MathML is a failed web standard
peterkrautzberger.org
MathML is a failed web standard
1–10 of 177 posts
Re: MathML is a failed web standard
#2MathJax looks excellent.
Re: MathML is a failed web standard
#3Blargh. It annoys me that MathML was even a thing. We've already had a perfectly-fine and widely-used markup language for mathematical formulae; it's called TeX.
Ideally, I'd prefer that HTML5 include a tag, or something similar, that takes TeX as input and produces formatted output. A side benefit would mean that every browser, and every system, in the world would have TeX installed!
At least MathJax can take TeX as input.
Re: MathML is a failed web standard
#4One big advantage of MathML is the ability to copy-paste from webpages directly into software like Mathematica. I don't think there is currently any alternative to MathML in that respect. The author argues (rightly) about styling and presentation of MathML, but doesn't propose an alternative to this copy-paste problem. May be TeX, when you're using MathJax at least ?
Re: MathML is a failed web standard
#5Blargh. It annoys me that MathML was even a thing. We've already had a perfectly-fine and widely-used markup language for mathematical formulae; it's called TeX. Ideally, I'd prefer that HTML5 include a tag, or something similar, that takes TeX as input and produces formatted output. A side benefit would mean that every browser, and every system, in the world would have TeX installed! At least MathJax can take TeX as…
TeX is analogous to Presentation MathML but doesn't solve the accessibility concerns because the semantic intent of the TeX source isn't always clear. (How do you read a \phantom{}? Or how can you know that "\left( 7 \atop 3 \right)" should be read as "7 choose 3"?)
Re: MathML is a failed web standard
#6MathJax looks excellent.
MathJax is very flexible and its output is certainly high-quality. However, it is difficult to configure and package in my experience, and it's pretty slow.
We used it for a few years at Khan Academy before building KaTeX (https://khan.github.io/KaTeX/) which is around 50x faster to render in our experience, not even counting download size. We ran an A/B test for MathJax vs. KaTeX on the Khan Academy site and people completed measurably more exercises with the latter, even though they look identical. (KaTeX doesn't support everything MathJax does though – we still fall back to MathJax for a small fraction of our content.)
Re: MathML is a failed web standard
#7MathJax looks excellent.
MathJax is great, but you may also want to look at KaTeX, which is currently much faster:
Re: MathML is a failed web standard
#8"We need to get together with CSSWG/Houdini TF/etc to work out solutions that help those developers who actually solve the problem of math on the web."
Unfortunately, Houdini seems to have deprioritized font metrics info currently so this might not come soon. See this issue for more details:
https://github.com/w3c/css-houdini-drafts/issues/135
(However, for math we often use webfonts anyway so it is practical to send down glyph metrics with the fonts. That's what both MathJax and KaTeX do.)
Re: MathML is a failed web standard
#9Blargh. It annoys me that MathML was even a thing. We've already had a perfectly-fine and widely-used markup language for mathematical formulae; it's called TeX. Ideally, I'd prefer that HTML5 include a tag, or something similar, that takes TeX as input and produces formatted output. A side benefit would mean that every browser, and every system, in the world would have TeX installed! At least MathJax can take TeX as…
TeX is analogous to Presentation MathML but doesn't solve the accessibility concerns because the semantic intent of the TeX source isn't always clear. (How do you read a \phantom{}? Or how can you know that "\left( 7 \atop 3 \right)" should be read as "7 choose 3"?)
The same way a sighted person does?
Re: MathML is a failed web standard
#10Earlier quoted context omitted.
TeX is analogous to Presentation MathML but doesn't solve the accessibility concerns because the semantic intent of the TeX source isn't always clear. (How do you read a \phantom{}? Or how can you know that "\left( 7 \atop 3 \right)" should be read as "7 choose 3"?)
The same way a sighted person does?
Yes, you could try to build algorithms that turn that markup into semantic meaning but now we're almost talking about computer vision. I don't think a solution like that is likely to ever work reliably, and it's better to encode the semantic meaning so that it doesn't need to be re-derived.