RFC: LaTeX compiling standardisation

[Danysan1]

I see lately lots of method are proliferating for compiling latex formulas into .md files. Sometimes these method clash against each other. IMHO we should decide together a common method, write it down in CONTRIBUTING.md and use only it fo future files.

Here's a list

Suppose this is the content of foo.tex.md:

foo $bar^2$ quz

On GitHub the formula isn't translated:

foo $bar^2$ quz

So until now many types of compilation have been used.

Type	Tools	Compilation	Output	Example	View	[Dis]Advantages
Raw	VS Code extension "Markdown+Math"		NO		Only as VS Code tab (screenshot)	No compilation needed; useless for GitHub; works only on VS Code
readme2tex	readme2tex	See CONTRIBUTING	./svgs/....svg new; foo.md new (example); foo.tex.md unchanged	Link	Ok on GitHub web, not ok on GitHub app (maybe doesn't accept svg?)	Not limited to VS Code; leaves original unchanged (useful for subsequent changes); creates partial duplicates; often original remains behind compiled; requires python module manual installation
Math>Image local	VS Code extension "Math to Image"	screenshot	/svg/....svg new; foo.tex.md changed (example)	Link	Ok on GitHub web, not ok on GitHub app	No duplicates; changes original (but leaving Latex formula in a comment); works only on VS Code
Math>Image remote	VS Code extension "Math to Image"	screenshot	foo.tex.md changed (example)	Link	Sometimes ok on GitHub web but sometimes no*, sometimes slower, not ok on GitHub app	No duplicates; changes original (but leaving Latex formula in a comment); with some hacking it can be done manually but mostly works only on VS Code
tex-builder.py	?	?	?	?	?	?

(* sometimes a bad link is generated and the image is not rendered. For example when I tried inserting $\left(\bigcup\limits_{c \ge 1} DTIME(n^c)\right)$ it generated the link https://render.githubusercontent.com/render/math?math=%5Cleft(%5Cbigcup%5Climits_%7Bc%20%5Cge%201%7D%20DTIME(n%5Ec)%5Cright) which does not work and caused the page to break.)

IMHO the best option is Math>Image local but I'm ok with the others

[montali]

You are absolutely right, and it is mostly my fault :(
I didn't think GitHub's LaTeX parser failed, so I'll agree with you on the Image local one!