48

This question concerns the following close reason:

This question was caused by a problem that can no longer be reproduced or a simple typographical error. While similar questions may be on-topic here, this one was resolved 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.

Various users have attempted to convince me that this close reason should be applied to any question that meets the following criteria:

resolved in a manner unlikely to help future readers

I see some problems with this way of thinking, namely:

  1. It ignores the rest of the close text.
    (which would apply to a much more selective group of questions)
  2. It leaves the closure of any given question open to the subjectivity of whether a given user believes that the question is likely to help future readers or not.
    (this seems like more of a reason to downvote than to close vote)
  3. I'm certain that most questions asked could be closed given this mentality.
    (until the point that they actually help a future reader)

But this keeps coming up, so I'm beginning to question my sanity.

Should we be using this reason to close vote all questions that are resolved in a manner unlikely to help future readers when the questions are not caused by problems that can no longer be reproduced or simple typographical errors?

3
  • 22
    This probably becomes a reading comprehension discussion which I'm terrible at but I risk it anyway: Isn't the close vote criteria just This question was caused by a problem that can no longer be reproduced or a simple typographical error.? Everything following it is more an elaboration but not part of any criteria for closing questions.
    – rene
    Commented Apr 17, 2017 at 19:11
  • 9
    To put your question in another way: In the absence of further qualification, isn't "resolved in a manner unlikely to help future readers" just a minor variation of "too localized"?
    – duplode
    Commented Apr 17, 2017 at 20:46
  • 5
    IMO the "unlikely to help readers" was the explanation given to justify the actual close reason of "typo" or "no reproduction". Most of the questions I read are unlikely to help me but are interesting at least.
    – TylerH
    Commented Apr 18, 2017 at 15:09

2 Answers 2

78

Short answer: do not pretend to be able to predict the future; close only based on the evidence in front of you.

The history

This close reason was originally proposed by Brian Diggs as a replacement for one specific use of the former Too Localized close reason; note in particular Kate's comment in the discussion there:

But not all typo questions are TL: imagine someone who muddles up comma and semi colon in C++ for example. The compiler error message may not make sense to the coder, but searching SO could lead to someone's typo question and thus enlightenment. Only when the typo is really not going to help anyone (you declared Count but are using Cont) should it be closed - and offtopic feels a little strange for those since you don't know that until it's been answered.

This led to the wording we have today; observe the note in the announcement several months later:

Note that the original wording proposed for this reason has been tweaked several times in response to confusion observed here on Meta and on Stack Overflow itself. The target remains the same though: "face-palm" problems that no longer affect the asker and whose solutions will never benefit anyone else.

Emphasis added: the intent here was clearly to create a descriptive reason for use when closing questions whose root-cause arose from a mistake that cannot be identified (and thus whose solution cannot be used) by future readers, even those making the exact same mistake.

In other words, the bold parts of this close reason are not optional. The critical flaw in the old "Too Localized" reason was that it invited close-voters to pull out their crystal balls and speculate on the unknowable future: is this a problem that someone else might have again? The results were often laughably inaccurate - we saw questions that were the target of multiple duplicates, with heavily viewed and up-voted answers, being closed as "too localized" based on someone's bizarre idea of what problems other people might encounter... This was clearly counter-productive.

The goal with the "typo" reason was to capture the more objective scenarios, those where the problem wasn't actually described in the question at all. Things like...

  • Asker questions his implementation of an algorithm only to realize that a loop wasn't being executed due to a stray semicolon.
  • Strange browser rendering that ended up being a misspelled JavaScript symbol name.
  • Suspected compiler / API bug ends up being the asker forgot to rebuild after making code changes.
  • etc. - anything where the stated problem is wholly removed from the actual problem.

Note that this does not cover a bunch of situations that folks sometimes assume it does:

  • Calling the wrong API function, or forgetting to initialize a parameter causes a specific misbehavior (which is noted prominently in the question): this is a mistake that someone else might make and if the effects are described then they can find it when they do.
  • An actual API / tooling bug (cough new XCode release cough) that causes things to break. This will almost invariably affect multiple people until it is fixed (usually in a subsequent release).
  • A typo where the error message - stated in the question - actually points to a typo with some accuracy. Again, this will almost certainly happen to someone else at some point (and thus is useful as a searchable solution or duplicate-target).

