear and therefore to write and maintain code for the long term. These facts, combined with the expressivity of Elm's type system and the quality of its parser combinator library (elm/parser), makes it an ideal vehicle for the MiniLatex project.

Plans. MiniLatex is still a research project and knode.io is still in development. That said, I've used MiniLatex on knode.io to write lecture notes, e.g, knode.io/427. So it is ready to be used by brave souls. If you would like to experiment with it, by all means do so. I very much value your feedback. You can write to me at [e-mail]. See the Plans section for more information about where MiniLatex is going.

Acknowledgements

I would like to thank the many people on the Elm Slack who have helped me with questions on this project, and who have made its success possible. I give special thanks to

• Evan Czaplicki

• Ilias van Peer

• Luke Westby

Evan suggested that I look at elm/parser tools for building a parser for another project. I did, and after a few weeks of study, I realized that I could attempt something much more ambitious: a parser for subset of LaTeX. That is how this project was born. Evan later suggested that I render to Html a rather than to regular HTML This change is part of what make rendering so very fast.

In starting this project, I had never used parser combinators before, so going from the definition of the type of the AST to a functioning parser was at first quite a struggle. The generous help that I received rom Ilias, whom I have never met, other than through the Elm Slack, helped me to find my way in this unfamiliar terrain.

As I made the transition from version 0.18 to 0.19 of the Elm compiler, I encountered what looked like a blocking problem in using MathJax to render mathematical equations. Luke provided me with a few beautiful lines of code that mde it all work.