6

Shog9 commented on his answer explaining the new close reasons, when asked for more clarity:

I will answer a separate question on this topic with a mapping of specific examples of questions to close reasons though. After I eat breakfast. If someone asks one.

So here is that question.

I noticed that the "minimal understanding" close reason is gone. Skimming through the discussion there, it seems like there's some confusion about how the various close reasons should be used (at least, it was confusing to me, because I always thought of "minimal understanding" as being similar to "too localized," but it seems to have been rolled into "too broad" now).

Can we make some kind of an easy-to-reference cheat sheet for each close reason with examples?

Something like this, maybe:


  • unclear what you're asking

    NARQ.

    Examples: #1, ...

  • too broad

    canhazcodez / homework.

    Examples: ...

  • primarily opinion-based

    shark v gorilla.

    Examples: ...

  • general computing

    configuring a server or network.

    Examples: #1, #2, ...

  • can't be reproduced

    too localized / typo / given code does not produce given error.

    Examples: ...

  • off-site resource

    "is there a library" / "which editor" / "where are the docs."

    Examples: ...


And just to clarify, "minimal understanding" been rolled into "too broad," correct?

  • 3
    homework is too broad? sounds like you want a mapping from what you object to and what you're going to say it is in order to close it anyway – Kate Gregory Jan 9 '14 at 21:48
  • 5
    The whole reason for the change was that people were making broad generalizations like this instead of actually thinking about what the close reason is stating and whether or not it applies to the question at hand. – Servy Jan 9 '14 at 21:48
  • 1
    Shark. No contest. – itsbruce Jan 9 '14 at 21:48
  • @itsbruce [citation needed] ;) – The Community Jan 9 '14 at 21:49
  • 1
    @KateGregory I want a mapping from what these things say to how they're generally intended to be used. – The Community Jan 9 '14 at 21:51
  • 3
    @TheCommunity There is no mapping. What they say is exactly how they're intended to be used. If you're using a reason to close something that the reason doesn't actually state, you're doing it wrong. – Servy Jan 9 '14 at 21:52
  • 8
    Do people here even understand that Shog9 said specifically for someone to ask this question? meta.stackexchange.com/questions/211080/…. Personally, I'd like to see Shog9's answer. – Lance Roberts Jan 9 '14 at 21:53
  • @LanceRoberts I guess not. – The Community Jan 9 '14 at 21:53
  • 2
    @Servy, what are you talking about, a lot of the reasons are very Unclear now, as many people have stated. – Lance Roberts Jan 9 '14 at 21:53
  • 3
    @LanceRoberts That comment is specifically saying not to do this. It's shog specifically saying that he's annoyed at people making mappings like is shown in the question here. "I'm tired of getting burned by folks taking pithy terms for questions and applying them in cases where they don't remotely apply" That comment is only talking about providing some good canonical examples of questions of each close reason. – Servy Jan 9 '14 at 21:58
  • 3
    @Servy, Here's the quote, emphasis mine: " I will answer a separate question on this topic with a mapping of specific examples of questions to close reasons though. After I eat breakfast. If someone asks one." – Lance Roberts Jan 9 '14 at 21:59
  • 1
    @LanceRoberts Yep. I saw that. That's not what this is though. This is asking to map close reasons to pithy terms for questions, the specific thing that shog was complaining about. – Servy Jan 9 '14 at 22:00
  • 3
    @LanceRoberts So you've just ignored the use of the pithy terms for questions here, or do they just not matter? If it were just asking for examples then it would be asking what that comment corresponds to. But it's not, it's asking for that in addition to encouraging the exact problem this change is trying to address. It's that second part that's a problem. – Servy Jan 9 '14 at 22:04
  • 1
    @Servy to my understanding the new close reasons were essentially the old close reasons, reworded to be less offensive to the OP, but still having roughly the same intended use. See "every close has its thorn." Why not map the concise old close reasons to the fluffy new ones where applicable? – The Community Jan 9 '14 at 22:20
  • 1
    @Troyen as far as I can tell, when this was proposed, it was purely about fluffing up the language and nothing else. I have asked low effort questions because I didn't feel like finding a windows box to test something on or whatever, but never had any "minimal understanding" close votes, presumably because it was clear that I understood what I was doing. How many questions can you find that show understanding but no effort that have been closed with "minimal understanding?" – The Community Jan 10 '14 at 8:10
15

