Should we have MathJax in MVP?

Reading through KaTeX’s supported language I didn’t notice anything I use regularly in MathJax missing (of course there are some things I use from packages in real LaTeX that aren’t there, but MathJax doesn’t have then either).

1 Like

Harking back to my previous post about the universality of MathJax and/or KaTeX, I recently found MermaidJS.
It’s a diagram rendering library. Primarily for UML, but it supports things like Venn diagrams, too. I find the flow chart rendering particularly useful, not only for technical things.

As for KaTeX, they’ve added quite a bit to the feature set, but apparently, there are still rendering problems with some fonts. MathJax adds a context menu to their rendered sections, KaTeX doesn’t. Personally, I never use that feature, but I know many people like the ability to quickly copy LaTeX code from a website into their own LaTeX source code.

2 Likes

Actually I use the copy LaTeX option on SE to copy formulas from text to comments or from questions to answers.

However if I understand correctly, KaTeX has an extension that gives LaTeX source directly from selection. If that works well, it would be sufficient for me (indeed, it would be superior because select, copy is quicker than right-click, select menu item, select submenu item, select, copy, close popup).

1 Like

One point favouring MathJax over KaTeX is that several SE sites already use MathJax; so if posts are imported to Codidact, some expressions won’t be rendered, or will be rendered incorrectly, with KaTeX. Another, less important, is that MathJax renders all AMS-LaTeX expressions used in the math environment—a set that many users are already familiar with. It would be a slight nuisance to have to familiarize oneself with the set KaTeX can render, or to try something one might expect to work only to find out it doesn’t. A third is that once the default MathJax is set up, it’ll (presumably) be easy to configure a new site to load any extensions its community want.

So. Consensus on this one? We try to implement MathJax before launching but it’s not an MVP requirement.

Speak now or forever hold your peace.

Just kidding. Speak now, though.

4 Likes

Does a MVP bike need gears?

If I take MVP literal then, no, it is not part of the minimum. We can work around it. But depending on who uses it, it would be very usefull. I would call it DesiredViableProduct.

The importance to include this depends on the community that is gonna use it. But if codidact is gonna do well then at some point there will/should be a codidact instance that may be gonna use (need) it. The question is how fast this is gonna happen in order that MathJax or KaTeX needs to be part of the first release.

In case people aren’t aware (I don’t recall anyone mentioning it), we do have browser extensions etc. available that individual users can use between MVP and whenever we actually implement mathjax (or equivalent), so I agree it definitely doesn’t need to be MVP.

On a related note, it would also be nice to have mathjax in chat but because of how we’re implementing things, I assume that’s no harder than putting it anywhere else?

https://meta.stackexchange.com/a/220976/349720

MathJax in chat, once we have it in Q&A is trivial.

But browser extensions? No. Browser extensions are great for a lot of things. But they are not good for content, such as including equations in a blob 'o text so that they look the same for everyone. In a small controlled environment (e.g., typically a small company internal system), that can work because you can limit use to specific browsers and mandate use of specific plugins for those browsers. (Though to be honest, even that is not a great idea any more as it gets messy the moment the boss decides to get an iPad, etc.)

3 Likes