# New close reasons: unclear, too broad, opinion-based, and custom sub-reasons off-topic

There will be changes in close reasons in the next few days:

• “not constructive” and “not a real question” are replaced by

• “primarily opinion-based”
• “off-topic” and “too localized” are replaced by “off-topic” with changed subheadings:

• up to 3 site-specific close reasons which will appear as subheadings of “off-topic”;
• “off-topic” with a free-form comment that must be at least 88 characters long (the prefix “This question appears to be off-topic because it is about ” is pre-entered, so you have to add at least 30 characters);
• migration (only to meta as long as the site is in beta).
• Closed questions will be labeled as “on hold” for 5 days, and “closed” afterwards.

Should we have predefined off-topic sub-reasons at all (otherwise, we'll retain a default unspecific one)? If so, which?

• Nice! Hm, I'd like site-specific variants for the first category; our "homework grading" and "homework without effort" pseudo-policies would make for good candidates. – Raphael Jun 13 '13 at 5:33
• @Raphael We have no homework policy that calls for closing certain questions for other than the usual quality control (i.e. the default close reasons). – Gilles 'SO- stop being evil' Jun 13 '13 at 7:22
• Yes, in principle, but the number of cases may warrant to move the default comment into the close reason. Or do you think only policied reasons should be there for clarity? – Raphael Jun 13 '13 at 10:46
• @Raphael The number of cases of closure? We aren't closing homework questions by any homework-specific policies, so why would there be a homework close reason? I'm not necessarily against a “you just dumped your assignment on us” close reason (as opposed to a “smells like homework” close reason, which I vehemently oppose), but we have no such policy. – Gilles 'SO- stop being evil' Jun 13 '13 at 10:52
• by the way are these official stackexchange changes in close reasons? is there any ref on that? close reasons are the same across all sites right? – vzn Jun 14 '13 at 18:31
• From Gilles' earlier comment: The close reasons are listed in faq (recently reorganized from /faq to /help). They are the same across Stack Exchange sites, except for the upcoming off-topic sub-reasons. – Raphael Jun 15 '13 at 21:21
• @Raphael Important note which wasn't explicitly mentioned in the original announcement on MSO: if we have custom close reasons, we no longer have plain off-topic! – Gilles 'SO- stop being evil' Jun 20 '13 at 12:21
• @Gilles That's a bummer. But I guess we can add our own unspecific OT, then? – Raphael Jun 20 '13 at 13:04
• @Raphael We're not supposed to (“off-topic reasons of the form "Things that are NOT X" will be discouraged”), but I don't see a better solution. – Gilles 'SO- stop being evil' Jun 20 '13 at 13:14

We could use a “programming question” close reason, for those students who think that any homework from a CS course is computer science, including programming assignments. (This only explains some of our programming questions; the really bad ones are often from askers who have been blocked on Stack Overflow due to low-quality contributions.) Here is the wording we have:

Questions about software development or programming tools are off-topic here, but can be asked on Stack Overflow.

Moderators can migrate a question to Stack Overflow, but only if the user is registered there and is not blocked from asking questions. In most circumstances, we prefer to let the asker repost. But if the question is a good one, and especially if there are answers here worth migrating, we can migrate.

I think we need an explicit fallback for the questions that are neither about computer science nor about programming. This is now live:

This question does not appear to be about computer science, within the scope defined in the help center.

There should be a closing reason we can use if somebody has cross-posted a question on three or four different stackexchange sites. I assume we still want to discourage this behavior.

• Good point, we were using too localized for that. I guess we'll use off-topic with a custom comment from now on. – Gilles 'SO- stop being evil' Jun 22 '13 at 18:29
• I think this one can be covered by "Other" where we can enter arbitrary reasons. – Raphael Jul 10 '13 at 7:51
• @Gilles Given the unpleasant experience with this question, I think we should create the "crosspost with an answer" aka "duplicate on another site" custom close reason so that the text fits the close reason. (Or we should make a case for allowing duplicate links to other sites.) – Raphael Jul 30 '13 at 10:36
• @Raphael On Unix & Linux we have This question has been **posted on multiple sites**. Cross-posting is strongly discouraged; see the [help center](http://unix.stackexchange.com/help/on-topic) and [community FAQ](http://meta.stackexchange.com/questions/64068/is-cross-posting-a-question-on-multiple-stack-exchange-sites-permitted-if-the-qu) for more information. But cross-posting is more common on U&L than on CS, do we really need to spend one of our predefined close reasons on this? I'd like to see a separate meta discussion with some statistics before proceeding. – Gilles 'SO- stop being evil' Jul 30 '13 at 13:57
• @Gilles We can redefine the reasons anytime, don't we? Meaning, as soon as we find a better one, we could ditch this. I think we have cross-posts quite regularly, although it should be tough to obtain reliable statistics. – Raphael Jul 30 '13 at 14:30
• @Gilles: As I understand it, we're not using all the slots for off-topic close reasons yet, and still have two slots open. I don't see what other reason would work better to fill this slot. Do you have any suggestions? – Peter Shor Jul 30 '13 at 16:21

I think the replacements proposed in the question for "not constructive" and "not real" are a good move. Although the possible motivations for these closure reasons are stated relatively clearly, I think splitting them into more explicit, individual categories will help the questioners focus their editing more effectively. I also think that it might help make people more cognisant of the fact that closure isn't meant to kill a question, but to allow the questioner more breathing room to fix a poorly asked question.

Along these lines, it might be worth considering some sort of "needs to be split" option for questions that actually house more than one real question. There is some overlap with "too broad", but there's a notable difference in that a single question could be too broad in the sense of "why do we do computer science", as compared to a handful of individually relatively focussed questions, that should just really be separate questions under the stackexchange model.

• Agreed: "Contains more than one question, consider rewriting or splitting into several questions" is a much more constructive annotation than "not a real question", when it applies. – András Salamon Jun 21 '13 at 9:40
• "Contains more than one question ..." could easily be a comment when we use the close reason "too broad". – Peter Shor Jul 31 '13 at 3:10
• In fact, this should be one of our reference templates with common comments. Does anybody know of any other SE sites with such a template, or do we need to write it ourselves? – Peter Shor Jul 31 '13 at 12:04

I suggest an additional closing reason for problem dumps with no effort shown and no further information/context ("solve my homework for me", with no effort shown, except that this isn't homework-specific). Candidate wording:

• When you ask a question here, we'd like to see that you have put in some effort towards solving your question yourself. If standard resources (textbooks, Wikipedia, search on this site, etc.) are sufficient to answer your question, that may indicate that your question is not a good fit for this site. You might consult those references, and if you're still stuck, edit the question to show what you have tried so far and where you got stuck.

See also https://cs.stackexchange.com/questions/how-to-ask for some related wording. See also discussion of Math.SE's policy on questions that are nothing beyond a problem statement (and don't show any evidence of effort): https://math.meta.stackexchange.com/q/9201