There have been questions that received answers before being closed, e.g. this one. I feel like there have been more, but I might be wrong.

Should we discourage people from providing answers to questions that are obviously or apparently (existence of [more than one] close vote) off-topic or otherwise prone to closing?

I know that there is such a policy on cstheory.SE.

Related question: What to do with existing answers on closed questions?

  • 3
    $\begingroup$ on cstheory we have a policy that we remove answers to clearly off-topic questions after they get closed. Clearly here typically means that the question was clearly not research-level. The reason for the policy at the time was that from time to time (near the end of semesters) we receive high number of off-topic questions from undergraduates which are answered before we can close them. I don't see any need for such a policy for Computer Science right now. $\endgroup$ – Kaveh Apr 10 '12 at 6:26
  • 2
    $\begingroup$ (btw, the policy was controversial even on cstheory, and it barely obtained the required vote.) $\endgroup$ – Kaveh Apr 10 '12 at 6:29
  • $\begingroup$ @Kaveh: Thanks for the reference. I don't think the scale of the problem is a good reason against a policy; better train the bear while it's young. $\endgroup$ – Raphael Apr 10 '12 at 6:31
  • $\begingroup$ I guess that is a German proverb. :) I personally prefer "If it ain't broke, don't fix it". $\endgroup$ – Kaveh Apr 10 '12 at 6:42
  • $\begingroup$ @Kaveh: It is not a proverb I know of, but yea, we (Germans) are known to value prevention. Pays off often enough. $\endgroup$ – Raphael Apr 10 '12 at 6:45

It's up to each poster to decide whether the question belongs on the site or not. If you think the question doesn't belong here, vote to close. If you think the question belongs here and you have an answer, post your answer. That others have answered should not influence you not to vote to close, and that others have voted to close should not influence you not to answer, other than the obvious way in which you may let others' arguments for or against the question sway you.

CSTheory.SE is a special case, as it gets many questions that are off-topic solely because hey are too basic, so that the audience would be capable of answering them.

Generally speaking, if you vote to close a question:

  • as duplicate, answer on the duplicate.
  • as off-topic, don't answer, because the site's audience will not be capable of evaluating your answer. Stack Exchange maintains the quality of its answers by peer review; answers to off-topic questions get no peer review and are therefore automatically suspicious.
    • Exception: if you think the question is likely to be migrated to a site where it will be on-topic, it's ok to answer without waiting; your answer will be evaluated on the target site.
  • as non-constructive, don't answer, because non-constructive means that there cannot be a balanced answer, or the risk of provoking arguments is large; answering a non-constructive question defeats the purpose of closing these questions.
  • as not a real question: the meaning of this close reason is that it is not reasonably possible to answer this question, so answering is logically inconsistent.
  • as too localized: this is a bit of a special case; I consider it bad form (but might leave a comment).

There's no need to do anything special with answers to closed questions. Closed questions (except duplicates) are normally deleted when it becomes clear that they won't be reopened.

  • $\begingroup$ I guess the point here is wether we should appeal to people to handle questions with (several) close votes differently than others, i.e. do not answer immediately even if they don't see a problem with the question (or choose to ignore that problem). $\endgroup$ – Raphael Apr 10 '12 at 14:28
  • 3
    $\begingroup$ @Raphael Why should the presence of close votes have any influence on whether to answer a question? Users who don't have enough reputation to vote to close don't even know whether there are close votes. If I see the close votes and disagree that the question should be closed, why wouldn't I answer and bookmark the question to reopen it if it gets closed? $\endgroup$ – Gilles 'SO- stop being evil' Apr 10 '12 at 15:01

So far I don't really see that answers to off topic questions are an issue for this site yet. Most other SE sites (probably around 99%) do not have or needed to discourage users from answering off-topic questions, I don't see why we need it either.

We are also in a better shape than cstheory because most off-topic questions are actually migrate-able if this even becomes an issue, while the same cannot be said about cstheory before cs.SE. Additionally, our set of off-topic questions are considerably smaller compared to cstheory, so it is inherently harder to ask an off-topic question here compared to cstheory. Our site resembles more like 99% of the other SE sites that don't adopt this policy, and cstheory is an exception among all of the SE sites, we are nothing like cstheory except we share some of the same topics.

Maybe if in the future, that this is indeed a real problem, we can talk about it again, but for now, my vote would be a solid no.


I generally agree with Gilles that people should answer questions they think they should give good answers to. If the answers are more interesting than the question, this is a sign to try and fix the question, not to close it.


I know I hated the policy on cstheory because I felt that (good) questions too basic for the site deserved an answer. This can clearly not be the case here, though; questions we close are (hopefully) clearly offtopic and/or really bad.

Reasons for implementing such a policy:

  • Protecting our scope -- If we answer even offtopic/bad questions, we can stop closing altogether. "But it received a good answer!" should never be a reason not to close.
  • Consistency -- The question might be salvaged, but it might be changed heavily in the process. Existing answers may become invalid.
  • Ensuring quality -- The quality measure when formulating a question has to be a YES-answer to "Is it good?" and not to "Can I ask it in such a way and at such a time that I will get an answer before enough close votes?"; we don't want the ensuing races.
  • Prevent reputation fishing -- questions about to be closed are obvious target to write a quick and dirty answer that get all (answer) votes this question will ever get. In the worst case, people inclined to write offtopic/bad posts answer each others questions, forming a subcommunity.

Reasons against:

  • If the question turns out to be salvagable or the problem was a misunderstanding, users might be offended.
  • It is hard to enforce beyond deleting answers.
  • Especially new users might feel that their contribution is not appreciated.

All in all, I am pro such a policy. Items one and three of the reasons against can be prevented (or at least softened) by friendly explaining the issue.


You must log in to answer this question.

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