Here is an attempt at a flow chart discussed in chat here:

enter image description here

  • 3
    "Has it been answered here before?" if No then why do we close it as duplicate? – hims056 Jan 15 '14 at 5:39
  • 3
    @hims, where would you get such a silly notion like that? I would never ever suggest such a thing. Unless it took more than 5 minutes to fix and was therefore outside the edit grace period. And then only if someone checked the revision history. (nice catch) – jmac Jan 15 '14 at 5:42
  • 1
    "Is it about programming -> no -> off-topic" is misleading, I think. At least half of the OT sub-reasons could apply to things that are about programming. So if it's not about programming, it's OT, but if it is about programming, it still might be OT. What I'm trying to figure out is if the stuff that was OT before is still OT, and we need to use a different close reason, or if some of it is no longer OT now that the close reasons are gone. – The Community Jan 17 '14 at 1:05
  • @TheCommunity, the subreasons are 'about hardware' 'about servers' 'was just a typo' (which may be about programming, but is an outlier reason in general), or 'a recommendation' -- so I think if it's about programming, it's generally speaking not going to be closed as off-topic unless it's a custom reason – jmac Jan 17 '14 at 1:27
  • So you think there are more write-ins than "typo" and "recommendation" combined? Or you don't think "recommendation" questions can be about programming? What about the old OT sub-reasons that didn't disclude stuff about programming, like "minimal understanding?" I don't have any stats handy but I have a feeling that plenty of OT closures have been about programming. – The Community Jan 17 '14 at 2:18
  • @TheCommunity, yeah, many off-topic closed questions were about programming -- and that was part of the issue and the reason for the change as I understand it. I just think that this mimics what my understanding of the close flow that Shog is talking about, not that it is what the community necessarily agrees with. I think closing bad questions as off-topic even though they are on the topic of the site and not closeable for other reasons doesn't make much sense, but I'm just one person so feel free to disagree. – jmac Jan 17 '14 at 2:30
  • I guess this is where the confusion comes from. If someone in an authoritative position were to say "stuff like this was off-topic before, but now it's not," it would all be perfectly clear and there would be no question. Instead there was kind of the implication that topicness hadn't really changed and that some close reasons had just been rolled together because they were too vague to be useful to the asker (but not necessarily because they no longer applied). What would be nice would be a clear answer as to whether topicness has changed. – The Community Jan 17 '14 at 2:51
  • 1
    You may be interested in Shouldn't "off topic" be only about...off topic?, @TheCommunity. – Josh Caswell May 19 '14 at 7:36
  • @Josh, thanks! Already read it before, but a good re-read. And good food for thought. I'm currently doing a bit of work thinking about our 'off-topic' close reasons on The Workplace, so this is great timing for me to have looked at that again. Making it clear to users is always a good thing. – jmac May 19 '14 at 9:31
8

I'll give you examples. I'll talk about the problems in those examples. But I won't classify them beyond the close reason itself: doing so leads to something akin to a hash collision - multiple distinct problems getting lumped into common category, with unhelpful results. Worse, it implies that you can skip identifying a problem entirely, and apply close reasons based purely on a superficial examination of a question - while this is often feasible, doing so in every case leads to cases where useful questions are closed based on irrelevant factors rather than the presence of actual problems on the site.

duplicate of

This is fairly obvious: a question asks for a solution to the same problem as a different question.

Example:

A is a duplicate of B

unclear what you're asking

Exactly what it says on the tin: it is not clear from the question what information the asker needs in order to solve his problem.

Note that many "unclear" questions are still questions - and they may even be answerable, if the answerer is willing to make sufficient assumptions regarding the problem. Answerers who are gamblers or psychics have a distinct advantage when it comes to answering these questions - and if such a question attracts a skilled answer from such a lucky / clairvoyant answerer, it may well be worth editing it to reflect the answer rather than closing it. The vast majority of these questions do not attract such answers, however.

