ENB: About mathjax, TeX and typesetting

[Edward K. Ream]

This Engineering Notebook post completes my retrospective of Knuth's TeX system and summarizes my study of MathJax.  I am writing this post so that I can move on to other projects.

About MathJax

I was mainly curious about how MathJax works.  The wikipedia summarizes MathJax well:

1. MathJax is nothing but embedded javascript.

2. "MathJax is meant only for math display, whereas LaTeX is a document layout language, MathJax only supports the subset of LaTeX used to describe mathematical notation."

As a result, MathJax has a relatively small size.

Confusingly, there are separate GitHub repositories for MathJax deliverables and MathJax sources. Most MathJax sources are in typescript.

The 3.0 docs are incomplete. The 2.0 API docs provide all I want to know. There are front ends (input code) for TeX and MathML and back ends (output code) for svg, etc. Front ends convert user input to a tree. Back ends convert the tree to their respective output formats.  This design is straightforward and needs little detailed documentation.

About TeX

The sources for TeX (web.tex) are here. The typeset (woven) version of web.tex is here. The web-to-outline script (in scripts.leo) converts web.tex to an outline.

Imo, neither outlines nor typesetting could compensate for pascal's deficiencies.

That's all I want to say about the program whose problems helped create Leo.

About typesetting

Imo, typesetting markup will always obscure documentation. Yes, one would prefer to see the rendered form of equations. The VR3 can render those equations, so nothing extra seems necessary.

Summary

Any day without typesetting is a good day.

I have no further interest in TeX, MathJax, or typesetting.  It's time to move on.

Edward