Today as I was answering a question I noticed that $\oiint$ doesn't work in Physics SE. Is it some kind of a bug or am I missing something obvious?
-
2$\begingroup$ Possible related post on meta.math: meta.math.stackexchange.com/q/9973 $\endgroup$– dmckee --- ex-moderator kittenMay 10, 2015 at 21:10
-
3$\begingroup$ It seems people have managed in various creative ways! $\endgroup$– Qmechanic ModMay 10, 2015 at 21:25
-
$\begingroup$ @Qmechanic That is utterly unbelievable! $\endgroup$– GonencMay 10, 2015 at 21:30
-
$\begingroup$ @Qmechanic - IMO, the OP of that question deserves being upvoted for the sheer effort he/she put in there! (I have, FWIW). $\endgroup$– 299792458May 11, 2015 at 9:07
2 Answers
I've noticed this as well. As stated in the answer to the Math Meta post dmckee linked, MathJax simply does not support this symbol. You could consider it a bug in MathJax if you want - or more accurately, a feature request - but it's not something for Stack Exchange to handle. So the place to report this would be on the MathJax website.
Unless and until this is implemented in MathJax, I would recommend just using \oint
, which gets the point across. It's quite common anyway for physicists to use a single integral symbol regardless of how many dimensions the integration region has.
-
$\begingroup$ This has been reported and it is documented in the MathJax github page as issue #566 - apparently it comes down to the symbol not being available in whatever fonts MathJax is using. If people want to apply pressure, that issue is probably the place to go. $\endgroup$ Nov 1, 2016 at 9:30
This was reported as issue #566 of the MathJax github page back in 2013, but I just re-raised it to see what they'd say. The devs indicated that they are working on the fonts for MathJax v3.0, which should arrive next year, and this might solve the problem.
In the meantime, they suggested that we add a custom macro to the MathJax configuration for this site, i.e. something of the following form:
You could use the equivalent of
\def\oint{\mathop{\vcenter{\mathchoice {\huge\unicode{x222E}\,}{\unicode{x222E}}{\unicode{x222E}}{\unicode{x222E}} }\,}\nolimits} \def\oiint{\mathop{\vcenter{\mathchoice {\huge\unicode{x222F}\,}{\unicode{x222F}}{\unicode{x222F}}{\unicode{x222F}} }\,}\nolimits} \def\oiiint{\mathop{\vcenter{\mathchoice {\huge\unicode{x2230}\,}{\unicode{x2230}}{\unicode{x2230}}{\unicode{x2230}} }\,}\nolimits}
to allow
\oint
,\oiint
, and\oiiint
to work with their respective unicode characters, with sizing reasonably for displaystyle, and being the proper TeX class to allow limits to be placed appropriately (though the lower limit will be a bit too far to the right, because with won't get the proper italic correction).
This seems excessive to me - there's no need to be hauling around that much specific TeX in the html header for every page on the site over such a seldom-used symbol. If people want to have that solution, they can just use the unicode directly:
$$
\mathop{\vcenter{
\huge\unicode{x222F}\,
}}
$$