29

The main Q&A on this (here's one, and another one, and of course this one all seem to be 3+ years old. I'd accept that this question is a duplicate of this question which seems to have comprehensively covered all the bases last year.

I'm not asking why we don't have it. That seems to have been covered a lot. I am asking are we at the stage where we can at least start talking about it? Here are a few examples of posts that would benefit from MathJax (hurriedly hashed together so they may not be the best) are (and feel free to add your examples to this list):

  1. Answer to: Determine if two rectangles overlap each other?
  2. Question: Gram Schmidt with R
  3. Question: Speeding up Julia's poorly written R examples
  4. Question: Ukkonen's suffix tree algorithm in plain English?
  5. Question: How to pair socks from a pile efficiently?
  6. Answer to: Plot a heart in R
  7. Answer to: Cannot rotate from one specific orientation to another
  8. Answer to: Creating a rotation matrix based on two vectors

Surely WebKit and native rendering of MathML have come a long way since 2009 and we might be at the stage where it is feasible to at least start discussing the inclusion of math markup on Stack Overflow? Are those really heavy dependencies still necessary? Is it still better to carry on uploading images where necessary? Every time this has been bought up, it's been a blanket NO.

One of the reasons frequently trotted out is that you shouldn't need math markup on a programming site or it's not often needed. Without any support for math markup how can you guage how much interest there is in using it to help illustrate questions or answers?

I'd love this post not to get closed and to generate some fresh debate given advances in native rendering of math markup in browsers.

Here's hoping.

11
  • 12
    I can see how reading "Before you flame me and close this", would instantly trigger a user's defensive flame and close response. Giving readers a little bit of credit and not insulting their intelligence goes a long way. +1 for the proposal though.
    – user200500
    Mar 27, 2013 at 1:40
  • 1
    And also suggestions on how it wouldn't have a real overlap with mathematics.SE. meta.stackexchange.com/questions/47107/… Mar 27, 2013 at 1:42
  • 6
    Those of us in the math-centric DSLs (e.g. R, Matlab, Maple, Mathematica, a certain subset of Python users) would probably really find this useful. My guess is that we don't constitute a large enough subset of the SO user base to get much traction on this issue, though.
    – joran
    Mar 27, 2013 at 2:29
  • 7
    @joran A lot of people in almost all 3D related tags frequently need to talk about transforms and matrices, so it helps there as well. This is actually quite broadly applicable.
    – user200500
    Mar 27, 2013 at 2:42
  • @Asad Good point. I hope that makes this more likely to be implemented.
    – joran
    Mar 27, 2013 at 3:48
  • I suppose that "I participate in CS.SE and Math.SE, and sorely miss MathJax on SO" doesn't cut it...
    – vonbrand
    Mar 27, 2013 at 4:05
  • 3
    For the record, MathJax does substantially slow down those sites that have it, so I would only consider it if it adds significant value.
    – user102937
    Mar 27, 2013 at 4:23
  • 9
    There could be a magic comment such as <!-- enable-mathjax --> that would enable it for a particular post, so that pages where it wasn't required wouldn't be slowed down.
    – Jeremy
    Mar 27, 2013 at 4:31
  • 2
    Or it could be reserved for certain tags, or activated by a mod after flagging, similarly to CW for questions (not sure if this last one is such a good idea, but just putting it out there)
    – user200500
    Mar 27, 2013 at 6:07
  • I assume that a downvote on this question should mean we are NOT at the stage where we should talk about including Math ML, or does it mean this is a stupid question? Mar 27, 2013 at 7:50
  • @SimonO101 Voting works differently here on meta. So the downvoters probably think support of math markup is unnecessary. Mar 27, 2013 at 8:53

1 Answer 1

3

Of your 8 examples:

  • I don't see the need for mathematical markup for 1, 2, 3.
  • 8 is squarely off-topic, it's a math question. There's a site for that.
  • 4 and 5 are algorithm questions. They're grandfathered in on SO by tradition, but they're not programming questions. I'd vote to close them as off-topic if I caught them coming in, but I prefer to let old borderline questions stand and only close the strongly off-topic ones. There's a proper site for algorithm questions now.

This leaves 7 and 6 as questions which are within SO's scope and would benefit from math markup. I don't think the need is common enough to warrant adding that feature to SO.

When a question is borderline between SO and CS.SE, I find that a good guideline for choosing which site has the most appropriate audience is whether it would benefit most from code markup or math markup. This isn't foolproof, but it's good enough that I think we're giving the right tools to the right audiences.

And yes, MathJax is still slow for some people.

6
  • seeing the need is somewhat subjective and whilst I greatly respect your opinion I disagree with it (aside from for 1, but then I did leave a disclaimer). Trying to follow the algorithm in 2 is difficult and would be aided by a formula. 3 is perhaps the best example of where a math formula would be ideal. The Question is about the best implementation of a particular algorithm. Having the mathematical definition of the algorithm in front of you allows those implementing new versions to work off the definition rather than suboptimal code of what the algorithm should do. Mar 27, 2013 at 11:25
  • I also take you point regarding 4 and 5. Still, these are 8 examples from the whole of SO, and without the support for it, need can't truly be gauged. I am very willing to accept that if it places too much load on SO then so be it, but I was under the impression that for instance Math ML could be rendered natively by the client-side browser and would not place a load on the system. Unfortunately I am not a web-dev and would happily accept that I am talking rubbish, but I just thought that as the technology has matured it might now be an option. Mar 27, 2013 at 11:28
  • @SimonO101: You may not be a web developer, but I would hope that you could use Wikipedia. There, you see that native MathML support is basically: Firefox, yes; Chrome, pre-beta; Opera, mostly; IE, no. Mar 27, 2013 at 11:34
  • 4
    @NicolBolas The sarcastic tone is quite unnecessary. Yes I can use Wikipedia. What is your point? Because IE doesn't support it the feature should not be available? Websites have served up different versions of a site for different browsers since forever. Where is the incentive for all those IE devs posting on SE to incorporate support to remain competitive. And, if you read the piece in any depth, you would see that support for Math ML in IE is as simple as installing a plugin. Mar 27, 2013 at 11:47
  • 1
    @SimonO101: My point is that you should do at least basic research into the current state of native MathML client-side rendering. I see the most popular browser having zero support for it, and I see the third-most-popular browser not having solid support for it. As for "installing a plugin", I'm not sure what your definition of "native" allows a plugin to be native to a browser. Also, if a post is unintelligible or significantly corrupt without some plugin, that's not a good thing. Are you really suggesting we ask people to install security holes in their browser just for MathML support? Mar 27, 2013 at 11:54
  • Sounds like the solution might be a MathML parser with a reasonable fallback for IE, whether that's a server-generated image, or something better. Apr 3, 2013 at 22:45

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .