2
$\begingroup$

Do we have a protocol on how to handle acceptance of answers on CW type questions (including those that aren't CW but of a similar flavor)?

There's clearly no correct answers for most CW questions. Some of the questions are explicitly big-list questions, so a unique answer wouldn't even make sense.

For example on StackOverflow, I think the convention is to always accept an answer, even on CW questions. On MO, the convention seems to be to not accept any answers on CW questions, or accept randomly based on whim.

There's probably no need for a rule to decide what everyone should do, but are there any suggestions for what would be a good thing to do?

| |
$\endgroup$
7
$\begingroup$

Accepting an answer has the important property of hoisting it to the top of the list of answers, regardless of what the distribution of votes is or when it was answered. If you think one answer does a good job of framing the whole thread, that would be a good reason for accepting it.

| |
$\endgroup$
4
$\begingroup$

That is why I do not like those big-list questions on a Stack Exchange website. My understanding is that if you post a question on Stack Exchange, you are expected to accept one answer. If you can foresee that you will not know which answer to accept, why do you ask it on Stack Exchange in the first place?

| |
$\endgroup$
  • $\begingroup$ I think CW was explicitly created for questions that don't have one answer, or where one should not gain reputation for posting an answer. So given that CW exists, I think the Stack Exchange overlords expect us to ask questions that don't have one answer. There questions are usually big-list questions where the aim is to collectively accumulate knowledge on some subject in an easy-to-access place on the internet. $\endgroup$ – Robin Kothari Sep 19 '10 at 1:24
  • $\begingroup$ @Robin: Do the questions like that have a specific goal? I do not mind occasional big-list questions, but currently I see them too often on the front page, and I decided to ignore all of them for a while. $\endgroup$ – Tsuyoshi Ito Sep 19 '10 at 3:26
  • $\begingroup$ I guess some big-list/CW questions are a bit subjective, and so the specific goal requirement has to be relaxed a bit. But many of them are not, for example the "problems between P and NPC", "When did computers help prove a theorem" threads, and "How many open problems from this list are still open?" threads. Some questions like "best paper to read" and "best algorithms" are subjective, but at least it is clear what an acceptable answer is --- any paper or algorithm. Upvotes will then decide whether the answer was good or bad. $\endgroup$ – Robin Kothari Sep 19 '10 at 3:53
3
$\begingroup$

I have never understood how to do this, and the listing of 'acceptance rates' creates IMO unnecessary pressure to accept. The problem is that the number of accepted answers is one of the many stats being kept by the SE folks during this beta period. Granted, it's not the only one, but still...

My recommendation is to go the MO way, and ignore it.

| |
$\endgroup$
  • $\begingroup$ For your information, the acceptance rate of each user is shown not only in the beta period. $\endgroup$ – Tsuyoshi Ito Sep 18 '10 at 22:17
  • $\begingroup$ yes that's correct. I meant that the global stats on accepted answers were one the factors being used. $\endgroup$ – Suresh Venkat Sep 19 '10 at 2:43
  • $\begingroup$ The list you mentioned has given me absolutely no pressure, because I do not know where that list is. $\endgroup$ – Tsuyoshi Ito Sep 19 '10 at 3:21
  • 1
    $\begingroup$ By the way, you may know this already, but community-wiki questions are not included in the calculation of the acceptance rate of each user. I do not know about the global stats. $\endgroup$ – Tsuyoshi Ito Sep 20 '10 at 14:41

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .