What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 128 Stack Exchange communities.

This is inspired by

On one hand, I think closing a question as a duplicate because the answers are duplicates is just dead wrong. The close reason is "exact duplicate" regarding the question, not the answers. When a question is closed as an exact duplicate, it is a reprimand to the asker -- saying essentially, "You couldn't take the time to check to see if someone had asked the exact same question already?"

On the other hand, I agree with GMan and friends that we are here to provide information, and copy/pasting answers is silly.

It would be nice for a way to say that the answers might be the same, despite the question not being the same.

EDIT: I think other possible solutions to this problem might exist. I'm most concerned about the problem getting solved; I don't care if it's this actual fix... any ideas?

share|improve this question
3  
There is already a feature in the moderator queue that identifies and displays copy-pasted answers for the moderators to review. That said, I think it's perfectly reasonable for one answer to be the correct answer for two or more different questions. –  Robert Harvey Jun 20 '11 at 19:37

2 Answers 2

If the answers are the same but they are fundamentally different questions, then they deserve to be different questions. I'll take an example from Gaming Meta, that worked well to illustrate this point to me.

I asked for an atomic UNIX operation on unix.se, and got the answer I was looking for - mkdir. It would be ridiculous if a question about how to create folders got closed as a duplicate of mine! - Kaestur Hakarl, Sep 4th 2010

Sometimes, genuinely different questions will have identical answers. Having the same answers hints that they might be the same question, but if they aren't the same question, I don't think they should be closed for merely sharing the same answer.

Copy/pasting answers is a bit different. I'm not sure how often you can completely copy paste an answer to a genuinely different question without needing some manner of tuning to address the question presented. In most cases that one might copy-paste without touching on the answer, it is likely a much closer duplicate question or the user is spamming the same answer (which gets flagged, as noted by Robert Harvey in the comments).

share|improve this answer
    
+1 -- but as for your last paragraph, I know of at least two recent examples. The first is the example I cited above. The second is this one got closed in favor of this one. Same exact answer applies, but the questions are very different -- the first is "how does std::forward work?", while the second is "why would I ever want to use std::forward?". –  Billy ONeal Jun 20 '11 at 19:54
    
@Billy The first one strikes me as laziness at sculpting the answer to better fit the target (it lacks any highlight to focus on what's important to the actual presented question). Second one... I can't tell which answer was the duplicated one, being only 1.5k and all. But this still strikes me as exceptionally rare in comparison to how often a strict copy/paste is not a marker for warning. –  Grace Note Jun 20 '11 at 19:58
    
@Grace: It's the one (currently) with 38 upvotes (the top one). –  Billy ONeal Jun 20 '11 at 19:59
    
@Billy In that case, that appears to be a case of "GMan gave the answer to a different question". It's only copy/paste because of GMan addressing your question before you asked it, in a different question. We had a similar question on Gaming, where someone basically went well beyond what was asked of the actual question. It's not exceptionally fair to the second asker (Who is going to think to look for the answer to Question A in completely different Question B?), but that's why we try to figure out a solution that'll properly get such things addressed. –  Grace Note Jun 20 '11 at 20:02
    
@Grace: I agree. :) (That's why I ended up asking for this even though it would cause a bunch of my questions to be closed... I'm of course open to other ideas to solve the problem) –  Billy ONeal Jun 20 '11 at 20:04
    
@Grace - In this particular case one question is "Why can we not instantiate a function template with a local type" and the other is "Why can we not instantiate a class template with a local class". The answer is that we cannot instantiate any templates this way, and for the same reason. That's why the answer is the same (and really the question too). –  Bo Persson Jun 20 '11 at 21:30
    
Although I generally agree with this answer, I think it should be noted that examples like the one quoted are straw men in this discussion. The bare "mkdir" is a terrible answer across the SE network. Any acceptable, genuinely helpful answer would go into more detail, and distinguish "Name an atomic operation" from "How do I create a folder?" by explaining the context. –  Josh Caswell Dec 27 '13 at 19:32

There's no reason to add this. The existing close reason "Exact Duplicate" already covers this case. I think that the full text of the close reason is quite relevant here, and also very clear:

exact duplicate
This question covers exactly the same ground as earlier questions on this topic; its answers may be merged with another identical question.

