22
$\begingroup$

Should question titles, which sometimes intentionally avoid LaTeX math for those who choose faster rendering, be edited to insist upon LaTeX math? A recent instance that prompted my question is "two tetrahedra in R^4." The original title used R^4 to mean $\mathbb{R}^4$, and it was edited, bringing a 1.5-yr-old post to the front page. I don't object to the editing—I just want to know what the community considers appropriate.

$\endgroup$
  • 5
    $\begingroup$ As some meta.MO users will know, I am strongly against such edits in titles, and against them by and large in the body of text. Each insertion of mathbb makes me want to change it to mathbf, more often than not $\endgroup$ – Yemon Choi Aug 6 '13 at 22:00
  • 3
    $\begingroup$ As part of question initiated by @YemonChoi on this general matter I specifically carried out a "poll" on almost exactly this meta.mathoverflow.net/questions/169/… at this moment 8 against 7 for such an edit (excluding my personal opinion, which would be against, but also see my conclusion). [Some of the comments there started to be on tagging, you can ignore them.] $\endgroup$ – user9072 Aug 6 '13 at 22:08
  • 8
    $\begingroup$ I rejected the edit, as too minor, but so it goes. Ridiculous that the body of the question, where R^4 was also used in many places, was not touched. If one is to make an edit on this question, the reasonable thing is to edit it fully, adding latex everywhere it is missing, and being consistent with the use of notation. (I do not have time for that right now, as it turns out, so I am not going to "improve it".) $\endgroup$ – Andrés E. Caicedo Aug 6 '13 at 23:39
  • 8
    $\begingroup$ I reverted the edit before seeing this thread. I hate unnecessary latex in titles because it makes the site slow on weak wifi or cellular connections on mobile devices. And it makes no sense to do it for long forgotten posts. $\endgroup$ – Benjamin Steinberg Aug 7 '13 at 1:22
25
$\begingroup$

Unlike Benjamin Steinberg, I have no objections to ${\rm\LaTeX}$ in titles of questions. I often use the site (even from mobile devices) where there is a sufficient speed for loading everything fairly fast.

I do, however, object to making this sort of edit to an old post. If you are going to improve the post, it should be an overall improvement, perhaps to another answer or so. But since a thread is bump, I expect a correlation between how old the thread is and how significant the improvement is, and in this case I don't think it's enough to merit such an edit.

$\endgroup$
  • 2
    $\begingroup$ -1 for 1st para and +1 for 2nd para :-) $\endgroup$ – Yemon Choi Aug 8 '13 at 0:16
  • 6
    $\begingroup$ +1 for first paragraph, and -1 for second. :-) $\endgroup$ – Joel David Hamkins Aug 8 '13 at 11:32
  • 1
    $\begingroup$ @Joel: My point in the second paragraph is the same as Andres' in the comments. If you are going to make an edit to an extremely old post, at least make it a good one. Not just one detail in the title. $\endgroup$ – Asaf Karagila Aug 8 '13 at 11:34
  • 18
    $\begingroup$ I appreciate any kind of improvement, even small ones. I like to see old questions resurrected occasionally---I enjoy reading them---and I don't mind if a lot of questions get bumped. I think we could be more tolerant for what are at bottom good improvements. $\endgroup$ – Joel David Hamkins Aug 8 '13 at 11:38
  • 1
    $\begingroup$ The point with the thread in question was that this was not really a good improvement, however well intentioned. Instead, it added noise (in the form of a clash of notation). $\endgroup$ – Andrés E. Caicedo Aug 12 '13 at 21:02
  • 9
    $\begingroup$ Joel, I think you meant (-: $\endgroup$ – Allen Knutson Aug 16 '13 at 19:45
  • $\begingroup$ I just made an edit to bump. Also to fix something, if one wants to take that point of view :) $\endgroup$ – Yemon Choi May 25 '16 at 17:02
  • $\begingroup$ @Yemon: I'm amused that you put braces there. :-P $\endgroup$ – Asaf Karagila May 26 '16 at 7:33
17
$\begingroup$

No, they should not. Many of us habitually read MO on mobile devices with unreliable wifi/cellular connections and if most titles take a long time to render it makes the whole experience irritating. Latex should only be used in titles when really necessary. In fact I think one should in general be a latex minimalist in questions and answers too.

