This looks nice, I personally would likely switch to it from HedgeDoc if TinyMCE really was “customizable and flexible” enough to have a vi-mode (it looks like it is flexible enough to add one with a browser plugin like wasavi, but I only use web editors on computers where I can’t install plugins let alone a good text editor that makes all web ones useless).
I had a look HedgeDoc (first time I heard of it), it’s very nicely done.
I think you are right that if you have a good text editor, no need to experiment too much with a web based one. And the market for people who need to typeset math on “guest” computers where access to a web based app is valuable would be very small.
The use case for my editor was people who want to publish math online (via MathJax) and want an easy tool for it. A side benefit was that it is somewhat useful for general mathematical typesetting. If I had to write a desktop text editor for math, I would take a different route and it would be much more powerful. But there are already many good software for that.
TinyMCE is quite flexible, and doesn’t require any installations (it’s plugins are just more JavaScript). I created custom keyboard shortcuts for mathematical symbols, one would put in a bit of JavaScript to add VIM like editing - probably take a weekend of hacking.
Didn’t see any plugins for modal editing - definitely something somebody should do! Overall I found TinyMCE a highly polished and stable foundation for online editing and it was a pleasure to work with.
The Online Encyclopedia of Integer Sequences has a lot of technical math writing that uses nothing but ascii. The fact that they have not adopted mathjax is sort of the mathematics equivalent of refusing to use JQuery or insisting on writing your own game engine. But I guess that is just not the culture.
It is unfortunate, makes it hard to read.[0] Although there's something to be said for a format like that shuts no-one out. Also, the OEIS has been around in different forms since the 1960s - "The database was at first stored on punched cards" – and it would be a huge effort to convert to LaTeX. "As of November 2020, it contains 338,526 sequences."
[0] Say what you will about LaTeX, it's amazing when you come across a mathematics paper from the pre-LaTeX era presenting complicated mathematical typsetting done...with a typewriter! - it's so very hard to understand and so unpleasant to read in comparison, that usually I don't bother. TeX + LaTeX was such a huge teknulogical advance.
I find mathematical content to behind the curve when it comes to digitization. The issue I guess is that math is much more expressive than text, and everybody has a different style.
There is no consensus offline on how to write math (for example I would recommend all variables in proofs to be English letters and not Greek letters - why add additional complexity in terms of digital input by using Greek letters, unless for specific needs, like the constant pi or capital Sigma). Smaller examples would be somebody may write a + b = 3 while somebody else will write x + y = 3 (should there be consistency? I personally don’t think so, but just flagging out that this is a variation too).
Well the problem is that it’s easy to run out of letters with only the latin alphabet.
Also some fields do have conventions. For instance in numerical linear algebra, many papers us the rules of (1) matrix in roman capital letters (2) vectors in roman normal-size letters (3) scalars in greek normal-size letters.
