tea.mathoverflow.net - Discussion Feed (Is there any consensus as to what an 'accepted' answer should be?)

Ilya Nikokoshev comments on "Is there any consensus as to what an 'accepted' answer should be?" (574)

Ilya Nikokoshev — Sat, 21 Nov 2009 06:18:47 -0800

It appears (e.g. this thread) that the "accepted answer" thing fits the working model of programmers much better than it does for mathematicians. We are using this feature, of course, where it fits us, but if it doesn't, I would say, feel free to experiment a bit.

I personally tend to think that it would be preferable for questions to eventually have an accepted answer that summarizes the current community knowledge. For example, when there are several complementary answers, you can write their summary as a community wiki post and accept it. If a much better answer appears, you can re-accept it.

Whatever you do, it makes sense to clearly say by editing the original question.

gmuller comments on "Is there any consensus as to what an 'accepted' answer should be?" (568)

gmuller — Fri, 20 Nov 2009 19:39:25 -0800

I have so far had no small amount of indecision regarding what kind of answers to accept to the questions I have posted. If I accept a partial answer or a less-than-perfect answer, I am worried that it will discourage subsequent answers. However, I don't want to wait only for perfect answers to accept them, because it feels like an insult to someone who provided a reasonably good answer to leave a question without an accepted answer.

I'm sure that there is no concrete answer, but I was wondering if the moderators had a rough outline of the direction and purpose they want the 'accepted answer' function to serve. It might also make a good remark in the faq, to give a rough outline of when to accept an answer.