Examples:

  • This question is unclear. In fact, it's not even clear that it's a programming question. It might be off-topic, it might be impossibly broad, or it might be trivial - who knows?

  • This question is also unclear - it mentions errors without including them, and hints at the presence of other problems without describing them. In fact, it has a very helpful answer that spends a great deal of time politely pointing out how utterly out of his depth the answerer is. This question is unlikely to be clarified because the asker has no idea what he's actually doing, but if he did it would likely be too broad as well.

too broad

Answers on Stack Overflow are limited to a paltry 30,000 characters. That's barely enough for a brief academic paper; a comprehensive tutorial on a subject of any size is unlikely to fit, and even a svelte book on any programming topic is right out.

In practice, very few answerers have the patience to write even that much. This is fine; there are other sites that cater to full-length articles. We're here to answer specific questions.

But some questions aren't specific. Some questions just lead to more questions. Others explicitly contain multiple, independent questions! A common motivation is an asker with a fairly basic understanding of a topic who wishes to leapfrog past the whole "study, trial and error" page and find that mythical Royal Road to programming. These are discouraged - rather someone ask multiple questions (or read the answers to existing ones), building their knowledge from the foundations up than require a personal tutor to walk them through the process.

Examples:

primarily opinion-based

You know what they say about opinions... There are a bunch of questions that are either designed or destined to collect raw opinions, devoid of anything verifiable. Sometimes, these are a lot of fun; other times, they're fertile ground for flame-wars.

Examples

off topic

These are supposed to be self-explanatory - hence the recent removal of a couple of oft-misinterpreted ones; if it's helpful, I can provide examples for them too though. The newest off-topic reason is:

This question was caused by a problem that can't be reproduced or a simple typographical error. While similar questions may be on-topic here, this one was solved in a manner unlikely to help future readers. This can often be avoided by identifying and closely inspecting the shortest program necessary to reproduce the problem before posting.

...and it's aimed squarely at this sort of thing.