Worth keeping in mind that we've always kinda preferred closing things as Duplicates than using some other reason, if the topic itself is reasonable; over the past few years, the system has been altered to make dup-closing far more efficient than any other form of closing... So if a problem can be solved in a general-purpose fashion, leaving it as a duplicate-target is much better than closing it, even if you personally hate the question and are disgusted by the person asking it. Something you might consider reminding the folks who argue for more liberal interpretations of these reasons...

27
  • 2
    "This will almost invariably affect multiple people until it is fixed", so once it's fixed, what we do? Note, I'm strongly biased towards deleting "fixed" bugs, since the only solution is to upgrade, these are potential waste of time if someone finds the same problem but for entirely different causes.
    – Braiam
    Commented Apr 17, 2017 at 21:35
  • 23
    @Braiam if it can still be reproduced by using an old version, it can still be reproduced. If the old version is no longer available (such as with a web service), then it can no longer be reproduced.
    – user4639281
    Commented Apr 17, 2017 at 21:40
  • 11
    If there's a work-around for a particular issue in a particular version, I'd err on the side of keeping that @Braiam. If there's no work-around, then at least documenting the existence of a bug in a way that prevents undue frustration can still be valuable. In any case, these are generally a much smaller problem than the one at which this close reason was aimed; a clearly-identifiable problem is always preferable to a misleading one.
    – Shog9
    Commented Apr 17, 2017 at 22:01
  • 33
    @Braiam. sigh unfortunately not everyone is in control of the software they must use. For example, the standard browser for many large companies is still IE8, for which the same argument applies. I honestly don't see what the problem is with leaving stuff open if it only applies to a small portion of the population, and everyone else can ignore it. If it's otherwise on-topic, why is there this obsession with closing every single question that could possibly fit a close reason?
    – user4639281
    Commented Apr 17, 2017 at 22:25
  • 6
    @Braiam because some software authors actually remove features and make incompatible changes when releasing new versions, in addition to fixing bugs. To reiterate the answer: not everyone agrees on your idea on what's more productive for other people.
    – artem
    Commented Apr 18, 2017 at 0:21
  • 10
    @Braiam Just because a solution doesn't work for you, does not mean that it won't work for someone else. Would you rob someone else of the knowledge that would help them solve their problem, just because it doesn't help you solve yours? And no, I cannot agree with your last statement. There are many situations where that is not the most productive course of action.
    – user4639281
    Commented Apr 18, 2017 at 0:34
  • 11
    @Braiam It isn't about what is sane, it is about what is on-topic. If Chrome 36 is still available then a work around for a bug that exists in Chrome 36 is still on-topic, and may still help future readers (though determining that is highly subjective, so only the on-topic argument counts).
    – user4639281
    Commented Apr 18, 2017 at 0:50
  • 18
    @Braiam, I've consulted for many clients who have no choice but to work with old versions of software that have many documented bugs (ranging from old versions of IE to old versions of a CRM or ERP platform). "Upgrade your stuff" isn't always an option - or, even when it is an option, it's sometimes an option that will take those clients several years to achieve (due to cost of new licenses, hardware, training users, upgrading/migrating customizations, corporate IT policies, etc). I'm always grateful to find good documentation/work arounds for these kinds of bugs, even when old
    – Dan Field
    Commented Apr 18, 2017 at 15:11
  • 2
    It should be noted that "too localized" worked much better, as it could also be used for questions of the nature "How do I print text without using the printText() function?", which are always artificial homework questions. Finding the answer to such questions is of no interest to anyone, including the OP (who might however not realize it), since the proper way to solve the problem in a real-world application would be to use the designated language feature for that task.
    – Lundin
    Commented Apr 19, 2017 at 11:25
  • 3
    But...but... "even if you personally hate the question and are disgusted by the person asking it" awwwww, spoilsport. Commented Apr 19, 2017 at 11:44
  • 5
    You're being either naive or nasty about this, @Braiam. Awful lot of people who need a fix today even if they can and fully intend to upgrade ASAP. And an awful lot of people for whom ASAP is measured in months if not years.
    – Shog9
    Commented Apr 19, 2017 at 23:12
  • 2
    @Braiam Yes, upgrading is the most permanent, and overall most effective solution, but that doesn't mean it's automatically the best and/or only answer in every situation. Large businesses can take months, or even years, to upgrade, for example, simply because they need to be absolutely certain that all of their assets will still work properly after upgrading; if they upgrade the instant a new version comes out, and it breaks some or all of their stuff, they're just going to have to restore from their backups, meaning that upgrading will have actively cost them time and money for no benefit. Commented Apr 20, 2017 at 18:56
  • 2
    Hobbyists might prefer an answer other than "Upgrade, yo!", too, especially if they're into retrocomputing or have their environment already set up the way they like it. In this case, upgrading is the mechanically best answer, but it's not the best for them if they specifically prefer their outdated version over the newest one. Commented Apr 20, 2017 at 19:04
  • 2
    So, actively going out of our way to deprive anyone that either chooses not to or is unable to use the most recent version of solutions doesn't benefit either Stack Overflow or the users, as it actively decreases the userbase's collective knowledge. Additionally, if the same bugs, or similar ones, show up again in future versions, then knowledge of how they were dealt with in the past could be useful for solving them again. Commented Apr 20, 2017 at 19:13
  • 5
    The easiest answer to most questions on The Workplace is "quit your job", @Braiam. Strictly-speaking, it's a valid answer - and sometimes, it's the best solution. Usually, it's just a lazy answer from someone who doesn't care.
    – Shog9
    Commented Apr 21, 2017 at 21:53
-15

