I'm really happy overall about the changes to closing. I think the new close reasons are much less full of jargon than the old reasons.

In practice, however, I'm seeing an actual issue with one specific close reason: The custom off topic close reason.

The specific issue I'm seeing is that users are using a custom version of that close reason to close questions that should really never be closed for the reason they're giving. For example, in this question, three members of the community came up with the following custom off-topic reason:

custom off topic close reason

Other: This question appears to be off-topic because it is about something that is to be read in a beginner's guide

I have a few issues with this close reason:

  1. There's no evidence given to support the claim made; the 'stock' off topic reasons all provide clear evidence as to why something is off topic, and action to take.

  2. It allows five people to close for whatever reason they feel like, so long as they can get 4 other people to listen to them. It provides an close reason that is not only going to be inconsistent, but has the potential to turn new users off because 'our rules keep changing'.

  3. Grammatically speaking, it doesn't flow.

We have well defined reasons to close, we should use those. We have well defined reasons to vote down a question, we should use those too.

In this case, I believe this question should be downvoted, but not closed. In the general sense, allowing a custom-end-user-filled-in close reason is dangerous, has the potential to splinter the community, and to turn off new people from asking their questions here.

If you agree (or especially if you disagree), please let me know why. Consider this a discussion about this issue and a feature request to have it removed (based on community consensus, of course).

  • 2
    I did state some very similar concerns when this was originally announced: meta.stackexchange.com/questions/184167/… Commented Jun 26, 2013 at 12:37
  • 3
    I'm concerned about removing the Other option before it is determined what to do with the Too Localized Q's that don't currently fit into any of the existing Off Topic categories. (But generally, I agree, the custom close option seems like a recipe for disaster.)
    – joran
    Commented Jun 26, 2013 at 13:24
  • 2
    @Lamak I don't think that should be closed; even though that instance may seem 'too localized' the existence of an error message that is commonly received is not too localized, and this shows why that error message happens. So more generally, it's a good question with an answer that will help future people. "Oh, I see this error message, it means I have a typo, and the answer shows where the typo would be in their instance, so I need to use that to see where the typo is in my instance." Commented Jun 26, 2013 at 16:09
  • 7
    @GeorgeStocker also note that the exact same typo could lead to dozens of different error messages, and that error message could also be caused by a completely different symptom. So I disagree that it will help future people.
    – Aaron Bertrand Staff
    Commented Jun 26, 2013 at 16:29
  • 5
    @George right but ones like these will not create enough breadcrumbs. They could also be closed as duplicates of dozens of questions that were previously closed as too localized. The consensus outside of this question, at least on dba.SE, seems to be that many silly typo questions that should be closed, and have always been closed, will now remain open, and those questions will create a lot of noise.
    – Aaron Bertrand Staff
    Commented Jun 26, 2013 at 16:32
  • 3
    @GeorgeStocker perhaps the crux of the problem is that everybody is tied to their own definition of "too localized." A lot of these questions that previously fell into the TL bucket are actually "too narrow," "too simple" or "RTFM."
    – Aaron Bertrand Staff
    Commented Jun 26, 2013 at 16:36
  • 2
    Also I believe my answer would be specific to the site where I'm a moderator (dba.SE). We certainly don't want to lose the ability to still use variations on TL via custom off-topic, regardless of what path is chosen for SO.
    – Aaron Bertrand Staff
    Commented Jun 26, 2013 at 16:38
  • 1
    @AaronBertrand "RTFM and Too Simple" have never been a reason to close on Stack Overflow -- partially because they turn away new users who don't know any better. There is definitely a force majeure that wants that to be the case, but that will hurt our cause and make us seem more insular than we already do. Commented Jun 26, 2013 at 16:49
  • 2
    @GeorgeStocker Okay. So, that was me. And in my opinion, there are RTFM questions, this particular one being one of them. I'm really sorry for the grammar, and yes, it's quite nonsensical a sentence to be honest, but where else to put this? As "too localized" was removed, most say that those should be marked as "off-topic". Since no other option in the off-topic category fits RTFM, I have opted to leave a (slightly more polite) comment and custom reasoning. (Apparently, at least 4 more users agreed with me.)
    – H2CO3
    Commented Jun 26, 2013 at 16:51
  • 1
    @H2CO3 This question was chosen because it's the most glaring example of a question that should not be closed that is closed under this new close reason. The question even went through the review queue and as (Thankfully) determined that it should be left open. If you look at Jaydles post, one of the major things Stack Exchange is trying to do is to be more welcoming. The premise of Stack Overflow was never to disallow RTFM questions. If you can find evidence that Joel or Jeff believed that (later it would be SE Inc), then share it. The custom reason allows people to invent their own rules. Commented Jun 26, 2013 at 16:55
  • 5
    @GeorgeStocker "The premise of Stack Overflow was never to disallow RTFM questions" - huh, that's serious. Then we answer them? What about the rule in the FAQ that says: "Search, and research. Have you thoroughly searched for an answer before asking your question?" That strictly and explicitly prohibits all RTFM questions.
    – H2CO3
    Commented Jun 26, 2013 at 16:58
  • 1
    @H2CO3 Note that "This question does not show any research effort." Is a reason to downvote. The downvote button even says so. It's not a reason to close. Commented Jun 26, 2013 at 17:00
  • 1
    I have no idea what you're asking. Could you elaborate on the "Why so"? Commented Jun 26, 2013 at 17:01
  • 2
    @H2CO3 Because it doesn't meet the critera lists for any of the given reasons for closure. Questions are not closed by default until given a reason to be opened, they are open by default until you can assert that they meet the criteria for closure. Which close reason do you think is appropriate, and explain any interpretation of that close reason as you feel is appropriate. Or do you think that all RTFM questions should be closed with the new custom reason?
    – Servy
    Commented Jun 26, 2013 at 17:04
  • 1
    @H2CO3 Servy touches on it, but there are dozens of meta posts related to this. If you feel like this should be different, then open a new meta question, get consensus, and let's see if it gets overriden by the SE team. I would be happy to address you fully there, since these comments are getting a bit off topic. Commented Jun 26, 2013 at 17:05