  • 4
    So what close reason would you give for something like this under the current scheme? Under the old scheme, I would have gone with "too localized" as it's unlikely to help anyone else (from the site's target audience). Under the previous scheme, it pretty clearly fell under "minimal understanding." Would it be "too broad" now, because we would need to fill in too many pieces for the OP? It's really not that many pieces though; here's the formula, here's how the code would look. Write in a close reason? Leave open? Something else? – The Community Jan 10 '14 at 7:11
  • 1
    Those are great examples, I sure hope you have some way of marking them to not be deleted in the future so they can keep being those examples. – Lance Roberts Jan 10 '14 at 7:26
  • 1
    "this one was solved in a manner unlikely to help future readers" kind of implies it's only relevant once an answer has been accepted, doesn't it? Maybe reword to something like "this one is unlikely to be useful to future readers?" – The Community Jan 10 '14 at 7:39
  • 2
    Thanks, Shog9 for your explanations, but I also have still problems like @TheCommunity. I review regularly the regex questions and there are so many, just asking for code, like this, this and this. I think it is clear that they should be closed, but the reason??? – stema Jan 10 '14 at 7:40
  • @stema that second one is a great example, I want to close vote but am not sure what reason to give. It has two off-topic close votes, but we can't see what they are, so they were probably old "minimal understanding" votes. I wonder if it would be closed by now if "minimal understanding" were still here. If stuff like this should be closed, there should probably be a good boilerplate reason, so it's clear to everyone that it really should be closed. – The Community Jan 10 '14 at 7:46
  • @stema The second one would be "unclear what you're asking" (the comments even suggest some confusion - is the OP asking for help with their tool? Why is it tagged regex then?). The third one could also be "unclear what you're asking" because it's not very specific. Not sure about the first one. And if you think the OP didn't put enough effort into their questions, make sure to downvote them. Like the tooltip says, downvotes are for questions that did not show any research effort, are unclear, or unuseful. – Troyen Jan 10 '14 at 7:49
  • 1
    @stema Actually, I'm not sure the first one is really close-worthy. – Troyen Jan 10 '14 at 7:56
  • 3
    @Troyen the second one seems pretty clear to me. "Here's what I want to do, here are the tools I'm using now." I would assume he just wants it to get done and doesn't care how, and shared all the information he had. A gedit answer or a sed answer would both probably satisfy him. In gedit, I'm almost positive escaping the backslash with another backslash would do the trick, but anyone with a minimal understanding of what they're doing would have tried that by now. "Unclear" seems like a stretch to me. If it should be closed, there should probably be a better (authoritative) reason. – The Community Jan 10 '14 at 7:58
  • @TheCommunity If you understand what's being asked, then it's not an unclear question (I just assumed it was because of the comments). Based on all the other discussions today, it sounds like it shouldn't be closed then. – Troyen Jan 10 '14 at 8:05
  • Agreeing with @TheCommunity. I have some linguistic reservations about this close reason, too. Why did you decide on "solved" where questions are concerned? – Mathias Müller Jan 10 '14 at 9:01
  • 1
    @TheCommunity: you're doing exactly what I tried to avoid - you're looking for questions that match a simple pattern instead of, upon finding a question and seeing a problem with it, looking for a close reason that describes that. Also, you're utterly missing the point of the typo reason even though I state it in the reason and link to a big discussion of it: they're blatantly obvious, you can put away your magnifying glass. – Shog9 Jan 10 '14 at 15:05
  • You have three different questions with (at least) three different issues, @stema: the first one is off-topic (belongs on either Ask Different or Unix & Linux), the second is missing a tag, and the third is unclear. – Shog9 Jan 10 '14 at 15:07
  • 1
    @Shog9, I'm not sure what you mean. Troyen linked a question, I found it via his link, I see problems with it (seems not useful to anyone except asker, very low quality, etc.), I want to close it, I'm looking for a close reason that describes it, I can't find one. I'm not looking for questions, I just happened to find one. The opposite of what you are saying is true, AFAICT. – The Community Jan 10 '14 at 20:50
  • As for the second part, I get the point of the typo reason. Trust me. That whole typo crusade stemmed from a question I posted in the first place, remember? I'm just saying that I think the wording could be confusing to someone else (probably an asker), not that it's confusing to me. – The Community Jan 10 '14 at 20:52
  • 1
    In my opinion it's a question that shouldn't be answered because answering that question just invites infinite permutations of questions like that. How do I convert liters to gallons in Visual Basic? It's just going to drive people nuts. I can't find a close reason that fits that reasoning, as I said in the initial comment. I still think it should be closed, and I don't want to have to resort to this kind of abuse to close questions. Should I write in my "infinite permutations" close reason then? – The Community Jan 11 '14 at 19:11
1

The cheat sheet is the entry in the help center that discusses closed questions.

What specifically in https://meta.stackoverflow.com/help/closed-questions is lacking?

(i.e. Rather than ask if we need a more detailed list, why not propose specific wording for one close reason - the one that is most lacking/misused/underused and build a case for why the current documentation is lacking?)

  • You see a bug perhaps, I see a feature. – bmike Jan 9 '14 at 21:51
  • 2
    Really? It lacks brevity? Three lines or fewer per reason is seriously not brief enough? – Jonathan Garber Jan 9 '14 at 21:53
  • @JonathanGarber what am I supposed to be looking at? I follow that link and see a page of more links. – The Community Jan 9 '14 at 21:55
  • 3
    @JonathanGarber that looks more like a reference for askers than for closers to me. – The Community Jan 9 '14 at 21:57
  • 3
    That page is lacking details. For instance, I learned from the comments on another thread that a question that just lists a set of non-trivial requirements and asks how to build it, with no example code, should be closed as "too broad". Is that right? I don't see how I could derive that from the text in the FAQ. In short: The FAQ needs more detail to help us closers know when to use the various close reasons, and when not to..... – D.W. Jan 9 '14 at 23:59
  • 3
    There's nothing fundamentally wrong with having a separate resource for closers. – Pëkka Jan 10 '14 at 3:00
  • In my opinion, there should be information on when to close on what grounds here: meta.stackoverflow.com/help/reputation - as this section is apparently concerned with "moderation". – Mathias Müller Jan 10 '14 at 9:55

You must log in to answer this question.

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