I will propose an alternate approach:

All questions that are reasonably likely to be of use to some unknown future reader should be kept.

Disk space is not expensive. Human effort (to rediscover answers) can be very expensive.

I hope that those with the luxury of always updating to the latest versions of tools will remember that some users (e.g. legacy projects or in regulatory environments) may need to continue using tools that are many years old.

6
  • 3
    I disagree with this because it again leaves it open to the subjectivity of determining what problems others will encounter in the future. You're still saying that you can use the close reasons for questions that you deem unlikely to help to future readers. I think the only reasonable option (as far as I can tell at this point) is the one presented in Shog9's answer.
    – user4639281
    Commented Apr 18, 2017 at 0:37
  • 39
    Storage space is never the issue; polluting the Internet with red herrings for future researchers is the problem we hope to avoid. There's another issue, often conflated, which involves folks getting all upset about others not being able to debug their own problems; as irritating as this is, closing doesn't solve it.
    – Shog9
    Commented Apr 18, 2017 at 0:49
  • @TinyGiant, please explain how you got "you can use the close reasons" from my answer. I wrote "should be kept." "Kept" is very different from "close". Commented Apr 18, 2017 at 0:50
  • 3
    What is the alternative to the content being kept? You're saying that all questions which you deem reasonably likely to be of use to some unknown future reader should be kept. What should happen to the questions you deem not reasonably likely to be of use to some unknown future reader? This all hinges on what you deem reasonably likely to be be of some use to some unknown future reader, which is a very subjective measurement, and is no different than the mentality that I'm asking about, just flipped.
    – user4639281
    Commented Apr 18, 2017 at 0:54
  • 3
    How exactly can you define reasonably likely to be of use to some unknown future reader? It's an impossible decision at worst, incredibly subjective at best
    – Liam
    Commented Apr 19, 2017 at 12:09
  • 1
    @Liam Hmm... we could define reasonably likely to be of use to some unknown future reader as "doesn't meet any of the specified close reasons", but then we're just going around in circles. I believe the implied answer is that everything should be given the benefit of the doubt, unless and until it can be explicitly proven that it will never be of use to any individual on Earth ever again... at which point it effectively becomes "never close anything unless you kill all humans first", which doesn't really solve the problem. Commented Apr 19, 2017 at 22:21

You must log in to answer this question.