7

Most of the criteria for closing questions (exact duplicates, non-questions etc.) are clear to me. Closing a question as being to localised is the one I don't think I understand. I suggested closing What could cause repeated NULL characters in an .aux file? in chat and in the comments Caramdir suggested that the question probably doesn't need to be closed. That's fine with me, but it raises the question of how do we decide this?

  • So what criteria should we use to close a question as "too local"?

Perhaps a related question (and I know I'm not supposed to do this) is:

  • What is the purpose of closing "too localised" questions in the first place?
8

I'll repeat what I've just said in 'chat' on this:

Closing a question as "too localised" says, for me, that here is a question that whilst technically on topic is so unique in some fashion that we cannot imagine anyone else ever asking it or needing to refer to it. And that it would be distracting to genuine questions or answers to have it exist on the site at the same level as other questions or answers (for example, it might keep coming up in the "suggested questions" list). Of course, it may come up again, and it was asked in good faith, so maybe we won't go so far as to delete it, but it certainly should sink without too many traces in the meantime.

  • this is true; it is also possible to have a question too localized in physical geography, though it is much more rare. (e.g. "any Tex experts in Greensboro, NC, USA?") – Jeff Atwood Mar 5 '11 at 21:50
7

Some criteria:

  • The issue cannot be recreated by other people (with a sane effort).
    This can be because a set of unknown conditions must be come together to cause this issue. For example it really only happens on one of the computers of the OP, but not on his/her others and he/she has no idea why.

  • The issue will only be an issue for a very limited time
    Things like: Can't install/update TeXLive because CTAN is currently down.

  • (to be continued)

5

Questions that are localized in time can also be considered as candidates to close. A made-up example:

What's the problem with <foo> in package <bar>, version <ancient|development>?

Ancient versions could be potentially acceptable, but questions on issues fixed in newer versions of the packages are not likely to be of much value to anyone but the original poster.

  • 3
    It's annoying that some people still use e.g. TeXLive2009 or sometimes even older packages. If the issue is fixed in the current version of that package using it should be the correct answer. Other people might have the same issue with the same old version, so why not keep it around. Really hot develop versions should be really considered as 'to localized in time'. An different thing are almost-stable commonly used but not yet officially released versions of e.g. PGF/TikZ. There v2.0beta was around for quite a time. – Martin Scharrer Mar 5 '11 at 23:03
  • 4
    @Martin: I agree with you, but sometimes people do have legitimate reasons to keep older packages (institution-mandated TeX installations, for example). However submitters need to be informed that potential responders would see little incentive to provide any support other than "update to the latest version". – Martin Tapankov Mar 5 '11 at 23:30
  • @MartinScharrer: To corroborate what Martin Tapankov wrote, at our institution we got (maybe half a year ago) our first ever update in years, and this was TeXLive2009! – Hendrik Vogt Mar 6 '11 at 7:14
  • @Hendrik: I understand, but it is not difficult to have your own updated texmf tree with the packages you need in a newer version. Of course if your institution or publisher insists on a certain old version you are out of luck. In this case this should be mentioned in the question. – Martin Scharrer Mar 6 '11 at 10:53
  • @MartinScharrer: You're right, but how many people will bother to do that? – Hendrik Vogt Mar 6 '11 at 11:40
  • @MartinScharrer: e.g. Ubuntu 10.04 comes with TeXLive2009, and requires manual installation for TeXLive2010. – Bruno Le Floch Mar 9 '11 at 11:48
  • @Bruno: But manually installation is very simple (telling Ubuntu to not install its outdated .deb packages is the difficult part). I have no idea why they not use TeXLive2010. There can't be so much binary dependencies and Ubuntu shouldn't care at all about LaTeX package versions. – Martin Scharrer Mar 9 '11 at 11:53
4

Just posting this as an answer so that it can be voted on:

I think that of the possible close reasons, "too localized" is the one we need least on our site. Even if the question mentioned in the OP is rather localized, I still don't see any harm in having it around. Why? The question is formulated so that I can understand it quite well, the OP gave an understandable answer, the post as a whole look quite clean, and in fact much better than other possibly more valid questions with several answers where the OP was misunderstood.

  • Jeff just gave an example of a question where I would indeed vote to close as "too localized". I'm waiting for the day that it happens :-) – Hendrik Vogt Mar 5 '11 at 21:59
  • Yes, I think this is the issue that prompted my second question. There may be very little need to close such questions generally. So the reasons to close are a kind of heterogeneous set with respect to the potential negative impact on the site. – Alan Munn Mar 5 '11 at 22:33

You must log in to answer this question.

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