10

I don't know if I made it very clear in my title, but here is my question, if I have a question, that can be perfectly answered by the answer to another question, should my question be marked as a duplicate of the other question just because the answer to the other one solves my one perfectly even if my question and the other question couldn't be more different? I've always just assumed that you copy the answer (maybe rewrite it in your own words) and then credit the answer you got all the stuff from.

But I saw a different approach to it today when @A.B. voted to close this question as a duplicate of this one just because the answer to it answered both questions, even though the two questions really couldn't, in my opinion, be any more different.

So which approach is it best to take in these circumstances?

Information Update:

Oh, I have just been informed that @A.B. has retracted his close vote, however my question still stands.

1

4 Answers 4

9

Most opinions I have seen about this say that questions are duplicates, not answers. Therefore, questions should only be closed as a duplicate when they are the same, not because the answer to one is also the answer to the other.

I, however, disagree. I see the XY problem as common enough occurrence that I consider whether:

  1. the OP is trying to achieve some greater target via this question, and
  2. the answer to some other question also answers this question, and
  3. that answer leads towards a solution (or, better yet, is the solution) for this greater aim of the OP.

If so, I vote to close as a duplicate. That's to say, that OP formulated their problem in a particular fashion is not very relevant if it is an XY problem - it could have been in another form quite easily.

I'm sure many would disagree with this view. :shrug:


For the particular case you show, no, I wouldn't have marked as a dupe - both A and B could very well lead in different directions (you just want to see more detailed history - the form not being important, the other user might have wanted to process their history, etc.).

1
5

I tend to agree with Muru. Answers also come into play when deciding whether something is a dupe. For example, consider these two (fictional) questions:

  1. How can I rename a file whose name contains a space?

  2. Why does my shell choke on whitespace in file names?

If the answers to Q1 contain a detailed explanation of why the shell has trouble with such names as well as a solution for renaming them, I would consider closing Q2 as a dupe of Q1.

Strictly speaking, the questions are not asking for the same thing. However, the point of the site is to provide answers. If the answer has already been provided, there's no need to repeat that information. Whether the question that gave rise to the answer is 100% identical to the one considered a duplicate is secondary in my opinion.

The whole point of duplicates is to i) avoid duplicating effort and ii) organize information better. Keeping a question that has a perfectly good answer elsewhere open just because it is phrased differently seems pointless. I am aware that sentiment on the SE network is often against this, but it just seems silly.

As far as I'm concerned, I will vote to close as a duplicate if the question has been answered elsewhere. To make that choice, I consider both the question and its answers.

12
  • As one might expect, I strongly disagree with this. Apart from good answers, findability plays a major role in the profit of a site. Some overlap hurts less than routing answers through other questions that might answer the question sideways. Sep 8, 2015 at 12:59
  • @JacobVlijm why is findability affected by dupes? They still point to the right place. If anything, they make finding the answer easier. Oh, and please downvote if you disagree, this is meta after all :)
    – terdon
    Sep 8, 2015 at 13:04
  • If I am looking for an answer, I won't look for questions that I don't understand they might give an answer. Most likely, I won't find the answer, unless after an exhaustive search through (as far as my understanding goes) related questions. Keep in mind that all available information on AU is probably in no more than 10% of the answers (is my assumption). I am convinced the other 90% is to meet the lack of understanding at the moment the question is posted, bad titles, tags etc. etc. Sep 8, 2015 at 13:14
  • Not a downvoter by nature, but here it goes :) Sep 8, 2015 at 13:14
  • 1
    @JacobVlijm the theory is that you search for answers that could answer your question, find the dupe, follow it and find your answer. And hey, downvotes are important! Especially here where we're trying to get community consensus.
    – terdon
    Sep 8, 2015 at 13:15
  • I think the theory is incorrect. At the moment of asking, I am not aware question B could answer question A. Sep 8, 2015 at 13:17
  • 1
    @JacobVlijm in my example, Q2 is closed as a dupe of Q1, not the other way around. So, if you look for "shell whitespace" and find Q2, you will then be redirected to Q1 which, despite appearing irrelevant, happens to contain the answer you needed.
    – terdon
    Sep 8, 2015 at 13:18
  • 1
    From the days I was looking for answers on a regular base, I can assure you, searching for the appropriate information in a question that also answers my question, among others, is no fun. Many times I switched to the internet to find the appropriate and specific information to prevent having to read through a lot of information I was not looking for. Again, if that was the idea, we could trash (at least) 90% of the q/a. Sep 8, 2015 at 13:26
  • 2
    If Q1 has seven different answers, only one of which adequately answers Q2 but is several answers down, it might be hard to find the answer to Q2 in all the answers to Q1. Sep 9, 2015 at 17:34
  • @tepples exactly my point against making answers (also) the criterion. Sep 9, 2015 at 20:36
  • @tepples the close voter should (always, not just for this case) have the sense to specify which answer they intended, if it's not the accepted one. Whenever I vote to close against How to customize the live CD? I always mention that it's Rinzwind's answer - since that's the one that works currently.
    – muru
    Sep 9, 2015 at 21:18
  • @tepples of course. All I'm saying is that answers should also be considered. Having one answer among 20 that also answers another question is not enough to make something a dupe target. Having the answer, a nice, long, highly upvoted and comprehensive one might.
    – terdon
    Sep 9, 2015 at 22:49
4

Here's how I like to boil it down:

  • If ALL answers this question is likely to receive would also be valid answers to the other question, it IS a duplicate of the other question.

Otherwise, it's not a duplicate. This includes when:

  • If there is overlap between the questions: if SOME answers to this question would also answer the other question, but other answers might not, then it's NOT a duplicate.

In the example case above, these questions are not duplicates according to my rules.

2

I definitely would say the only criterion to mark as a duplicate is the question.

There are many cases answers answer more than one question. For various reasons, that is irrelevant to the decision to mark as a duplicate. Clarity and findability are the most important of them. Someone with question A is not going to look for his answer at question B, although the answer to B might solve A as well.

From a broader perspective: also in many cases, in the answer to one question, the answer to another question can be concluded, although the answer nor the question is the same. Would we also mark them as a duplicate?

Let's keep it clear; a question only is a duplicate if the question is an exact duplicate.

You must log in to answer this question.