$\endgroup$
  • 18
    $\begingroup$ I take your point, Ben. But equally, many of us habitually read MO on reliable high-speed broadband, and so prefer to see properly typeset mathematics rather than pseudo-tex code. I don't think users should feel inhibited from using Latex in their questions and answers, but I think it would be a sensible compromise to be very sparing with it in titles, for the reason you give. After all, how many of us register the difference between R^4 and $\mathbb{R}^4$ anyway? $\endgroup$ – Tom Leinster Aug 7 '13 at 15:07
  • 3
    $\begingroup$ @TomLeinster, if people want to make their title in latex, I won't go and deliberately deTeXify it, but my gripe is when people TeXify a readable title that may have been purposely $deTeXified$ to be readable quickly. $\endgroup$ – Benjamin Steinberg Aug 7 '13 at 15:41
  • 1
    $\begingroup$ @WillieWong, you are right. I just tried it. $\endgroup$ – Benjamin Steinberg Aug 7 '13 at 15:54
  • 8
    $\begingroup$ @TomLeinster I'm afraid your comment reads a bit as "It works for me so what's the problem?". Even on a decent speed broadband, I get a flicker as MathJaX renders it from TeX to MathML which is very annoying, and the page adjusts as the rendering occurs which makes it difficult to quickly scan through the question titles, so my vote is for no TeX in titles at all. $\endgroup$ – Loop Space Aug 7 '13 at 18:45
  • 5
    $\begingroup$ @AndrewStacey: well, Ben was arguing that the community should adopt a certain behaviour because of his own personal way of using MO, and I was doing just the same. I think we're both being completely reasonable. Surely lots of people do read MO on flaky connections, and surely lots of people read it on good university connections. In the absence of data, we can only relate our personal experiences, express personal preferences, and make educated guesses as to what others want. $\endgroup$ – Tom Leinster Aug 7 '13 at 19:17
  • 6
    $\begingroup$ @TomLeinster The problem with that is that the two extremes are not balanced. If there is no MathJaX in titles then people on fast connections are hardly inconvenienced at all, but if there is a lot of MathJaX in titles then people on slow connections are greatly inconvenienced to the point that the site becomes unusable. I remember when jsMath was first put on and I was on a very slow connection in Rio and MO became unusable because I couldn't get past the question page. So I don't think you're being equally completely reasonable! $\endgroup$ – Loop Space Aug 7 '13 at 19:50
  • 4
    $\begingroup$ @AndrewStacey I essential agree with you (and on my notebook the question-page regularly "freezes" everything, for quite a while), but also note that Tom Leinster suggested to be very sparing with TeX in titles, so I feel his position is rather balanced. $\endgroup$ – user9072 Aug 7 '13 at 20:01
  • 1
    $\begingroup$ I agree with quid that my position is rather balanced :-). I'm suggesting using Latex very sparingly in titles, after all. $\endgroup$ – Tom Leinster Aug 8 '13 at 0:21
  • 1
    $\begingroup$ Anyway, Andrew, I bet your connection's faster than mine (which is nothing special). I think this is also to do with personal preference: would you accept a bit of delay and flicker if the reward was properly typeset maths? I suspect you wouldn't, and I would. $\endgroup$ – Tom Leinster Aug 8 '13 at 0:33
  • 1
    $\begingroup$ Now that we have SE 2.0, is it possible to implement a feature that de$\LaTeX$ifies titles? $\endgroup$ – Tom LaGatta Aug 8 '13 at 6:43
  • 1
    $\begingroup$ I am not well-informed on technicalities here, but I think--mainly from observation how my devices behave--that the quality of the internet-connection is not the only or even the main issue. Rather the page when there is a lot of MathJax seems pretty demanding on the hardware. $\endgroup$ – user9072 Aug 8 '13 at 18:36
  • 1
    $\begingroup$ @TomLaGatta I am not sure how to interpret the 'now' or what you mean precisely, since before the move MathJax rendering in titles (at least in lists) was optional with default off. But I agree(?) it would be nice to get this back. I actually asked for this before the move, if it were possible; likely fell a bit through the cracks, maybe worth reasking.) $\endgroup$ – user9072 Aug 8 '13 at 18:40
  • 5
    $\begingroup$ I would recommend against using HTML to format mathematics (like <b>R<b><sup>4</sup>), as this loses the semantics of the mathematics. That makes it harder to programmatically recognize the mathematics within the page. That makes it harder for things like screen readers (e.g., ChromeVox) to properly speak the mathematics, and makes it harder to copy and paste the math into CAS like Mathematica, and harder for search engines to properly identify the mathematics, etc. Although I understand the need for practicality, it is also important to consider possible reuse of the page in the future. $\endgroup$ – Davide Cervone Aug 12 '13 at 12:03
  • 3
    $\begingroup$ Also, we are looking at speed improvements for MathJax (and mobile devices and networks also increase in speed), and in the future these short-term solutions may become liabilities (consider the double-backslash issue currently being dealt with). I think it is better to write what you mean than to fake it to get around current technical limitations. $\endgroup$ – Davide Cervone Aug 12 '13 at 12:06
  • 2
    $\begingroup$ @YemonChoi, nothing has happened to that, but I'm you to not forget that your pages are used in other ways as well. E.g., some of your human readers may need assistive technology (like screen readers) to "view" your answer, and in that case, marking the mathematics as mathematics (as opposed to just some additional text) can make a difference in their ability to understand what you have written. A visual approximation is not always good enough. ... $\endgroup$ – Davide Cervone Aug 19 '13 at 16:21
11
$\begingroup$