3 Answers 3


This is probably my biggest fear with offering an "other" option (and, incidentally, why the "other" text appears as a normal, attributed comment and not as part of the close banner). That said, I see two directions this could head in right now:

  1. Folks who were already (ab)using the TL and OT close reasons for stuff like this are now more explicit about what they're doing, and the problem gets tackled head on (you know this isn't the first time that particular user has voted to close a question for that reason...). Note that this question went through the close review queue and was left open...

  2. The scope of the site - at least within certain areas - changes, possibly in ways we would prefer it didn't. We can try to offer guidance, but ultimately it is the community that decides what kinds of questions they want or don't want.

It's worth noting that these comments can be flagged and deleted, at which point they're no longer available to future voters.

For what it's worth, here are the "other" reasons used so far on SO, along with the number of times they've been used (these are vote counts, not post counts) and a link to an example of a question where they were used:

  • 2
    I feared it could be used to turn away questions that we explicitly want, and now you've confirmed those fears. Thanks for digging this up. Is this available in the moderator queue for us to review when custom reasons are used or is that kept at the SE level solely? Commented Jun 26, 2013 at 16:20
  • No, there are no proper tools for this yet - I'm just writing queries. I do want some tooling for this built in at some point for mods (and, ideally, 10K users), but that'll have to wait until we smooth out some of the more pressing issues.
    – Shog9
    Commented Jun 26, 2013 at 16:21
  • 5
    "This question appears to be off-topic because it is about Dentists" -- what??
    – nhinkle
    Commented Jun 26, 2013 at 16:22
  • @nhinkle I'll quote one of the comments for that question: "Well indentation does not really have any relationship to dentists ." Commented Jun 26, 2013 at 16:26
  • @JanDvorak please tell me it was tagged indentation...
    – nhinkle
    Commented Jun 26, 2013 at 16:27
  • @nhinkle yep. Someone also suggested adding the string tag (the question was about cheap teeth removal) Commented Jun 26, 2013 at 16:29
  • So out of that list, there are at least 3 that should have stuck around (seriously, not too localized, again and again. Each of these questions may have a typo in them, but the information in the question (the error message, and the context) help future googlers determine what could cause this issue and where to look, even if the specific code changes. Commented Jun 26, 2013 at 16:29
  • Those recommendation ones should probably be under "tool or library recommendations".
    – KatieK
    Commented Jun 26, 2013 at 16:32
  • 3
    @George with probably much more specific knowledge in SQL Server, I disagree about this one. That error message actually almost always comes from a completely different problem - a misunderstanding of syntax, rather than an errant close bracket. I've been using SQL Server since 1997 and I can tell you quite honestly that this question is very unlikely to help other users.
    – Aaron Bertrand Staff
    Commented Jun 26, 2013 at 16:34
  • @AaronBertrand We disagree then (and I'm not a neophyte with SQL server either). Commented Jun 26, 2013 at 16:51
  • 3
    Though I won't comment on the last of your examples I thought that typo questions were always too localized @George. Why are the two you've commented about different? Commented Jun 26, 2013 at 18:20
  • 4
    @GeorgeStocker: You're welcome to post a Meta question about this, but there's an awful lot of community support for removal of typo questions. Jeff also posted an answer supporting this viewpoint a long time ago.
    – jscs
    Commented Jun 26, 2013 at 19:26
  • @JoshCaswell I'm applying the first sentence of what Jeff wrote in that post, "If it's about a specific error code that can be somewhat narrowed down, it's ok-ish." Commented Jun 26, 2013 at 21:05

If you're going to remove the ability to choose a custom off-topic reason, please consider this change on a site-by-site basis, where the site has demonstrated issues or abuse with those reasons. I don't think the concerns you're bringing up will be a problem on smaller or niche sites, while I do acknowledge that they can and will exist on the mother-of-all-SE-sites.

