# Should identical answers on questions which are abstract duplicates of each other be allowed?

Let's say that a question is posted, asking how to find $$\sum_{k=1}^n2k^2.$$ It is proposed that the question be closed as a duplicate of a much older one, asking how to find $$\sum_{k=1}^nk^2,$$ but before the closure is complete, a user posts identical answers on both questions explaining how to find something like $$\sum_{k=a}^nbk^2.$$ My question is, should this be allowed, or should the user be asked to delete one of the copies? Something similar to the above (though with different math) occurred recently, and I'd just like to get some community input. I am strongly in favor of deleting the copy on the closed question.

Martin Sleziak has found some relevant meta.SO links and posted them in the comments, and I will include them in my post below. Of course, we have no requirement to follow the SO community's policies.

User posting exact duplicate answers to multiple questions

Can we hash answers and check for duplicates (upon submission)?

-
Just to be clear - the two answer were exactly identical, word by word? Were they posted at the same time? I am asking to find out whether the possibility that the user simply posted new answer with similar wording is out of question. (Some users here have plenty of answers, it is possible to forget about answer you posted or it might be not that easy to find it. For example, I remember posting some explanation about Hamel basis in one question and then realizing I've posted very similar stuff earlier to a related question and forgotten about it.) – Martin Sleziak Jun 22 '12 at 6:23
I don't see what the harm is in having the answer in two different places. – Gerry Myerson Jun 22 '12 at 6:37
@Martin: yes, they are word by word identical (discounting some subsequent minor edits); and they are posted within minutes of each other. – Willie Wong Jun 22 '12 at 9:04
@Gerry: I suppose some people may see it as reputation gaming. – Willie Wong Jun 22 '12 at 9:05
@Willie, some people may, and they might even be right. But if the answer is good for two questions, why shouldn't it appear for both questions? People looking for an answer to the question, their finding the good answer shouldn't depend on which version of the question they chance upon. – Gerry Myerson Jun 22 '12 at 12:42
@Gerry: I think that is sort of the point of "closing as duplicates". – Willie Wong Jun 22 '12 at 13:10
I don't think that there is a good answer to this unless it is that we always merge duplicates. Arguments for both positions have merit. I think it would be very unfair to force the user to delete one of the answers, because there are many examples of such duplicate answers already, including probably answers by almost everyone on the front rep page. Whether or not this is allowed should not depend on the timing. The problem is really that we don't handle duplicates well. They really should be merged into the latest version, so that the question gets exposure to new members. – Gone Jun 22 '12 at 13:49
I'm afraid I caught myself doing this yesterday. I had a recollection of having typed "the same" answer earlier but couldn't find it at first, because one parameter was different. Later I deleted the new answer and added a link to the earlier question. I guess that could also be viewed as fishing for reputation, so it is not ideal. As years tick by and the regulars start forgetting about their (and each others') old answers, this is bound to get worse. If somebody can suggest a good way of dealing with this, I'm all ears. Meanwhile, let's just assume that everybody is acting in good faith. – Jyrki Lahtonen Jun 22 '12 at 15:28
I see it as very natural that at different times one might post very similar answers to very similar questions. But posting to an old problem and a new one more or less simultaneously seems unreasonable. – André Nicolas Jun 22 '12 at 17:27
+1 @Bill if we were to have a robust and aggressive policy on merging duplicate questions, we can reasonably argue against such behaviour by virtue of "we don't want two copies of the same answer appearing on the same question post merger". As it stands (especially since we sometimes let abstract but not exact duplicates remain as individual questions not linked to one another), I think at least Zev, Bill, and I agree that right now the issue is quite murky, and hence the question here seeking community input. – Willie Wong Jun 25 '12 at 8:39

I don't see a problem with it.

I think our main focus should be on how to best help people find answers to their questions, and having a good answer to every question that might get Googled is a step in that direction.

If a few users post the same good answer to multiple questions hoping to "game" reputation, who cares? There are fewer unaswered/poorly-answered question on the site, and if they get upvotes, somebody found their answer helpful. I don't see a perverse incentive.

