Compare the FF MathML rendering:
with the SVG and PNG fallback image
for the example request
https://en.wikipedia.org/api/rest_v1/media/math/formula/102683d435b2bd37f214e2527e70098322c273ad
Physikerwelt | |
Aug 21 2016, 3:35 PM |
F4382214: 102683d435b2bd37f214e2527e70098322c273ad.svg | |
Aug 21 2016, 3:35 PM |
F4382212: mathml.png | |
Aug 21 2016, 3:35 PM |
F4382215: 102683d435b2bd37f214e2527e70098322c273ad.png | |
Aug 21 2016, 3:35 PM |
Compare the FF MathML rendering:
with the SVG and PNG fallback image
for the example request
https://en.wikipedia.org/api/rest_v1/media/math/formula/102683d435b2bd37f214e2527e70098322c273ad
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Resolved | Physikerwelt | T105310 Need a way to generate MathML for contour/surface/volume integrals | |||
Open | None | T143520 Improve mathoid font to support new integral symbols that work well in MathML, but have problems in the fallback modes |
As usual, this is purely a font issue (on both ends).
Since mathoid configures mathjax-node to use a math font that does not include glyphs for those unicode points, the rendering can be poor.
On a machine without compatible fonts, Gecko's rendering is just as poor.
@Pkra do you have a suggestion for a particular font? I know that there has been a lot of discussion and research about math fonts, but I honestly don't know the results. Is the documentation
Note that not all mathematical characters are available in all fonts (e.g., Neo-Euler does not include italic characters), so some mathematics may work better in some fonts than in others. The STIX-Web font is the most complete.
from http://docs.mathjax.org/en/latest/options/SVG.html#font still up to date or are there some tricks we could play to gradually improve the default font "MathJax TeX" to solve the problem at hand.
do you have a suggestion for a particular font?
I don't think I'm a position to make a suggestion. Too many factors come into play that are up to MW and WM.
from http://docs.mathjax.org/en/latest/options/SVG.html#font
See also http://docs.mathjax.org/en/latest/font-support.html
are there some tricks we could play to gradually improve the default font "MathJax TeX" to solve the problem at hand.
There are no easy tricks for contour integrals that are reliable, I think (but I might be wrong since my TeX hacking foo is limited).
There are at least three paths MW/WMF could consider: switching to a different font that supports glyphs, adding fonts, and extending MathJax's font fallback chain. All of these require varying amount of work (the first one primarily testing).
Maybe a virtual F2F of relevant people at MW/WMF would be helpful.