# There needs to be a clean way of dealing with abstract duplicates.

(Edited) I found that these four question are very similar: (One) (Two) (Three) (Four), and are abstract duplicates in my opinion.

However each question is "slightly" different then the next, perhaps adding in the word "elementary" or specifying a specific case of the more general formula. I think there needs to be some way to merge these questions as reader of any one would be equally interested by any of the answers contained in the others.

The current system, as outlined here, https://meta.math.stackexchange.com/questions/1756/coping-with-abstract-duplicate-questions suggest doing something along the lines of

Rewrite a more general question which encompasses all the others, flag all the other posts to be closed, and then repost all the answers on the new question as community wiki.

This feels like fixing a hole in a wall with duct tape. It is not pretty, and is not a long term solution.
Using the above method is very time consuming, which makes users less likely to do it. Also, the decision would come down to a user and the moderator, as opposed to a group of users. (Example: Voting to close/open) For this particular case, I am not 100% sure if it should be done, and would like to see that at least a few other users agree with me.

Main Point: I think there needs to be some mechanism to merge questions. I know that closing questions is suppose to fill this purpose, but it often fails to do so. Perhaps the ability to vote to merge should require more reputation, or after 5 votes it additionally must be ok-ed by a moderator. (as it is suppose to be used rarely) I understand there are ways to work around the current system, but none of these are satisfactory.

• Ironically this question looks to me as an abstract duplicate of Coping with abstract duplicate questions. – Asaf Karagila Feb 29 '12 at 17:37
• @AsafKaragila: Should I just rewrite a more general question which encompasses all the others, flag all the other posts to be closed, and then repost all the answers on the new question I created as community wiki? This seems to be extremely time consuming, as this is only dealing with 1 question. Perhaps I should change my question to a feature request which asks for some simple way to merge questions. (a moderator only type feature) If I can think of a good proposition for how to do this, I will post it. – Eric Naslund Feb 29 '12 at 18:56
• If merging is not an option, how about editing each and linking to the others (and possibly to this thread)? – The Chaz 2.0 Feb 29 '12 at 20:02
• @Eric, there is a moderator tool to merge questions. And it has been used a few times in the past. But merging can be problematic when the two questions are not exactly the same. (Minor difference in notations, for example.) – Willie Wong Mar 1 '12 at 11:35

I support the option to merge questions than closing them as abstract duplicates.

For instance, the questions

are essentially the same question, just that OP in one of these questions, wants to solve the problem by a specific method. There are lot of examples like this.

Another advantage of having an option to merge is that, in some instances, a question gets closed as an abstract duplicate of another, but some of the answers provided to the closed question are way better than the answers provided to the original question.

The only issue might be which OP gets the up-vote when the question is merged, though this can be addressed by awarding the points to the question which has an earlier time-stamp.

Similar to the current system where $10$K+ users can vote to close questions, it would be good to have the option to merge available to the $10$K+ users as well. i.e. the questions will be merged when $5$, $10$K+ users vote to merge the question.

• Though the links are generated from comments and close-as-duplicate votes, plus some voodoo about search keywords. – Willie Wong Mar 1 '12 at 11:28
• I personally like it to scan through the links. Sometimes you find what you're looking for or something "really" interesting, that you thought about. – draks ... Mar 1 '12 at 11:44
• @WillieWong I guess that the "some voodoo about search keywords" part of your comment applies only to related questions, not to linked questions. – Martin Sleziak Jun 24 '12 at 17:37
• @Martin: you are right. – Willie Wong Jun 25 '12 at 8:06

In my opinion, once we went to accepting homeworks and exercises from books as legitimate questions, there is no way to put the repetitions under control any more just because students are reading the same textbooks all over the globe and their professors take the exercises from a bit wider but still quite limited collection of canonical texts. Research questions are mostly unique, but training questions tend to be the same and I know people who even believe that they should be the same (i.e., that making exams from scratch and trying to incorporate a novel twist into each problem is actually a bad idea). So, unless we want to repel all struggling students (and the consensus seems to be that we do not), we'll have just to accept that the same questions will appear on the site infinitely many times no matter what.

• Disagree. A) "we" have never went to accepting homeworks and exercises whole sale. B) disallowing repeats of questions does not amount to repelling new users, it simply means that the focus should be in searching as opposed to asking. – Jyrki Lahtonen Jun 3 at 3:34
• The site is just a huge database. The early years were mostly about building the database. Gradually the use should shift towards using the database. Out of inertia, users fail to acknowledge that simple fact. – Jyrki Lahtonen Jun 3 at 3:36
• And also, many of us feel that database maintenance trumps other concerns. The struggling students should adjust their use of the site to fit the needs of the database rather than the other way around. – Jyrki Lahtonen Jun 3 at 3:40
• @JyrkiLahtonen "we" have never went to accepting homeworks and exercises whole sale A) OK, looking at the flow right now. 27 out of 75 first questions are homeworks or exercises (with ratings from -1 to 5, mainly 0). If that is not "accepting wholesale", what is? B) Encouraging search is an excellent suggestion if you can do it in some efficient way. Any idea how to enforce it? – fedja Jun 3 at 5:45
• It simply means that "some users" are accepting of homework. True, a need for compromise exists. Many have been proposed, and some (like the requirement for contex) de facto established. But those same users ignore the compromise. Also, the need to weed out duplicates is rather unrelated to our homework policy. – Jyrki Lahtonen Jun 3 at 5:50
• @JyrkiLahtonen the need to weed out duplicates is rather unrelated to our homework policy Here is where we disagree: IMHO, the duplicates are an inevitable consequence of the current homework policy. If it continues, you'll just not have enough strength to weed them out faster than they arise (I'll be delighted to be proved wrong here, of course). – fedja Jun 3 at 6:03
• Regarding oft-asked standard questions, I think there is a difference between: (A) duplicate questions getting asked. (B) duplicate questions getting asked and answered (C) duplicate questions getting asked and answered and people putting up a fight over their dupe status and/or removal. // I tend to agree that (A) is by and large unavoidable, (B) is tricky to avoid completely but I think we could push back a bit more against this and it might have some effect. (B) is not great yet it's in fact (C) what mostly annoys me at times, and we do have the mean to push back against this. cc @Jyrki – quid Jun 4 at 13:33
• Separately I am not sure if you had noticed that the meta question is more than seven years old. :-) – quid Jun 4 at 13:35