Aside from rendering speed, there are other reasons why MathJax is undesirable in titles.

  • Titles are also displayed in places where MathJax does not render, for example, in the browser's title for the page, and in the lists of questions in the StackExchange apps.
  • The URL is created from the title and then will contain it with some symbols removed.
  • MathJax titles make searching more difficult, and might affect search engines negatively due to the relevance of the title and URL.

More generally, a title heavy on formulas is more often than not badly chosen, in my opinion.

$\endgroup$
  • 1
    $\begingroup$ I'm not sure it's true about the search engines. I have in the title of my current window the plain MathJax code. Certainly Google's crawler sees the same. $\endgroup$ – Asaf Karagila Aug 7 '13 at 18:55
  • $\begingroup$ I said that some characters are removed for the URL, for the title of the page I said is is does not render, which you confirm. And, I still believe that in general a couple of words are better for searching than some symbols and/or their corresponding source (not last since often there are somewhat different ways to get more or less the same visually). $\endgroup$ – user9072 Aug 7 '13 at 19:06
3
$\begingroup$

This is only marginal to the discussed issue (but I still think it is worth mentioning).

There were some discussion on this meta about possibility of not showing questions from MO in the network wide hot questions list: Measures to separate math overflow from the rest of the stack exchange network (Related: Featured MO questions on the hot list: what benefits, if any, do these bring?) This suggestion received some support, but it is not very likely that this is going to happen anytime soon.

Adding MathJax to the title prevents that particular question from being shown in the hot network question list: Are the questions with TeX markup in titles excluded from Hot Network Questions? So some users can consider this as an advantage of having LaTeX/MathJax in the title. (Depending on their position about the issue of MathOverflow and hot network questions.)

$\endgroup$

You must log in to answer this question.

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