-
I downvoted not only because I do think some action should be taken, but also because I generally disagree with the notion of "who cares about games?", this is a slippery slope that tends to get very sloppy, very fast. True, mathematics is a natural barrier, but that too can only hold for so long. We should do our best to avoid games, especially games in the form of answers (edits, votes, reviews, things which are in the "background" are somewhat tolerable, but not in the answers which are the front part of the site). – Asaf Karagila Jul 2 '12 at 21:55
My point about games is the following: individuals want to maximize their reputations. The site wants good answers posted to every question. The site works as long as these two interests are aligned. A problem occurs whenever there's a way to gain reputation without helping, or by harming, the site. People who post duplicate (good, relevant) answers, even out of self-interest, end up helping the site, which is why I don't see a problem in this case. – user7530 Jul 2 '12 at 22:23
-1: It seems to me your argument would also support never closing any questions as duplicates. What I think you're forgetting is that the site also wants to be well-organized and searchable, and this is the (primary, at least) reason for not allowing multiple identical copies of content. – Zev Chonoles Jul 3 '12 at 2:04
@Zev Chonoles: My searching abilities are at best modest. However, various questions, with various wordings, about very similar topics would seem to increase the likelihood of a hit. – André Nicolas Jul 3 '12 at 2:47
@Andre: Are you therefore not in favor of closing questions as duplicates? Having one general question and answer, created to absorb abstract duplicates, seems to me to be far more preferable than a sea of barely-different questions, each with its own mixture of the various ways there might be to solve the problem. It requires more maintenance on our part, and I agree that it can often be the case that a person will not see the general post in a search, but at some (IMO very early) point, each new abstract copy of a question produces another hit in a search but less utility for a user. – Zev Chonoles Jul 3 '12 at 3:01
@ZevChonoles: Generally I am not in favour of closing. (Exception: when someone resubmits a question a short time after first asking it.) Abstract duplicates don't mean much for people with insufficient experience or skill in abstraction. An answer to a concrete problem that the person has (one hopes!) worked on teaches much more. Even better is several somewhat different approaches. A link to other similar questions is certainly useful. The closing of duplicated questions is in any case somewhat haphazard, sometimes quick, sometimes after several answers. Usually it doesn't happen. – André Nicolas Jul 3 '12 at 3:19
Suppose that we adopt a policy of allowing a user to repost their answer on any question where it serves as an answer. Do you have a suggested delay, or can someone repost a sufficiently general answer on hundreds of questions within minutes? What about someone who skips the middleman altogether and, (ab)using the policy of allowing one to ask and answer one's own question, posts an arbitrary number of questions $$\text{Solve }\int x+1\;dx?$$ $$\text{Solve }\int x+2\;dx?$$ $$\text{Solve }\int x+3\;dx?$$ $$\cdots$$ and then posts an identical general answer on all of them? – Zev Chonoles Jul 3 '12 at 5:19
After all, there are hypothetical people out there (unable to see any generality at all) who might be faced with any one of these versions. Is a user who does that "helping"? Without closing these questions as duplicates / merging them, an unscrupulous user would be able to accumulate an essentially arbitrary amount of reputation for knowing one actual thing, while polluting the site in general (when you search, and are confronted with hundreds of these posts, which of them had that one other answer you're thinking of...?), and the front page in particular if they do it fast enough. – Zev Chonoles Jul 3 '12 at 5:19
@ZevChonoles: Extreme cases, and more particularly extreme hypothetical cases, make bad law. And downvoting would surely discourage such egregious behaviour. – André Nicolas Jul 3 '12 at 7:41
@Andre: I would hope so. And regardless, we would correct any damage done by closing (or merging) those questions as necessary. But I don't see what reason there would be for people not to repost their answers to as many similar questions as possible. Sure, a halfway decent person would do it slowly, so as not to crowd the front page, but reposting answers is "helpful" to the OPs and will get them huge reputation gains compared to any downvotes they may receive, so I am imagining (eventually) every similar question will have the same list of tens or hundreds of answers. – Zev Chonoles Jul 3 '12 at 7:54
Why would reposting an answer give huge reputation gains? Old answers certainly do pay reputation "dividends" but the rate of gain is very low for questions that already have answers with many upvotes. – user7530 Jul 3 '12 at 8:04
Each reposted answer would bump the relevant question to the front page, giving it a chance to be voted on by an unsuspecting public (I don't think there's been any discussion yet about disclosure - can we agree that if someone reposts verbatim an old answer of theirs, this should be announced, with a link?) The reposting of the answer many times is what would give huge reputation gains. If an answer is reposted on a hundred similar questions and half of those reposts get a single upvote, that's 500 points. – Zev Chonoles Jul 3 '12 at 8:12
500 points? Arturo gets that many in his sleep! – Gerry Myerson Jul 3 '12 at 13:02
@Gerry: But not from posting duplicated answers, rather he gets it from well-written answers. – Asaf Karagila Jul 3 '12 at 13:27
@Asaf Probably every high rep user has many duplicate answers (partly due to the fact that it is often quicker to repost an answer than to locate an old one with the poor built-in search) – Gone Jul 3 '12 at 22:47