12

One very enjoyable feature of mathoverflow is that the math just works. You enter it using the usual LaTeX, and then the jsMath magic does the rest. One of the most frequent activities for many mathematicians is checking abstracts on arXiv. So you open an abstract and you see all these dollar signs which you have to mentally decipher.

Question: Is there a way to set up jsMath, Firefox, etc., so that you can see formulas when viewing abstracts on arxiv.org?

This may be a meta question, but who reads meta threads, right? And there is a high concentration of experts here who may be able to answer this. That would improve quality of life for many mathematicians, I am sure.

Notes on the comments:

  1. Yes, people should make an effort and write abstracts on arXiv without using math. symbols, preferably. But the easiest thing to do, really, is to reuse your article abstract for the arXiv abstract, and that's what people do. And you can easily use some math symbols in your article abstract.
  2. Is it a question for arXiv people and not for MO? Maybe. But this is a very active community of people who encounter this problem on a daily basis. I hoping there is someone clever enough that solved this problem for themselves already.
VA.
  • 12,929
  • 5
    You should contact the arXiv people, really. – Mariano Suárez-Álvarez Mar 29 '10 at 01:25
  • Or the people at jsMath - http://www.math.union.edu/~dpvc/jsMath/ – François G. Dorais Mar 29 '10 at 01:29
  • -1 for not being particularly site-appropriate IMHO.

    Also: maybe I'm prejudiced by brief dalliance (in a former professional life) with web accessibility issues, but I'm not convinced there is a need to plaster JS cruft over everything. By their nature, abstracts cannot be too long, which usually limits the amount of LaTeX in them which one needs to parse; and if it looks like it might be interesting, then cutting, pasting and TeXing is not too arduous

     – Yemon Choi Mar 29 '10 at 01:38
  • I have no vote on this MO question, but being an old school kind of a guy, I think abstracts should involve symbols minimally, if at all. An abstract should make sense in the common tongue of the author. – Scott Carter Mar 29 '10 at 02:20
  • 2
    Well... hopefully, the author of a math paper speaks math-ese :P – Mariano Suárez-Álvarez Mar 29 '10 at 02:31
  • 3
    This is a very borderline case in my opinion, but I'm voting to close as 'off topic' on the basis that this question should be addressed to the arXiv directly. (Addressing it here will serve no purpose as far as I can tell.) I would be fine with a variation on the following instead. Is there a user-side method to render latex formulas on arbitrary websites? – François G. Dorais Mar 29 '10 at 03:11
  • 1
    (I hadn't seen your addition when I wrote the above. Your second item is along the lines of my proposed variation. I suggest you follow up with another question.) – François G. Dorais Mar 29 '10 at 03:14
  • 2
    I have not tried to do this myself, but one ought to be able to accomplish this with a Greasemonkey script. You might ask on meta.MO, simply because much discussion of jsMath already goes on there. – Reid Barton Mar 29 '10 at 04:17
  • 3
    Francois -- this post may be off-topic, but it may be helpful for the community, since it may attract the attention of those in charge of the archive. On the other hand I do wish there were an alternative to jsmath given the amount of bugs. – algori Mar 29 '10 at 04:17
  • The best way to attract the attention of those in charge of the archive is to contact them! – Mariano Suárez-Álvarez Mar 29 '10 at 04:35
  • 2
    I'd also suggest contacting the arXiv directly, via www-admin@arxiv.org, but I'm almost certain they won't be interested: they're very concerned by accessibility and bandwidth issues, and jsMath isn't exactly helping in these regards. – Scott Morrison Mar 29 '10 at 04:38
  • 1
    The bandwidth issue could be "solved" by having google incorporate jsMath in their Google AJAX Libraries API, so that they serve jsMath. This would probably not hurt MO, and people using it from the beaches of Copacabana! – Mariano Suárez-Álvarez Mar 29 '10 at 05:03
  • 7
    FWIW I've noticed that most LaTeX in ArXiv abstracts tends to have undefined macros in! That makes the problem much harder. You typically get things like "Consider the curve $ \C:y^2=x^3+1 $ . In \cite{BSD} it was conjectured that...". You're never going to be rendering this in your browser ;-) (unless your browser reads the preamble of the TeX source before processing the abstract!) – Kevin Buzzard Mar 29 '10 at 06:54
  • 2
    You can add the preamble into jsMath (in the macros section I believe). Bandwidth is not an issue - the browser downloads the library practically only once. And finally, making a Firefox addon that does this is simple, and I take on the challenge to do it during passover. – Dror Speiser Mar 29 '10 at 07:06
  • 1
    superuser.com is a better site than here for this question. – Theo Johnson-Freyd Mar 29 '10 at 14:44
  • Now on SO: http://stackoverflow.com/questions/2701841/rendering-latex-on-third-party-websites – Charles Stewart Apr 23 '10 at 21:40
  • This feels off topic to me, but I can't really put my finger on why it shouldn't be on MO. I've started a discussion at http://tea.mathoverflow.net/discussion/363 – Anton Geraschenko Apr 24 '10 at 05:48

1 Answers1

5

You may be interested in looking at the following greasemonkey script:

http://www.gold-saucer.org/mathml/greasemonkey/

It doesn't look as though it's been worked on recently but that doesn't stop it being usable. The basic idea is to overlay a simplified latex->MathML converter on top of a webpage. Since it is a greasemonkey script, it is entirely user-controlled and can be applied to any webpage, arxiv or otherwise.

Of course, there is always the problem of non-standard macros in abstracts and I completely agree with the sentiment that abstracts should be 100% legible without requiring extra parsing.

(In addition to posting this answer, I am add my vote to the list of "to close" as I agree with the various sentiments on that line in the comment thread. That I'm answering the question is not paradoxical since this isn't actually an answer to the question that was asked. I'm not convinced that I would want to see the question to which this is the answer also posted on MO, but think that the easiest way to prevent it being asked is to answer it here. I'm community-wikifying this answer not because I want others to be able to edit it, but to underline my opinion by forgoing any reputation for this answer (not that I assume it would actually get any!).)

Andrew Stacey
  • 26,373