3

For some time there has been a thread on meta discussing the desire for MathJax support on Cogsci.se in order to support the display of mathematical equations in questions and answers.

Recently I received the following post from Rebecca from StackExchange:

Enabling MathJax supports means that on every single page on this site, the MathJax scripts will be included, and they are not small. Due to this, we want to see that MathJax is being used significantly on the site - otherwise, we are just needlessly slowing down the user experience on the site.

In the list of questions included here, I see stuff that would use MathJax in only a couple of the posts. Poking around the rest of the site, I'm seeing it used extremely minimally.

Given the extremely small use I see on the site in the current questions, I'm not sure I see the benefit in enabling MathJax.

I can appreciate that:

  • There is a trade-off between the user experience (i.e., the time for a page to load or appear loaded) and functionality (i.e., supporting MathJax in order to permit display of mathematical equations).
  • It's ultimately stack exchange's call as to whether this trade-off is worth while.

That said I think a few points should be mentioned

Equation support is important to this site

  • I'm not sure if Stack Exchange fully appreciates the importance of mathematics in the disciplines of psychology and cognitive science.
  • There is a bit of a Chicken and Egg things going on. The capacity to display equations will encourage users to incorporate equations into answers. It's difficult to demonstrate the importance of equation support when the site lacks support for equations.
  • The site also has aspirations of scientific credibility. It seeks to engage academics and competent researchers. Equation support is both a differentiator with other forums and also reflects the standards of rigour which we hope will appeal to our target contributors.

Reflections on load times with MathJax

  • I find the user experience on http://stats.stackexchange.com (a MathJax enabled site) to be fine. Pages seem to load quickly.
  • Prof David P. Cervone has provided an extensive post on meta about the actual page loading implications of the MathJax script. By that analysis the effect on page loading appears to be fairly minimal and mostly related to pages that use mathematical equations. And surely questions that require equations can justify the additional load time.

Thus, I am aware that there are trade-offs, but I'm hoping to have a dialogue both with users of this site and StackExchange about whether all things considered the trade-offs justify inclusion or not of Mathjax.

Questions

  • How can the importance of mathematics to the site be effectively communicated to the StackExchange team?
  • Does the StackExchange team appreciate the importance of mathematics to the site?
  • Can the StackExchange team communicate more clearly the performance cost in terms of site load times and user experience of MathJax support and why the trade-off is not worth it?
  • Finally, what do other users of the site feel about the relative importance of mathematics support and page load times?
2

This is a great question.

I totally agree with the chicken-and-the-egg problem. I know that this site has no MathJax support and so I don't even try to write equations, because it is so painful to try to write math without MathJax when you are used to having it (as on cstheory.SE, cs.SE, math.SE, stats.SE, scicomp.SE... etc etc etc). I also avoid writing certaion kinds of questions on this site, because I don't know how to formulate them without math. I feel others who have technical questions, might feel the same way. Without these technical questions being asked (something we desperately need) we are not generating the need for MathJax in answers.

I am more than willing to sacrifice a little bit on load times for the potential benefit of more technical activity on this website. Further, I agree that this is an essential differentiating feature for a scientific site that wants to be taken seriously.

2

I have to say, Rebecca confuses me greatly:

Due to this, we want to see that MathJax is being used significantly on the site - otherwise, we are just needlessly slowing down the user experience on the site

...Uhhh, we don't have MathJax, so of course it's not being used!

I'd like to ask the Stack Exchange team to reverse their position: give us MathJax support now, then if we haven't used it enough, then discuss taking it away. Otherwise how are we to know if "MathJax is being used significantly on the site"?

Also I fundamentally disagree with their assertion that the MathJax scripts [...] are not small. [...] we are just needlessly slowing down the user experience on the site." In my testing the MathJax scripts do not cause any human noticeable differences in load times.

Here is what Chrome's inspector shows for the loading times on one of our questions:

CogSci Load Time

Total load time for that question on our site was 767 milliseconds.

Here is what Chrome's inspector shows for the loading times on Cross Validated:

CrossValidated Load Time

That's 878 milliseconds. Less than a second to load the page. Only ~120 milliseconds longer than our site!!! And this is from a residential cable connection (Where I am only getting 30% of the speed I am paying for, but that's another matter):

Speed Test results http://www.speedtest.net/result/1876874382.png

OK. So according to SpeedTest.net my connection is faster than 87% of others in the US. However most other users of this site will be using connections similar to mine. Many ideal users will be students or professors at a university, and will in fact get much better internet speeds than I do.

I see no validity to the claim that MathJax support will cause too much load on our site.

  • Times of under 100 ms are recognized as "instant" but since the total time is over 100 ms extra load time DOES have an impact. The raise in bounce rates is most noticeable during the first second of load time – Ben Brocka Apr 5 '12 at 14:11
  • Notice how on the graph you just linked to the difference between 0.7 and 0.8 seconds is next to nothing? Probably a 1% increase at most and I doubt that chart is accurate to that degree. – Josh Apr 5 '12 at 14:26
  • Notice how the majority of US users have 1/6th your internet speed or less? Stop focusing on 100ms, it's not the only possible result. – Ben Brocka Apr 5 '12 at 15:06
  • 2
    The thing I like about the discussion (between Ben and Josh) here is that we are adding some data to the question of the trade-off. – Jeromy Anglim Apr 6 '12 at 7:32

You must log in to answer this question.

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