So I just asked http://mathoverflow.net/questions/22837/what-are-the-uses-of-the-homotopy-groups-of-spheres and jc and Andrea spotted that it was sort of related to http://mathoverflow.net/questions/16495/applications-of-homotopy-groups-of-spheres. I like to think that my version was a better question (well, I would, wouldn't I? But it also seems backed-up by the comments and answers to the two questions) and I think it has also garnered better answers.

So my main question: should I do anything about this? Should the moderators? Should I walk down Olav Tryggvasons gate in sackcloth and ashes?

Here are the options I can see:

• Do nothing. This is a lame answer, but a definite possibility. At the very least, we should edit both questions so that they link to each other. This would still have the disadvantage that there won't be a canonical place to add a new answer or to read answers.
• Close one as a duplicate of the other. This makes sense, but of course there's the issue of which one to close. Your duplicate has far more activity (in the form of answers and comments) than the original. Also, it might make sense to merge the contents of the two questions into the one that remains open.
• A moderator can merge one question into the other. This would move all the answers over, so there would be a single page where you could read answers and post new answers. The disadvantage is that one of the questions (along with its comment thread) will be completely removed. This could be mitigated by editing the content of one question into the body of the other.

The last option is the most appealing to me if it can be done without anybody objecting. But it's an irreversible procedure (unlike closing), so I really want to make sure it's the right thing to do.

Since the second question is more articulated than the first and got better answers, I think things may be left as they are now. Also, I do not see the difference between options 2 and 3 of Anton.

A more symmetric solution: close both questions, and open a third question which refers to both of them and accepts answers to either. (Moderator(s) use their judgement as to how to introduce both questions. Community uses their judgement about how to frame the answers.) Try it and see what happens.

Gerhard "Ask Me About System Design" Paseman, 2010.04.28

@grp: I'd rather leave both questions open. I feel like increasing the number of pages you have to visit to get to the answers is a pretty bad thing. Right now it's two. I'd like it to go down to one, but I really don't want it to go up to three.

Morally, I'm in the wrong as mine was the second question and I should have seen the first. However, joking and modesty aside, I do think that mine was the better question so I would want mine to be the dominant question in a merge (if a merge is chosen).

It's possible that if I'd seen that original question then I wouldn't have asked mine. However, it's also possible that I would have seen it, read it, thought something similar to what Harry just said, and then been in a quandary about what to do. Pete's challenge was just so irresistible to my inner algebraic topologist (and, clearly, to others) that I'd want to have somewhere to record good answers. So my choice then would have been to either post the question anyway (with a link to the original and an explanation of why I didn't like it) or to edit the original, essentially imposing my question over the top (and thus making the original answers look daft).

So it goes back to the lesser-of-two-evils question: is it worse to post a question that is essentially a duplicate, but somehow better phrased, or to edit an old question (the age should be taken into account here: I think it is more acceptable to edit a fresh question than an old one, especially when the old one has some comments and answers)? I can anticipate several answers on the basis of what one should do in the best of all possible worlds, but there's also a pragmatic angle: if I'd just edited the original question, would anyone have looked at it? I judge it a higher priority to have good answers (to good questions) than to have a minimum number of links to click when searching for a particular question. Thus I would accept a small amount of crud to avoid missing the gems.

(To be clear: I'm arguing from the lesser to the greater here; I wouldn't classify the original homotopy-of-spheres question as "crud" nor mine as a "gem". Also, note that I said "a small amount"; clearly either extreme is wrong and the question is more about where to draw the line.)