Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.
MathJax 2.2 beta is now out http://www.mathjax.org/mathjax-v2-2-beta-now-available/, with the final release in a few weeks. Notably, it includes support for AMS commutative diagrams http://ctan.unsw.edu.au/macros/latex/required/amslatex/math/amscd.pdf.
Our current MathJax setup doesn't automatically upgrade to new versions (and is still back at 2.0). Two questions:
It seems we are a pretty good candidate site for trying out their beta for them --- we have lots of eyes, people care about the typesetting looking right, and there's a reasonable chance that if problems are noticed feedback would at least reach meta. On the other hand maybe it's not worth the bother, or on the whole we'd prefer the safety of a stable version.
I certainly think automatically updating to the latest stable version makes sense. Probably it's best if there's a way for you or Anton to manually rollback updates in case something breaks dramatically.
I have no objection to using MathJax 2.2 immediately. I will say that the amscd documentation could include, you know, some slightly more readable discussion of what the commands are.
You (Scott Morrison) are better placed to tell, but if we migrate to SE2.0 soon(?), as far as I understand, this will be obsolete (as we'd use "their" MathJax setup). Then, perhpaps to have the same version as there (I do not know which one that'd be) could be a plus, to avoid having a potential additional issue in the migrating process.
Quid, don't be such a pessimist! For something as important as MathJax, we won't let anyone dictate what is best for us.
The question is therefore: Which MathJax version is best for us?
The MathJax people don't appear to think using the beta version is a great idea for large sites like ours. I am generally more adventurous and I think out site would provide a great beta testing ground for MathJax. I'm also very much looking forward to commutative diagrams on MO!
@François G. Dorais: While I think I have to admit to rather being a pessisemist, I did not mean to make a pessimistic statement here, which then in turn sort of confirms the first part. Only, I wanted to caution--possibly unnecessarily so--against making an effort related to something that soon is not (necessarily) relevant anymore. [Off-topic but to avoid a misconception: I am rather looking forward to the move.]
... and we're working on catching up :-)
By the way, there's also a pretty amazing plugin XyJax that allows you to use most of xypic within MathJax. Personally, I don't use xypic (I was never able to make the syntax stick in my head between attempts), so it's maybe not that exciting.
I think we should be very cautious about deploying 3rd party plugins, which will likely fall by the wayside, leaving us with unrenderable posts. But maybe it's worth thinking about.
We're now on 2.2-beta. Bug reports welcome here or in the main MathJax sticky thread. (Try restarting your browser first; there are often transient issues because your browser tries hard to cache javascript.)
Does someone want to try an AMScd commutative diagram, and report back?
It doesn't work ( http://i.imgur.com/5PmOG1y.jpg ), or should I kill all my browser caches with fire?
EDIT: Oh, it just seems to take half a minute to load (on FF and on Chrome likewise). This doesn't look very good...
I'd also like to report that the rendering is currently VERY slow.
Alberto and stankewicz: is rendering equally slow on math.stackexchange? I'm curious if the problem is with MathJax itself, or with something about how it is configured/included on MO.
Ok. Given that using 2.2 is causing problems for some people, and migration to SE 2.0 is imminent, I've rolled back to using MathJax 2.0 with our previous configuration for the time being.
1 to 17 of 17