There are two parts to this: first, the part on which you are focusing, that the questions themselves be equivalent, and second, that the answers can be usefully combined. You're completely ignoring the second half. Your comment below that "the full text of the close reason is irrelevant" absolutely boggles my mind. You also say below "There's no way around what the title of the close reason is!", but the title of the close reason is just a title, not a rationale -- it's not even a complete sentence.

Grace Note's example is excellent (as is the whole of the quoted post). Identical answers to two questions do not ipso facto cause the questions to be duplicates. This has been said in other places, and I completely agree with everyone on this. However, the solutions to the question do have to be taken into account. There were a rash of questions the other day in that were superficially different:

These all have the same root cause, despite each having its own particular take on the problem. (There are plenty of others floating around, too; maybe one calls sleep(), perhaps one uses a UILabel and the other a UITextField.) The answer, and the underlying issue, is identical in all cases.* In other words, they "cover exactly the same ground". They should be closed as duplicates so that one authoritative answer can be established, polished, and referred to whenever the next person has this difficulty.**

An important point here is that the asker of a new question doesn't have to know beforehand that the other question applies (as long as the explanations on the other are sufficient to cover the new one). Two people asking about float arithmetic and precision might not see the connection between their questions, but the explanations are very likely to be usefully equivalent. Which leads me to your statement:

When a question is closed as an exact duplicate, it is a reprimand to the asker

I strenuously disagree with this. The act of closing (especially as duplicate) has little to do with the poster, and everything to do with the maintenance of the site.

Closing a question as a duplicate of an earlier question benefits everyone. The asker still gets an answer, future searchers can find the information they want in a single place rather than having to comb through three dozen different pages, and regulars don't have to rehash the same questions and answers over and over. As time goes on, the answers on the linked question accrete more information, so that their worth only grows.

The web at large already provides a million variations on the same answer to the million different ways to express any given problem. One of the most valuable features of StackOverflow is that all the question variations can be pointed at one most comprehensive and useful explanation.

In short, I think that the problem you have identified is not actually a problem. Questions whose causes are identical, and which therefore have equivalent answers, can already be closed as duplicates.


*You can see this if you read through the answers, although of course it may not be completely clear if you're unfamiliar with the framework.

**Cf. the blog entry The Wikipedia of Long Tail Programming Questions

When a question is closed as NARQ or Not Constructive, it incurs an auto-downvote. Notice that this is not true of "Exact duplicate".

share|improve this answer
    
"The answer, and the underlying issue, is identical in all cases.* In other words, they "cover the same ground". They should be closed as duplicates so that one authoritative answer can be established, polished, and referred to whenever the next person has this difficulty.**" <-- But the questions are still not duplicates. Period. The answers may be, but the close reason is talking about the question, not the answers. More to the point, this does not address my feature request here -- you're answering another question. –  Billy ONeal Jun 20 '11 at 22:40
    
I realize I was on the rambly side, but my point, as I said in my last sentence, is that your feature request is unnecessary because the existing close reason covers the cases you're talking about -- the close reason does include consideration of the answers. –  Josh Caswell Jun 20 '11 at 22:42
    
Further, the "its answers may be merged with another identical question" is not explanation of "here are all the cases where this may be used" -- it is just an attribute of duplicate questions. If the question is an exact duplicate, than of course the same answer applies. However, all of the objective-c questions you posted are not closed as duplicates -- they're not! But they do have similar answers, and perhaps they should be closed for that reason anyway. –  Billy ONeal Jun 20 '11 at 22:42
    
@Josh: No, it does not. The questions are not duplicates. Therefore, calling the questions duplicates is wrong. There's no way around what the title of the close reason is! The close reason does not include consideration of the answers. –  Billy ONeal Jun 20 '11 at 22:43
    
You're ignoring the full text of the close reason. –  Josh Caswell Jun 20 '11 at 22:44
    
@Josh: Read my comment. The full text of the close reason is talking about attributes of duplicate questions. All duplicate questions' answers may be merged, but not all questions with mergable answers are duplicates. The full text of the close reason is irrelevant. –  Billy ONeal Jun 20 '11 at 22:45
    
All duplicate questions' answers may be merged, but not all questions with mergable answers are duplicates. It does not matter if the answers are mergable, if the question is obviously not a duplicate, it's not a duplicate. You can't ignore half of the close reason because of the one sentence. That's like saying C# and C++ are the same because they both use curly braces. You're saying the two things have the same attribute and therefore they are the same -- but that's obviously false. –  Billy ONeal Jun 20 '11 at 22:56

You must log in to answer this question.

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