On our site (dba.SE), we mourned the loss of Too localized, as it allowed us to quickly close questions that are not helpful to other users or are simply not welcome. These include:

  • syntax errors (missing bracket, semi-colon, wrong keyword)
  • too localized (how many rows will this affect?)
  • RTFM questions that ask for documentation regurgitation (how does database mirroring work?)

We also still need a way for our users to vote that a question is too simplistic for our site, and that it should get kicked to StackOverflow. They can do this with a custom off-topic reason, but they can't with a canned off-topic reason because the only migration path (for non-moderators) is to our own site's meta.

FWIW we are still finalizing our list of custom off-topic reasons and TL is one of the candidates.

One of the SO questions you said should not have been closed (and is not "too localized") is this one:

SQL Join Issue - msg 4145

I disagree entirely. That specific error message is actually one that is usually from a different cause, and the missing bracket can often lead to a variety of other messages, depending on where in the query it appears. The question is really "be my second set of eyes and help me debug my code." The number of future users who will make the same typo and lead to the same error message is going to be ridiculously low, and perhaps zero. And I'd rather not have our users spend their time trying to find a duplicate question that was closed as Too Localized before these changes took place when they could still do so. For example - and note that the error message here came from a completely different kind of typo.

I do realize that moderators are not always equipped with specific knowledge about a particular technology, and may not recognize the difference between a common issue and one that is way too localized to be useful to others. But don't take away the right for users who are active in the question's tags to identify those same types of questions, and please don't do so network-wide if you're going to do it at all.

  • This would never (and should never) affect the ability of a moderator to create custom reasons for their site; only for end users to create custom rules per question. Commented Jun 26, 2013 at 16:58
  • 1
    @George the only custom reasons I can create for my site fall under Off-Topic > Other. Aren't you suggesting that that functionality be taken away?
    – Aaron Bertrand Staff
    Commented Jun 26, 2013 at 16:59
  • @GeorgeStocker That ability should not be revoked either.
    – H2CO3
    Commented Jun 26, 2013 at 16:59
  • 1
    I never claimed that the question should be kept around because of the typo itself, rather because the error message with the context of the code and the error location would help other users put that error in context, and see if it applies to them. Commented Jun 26, 2013 at 16:59
  • 1
    @George but there could be dozens of ways the error message could be associated with a typo, and sometimes it isn't even due to a typo. Should we have a question for each possible combination?
    – Aaron Bertrand Staff
    Commented Jun 26, 2013 at 17:00
  • 1
    @AaronBertrand We have knowledge and experience, two things question askers often lack. If we got rid of all questions that would produce these error messages (because they're typo questions) there would literally be no place for someone to get help. The best thing we can do is keep them around, and when there is duplication, close and link to it. Typo problems make up programming problems, and the compiler doesn't speak beginner. We can, and we do. Commented Jun 26, 2013 at 17:07
  • 5
    @George that's not what I suggested either. I explicitly talked about that exact question because (a) you mentioned it and (b) it is a typo question that also has the attribute that it will be unlikely to help future users. I did not suggest that all typo questions are too localized. Merely suggested that sometimes a typo question is too localized. Now, I could create a custom close reason called "too localized" but (1) I think that goes against the spirit of this change in the first place, and will still lead to ambiguity and (2) I don't want to waste one of my custom reasons on it.
    – Aaron Bertrand Staff
    Commented Jun 26, 2013 at 17:13
  • @George so what I'm suggesting is that for our site at least I want the "Other" write-in reason to remain.
    – Aaron Bertrand Staff
    Commented Jun 26, 2013 at 17:14
  • Syntax error is a symptom of lack of understanding of the topic (a valid off-topic reason) or a lack of sleep (off-topic on current site, maybe on-topic on personal productivity or workspace). If you already have a hammer, you don't need to hammer nails with rabbit legs. Commented Jun 27, 2013 at 11:39
  • 2
    @GeorgeStocker: I disagree. Stack Overflow is the place to get expert answers to your questions (per the Stack Exchange motto). You don't need an "expert" to tell you where your syntax error is. On most cases, you need a decent (and FREE!) editor. Tailoring answers to typo isn't what an expert would want to spend his time on. And I think that it's the best interest of Stack Overflow to preserve the experts, and not the typo-asking beginners. Commented Jul 6, 2013 at 15:54

That's a hole that could allow for bad closures and bad close reasons, I agree. But please don't just get rid of it. With the split of not a real question into unclear what you're asking and too broad and the loss of an unmodified off topic reason, we lost the means to close wildly off-topic questions, like the too-frequent Change something on my Facebook timeline, and the occasional question from outer space.

I've seen a moderator close a question as This question does not appear to be about programming, within the scope defined in the help center, without selecting a subreason. Lose the custom reason, and we might need something like that back.

  • That's done by selecting the custom close reason, and then deleting the comment afterward. Commented Jun 26, 2013 at 15:37
  • George is recommending that the custom close option be removed, @Michael. Commented Jun 26, 2013 at 15:45
  • 2
    I am; but if the custom reason said what you mentioned: This question does not appear to be about programming, within the scope defined in the help center, that's useful, and covers all the ground needed. Commented Jun 26, 2013 at 16:05

You must log in to answer this question.

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