7
$\begingroup$

I have been seeing answers accepted a very short time after being posted.

Waiting less than 24 hours means that people in some time zones may not have the opportunity to read a question before an answer is accepted. This does not explicitly prevent an answer but I think it can discourage both reading of the question and posting of additional answers.

If users intend to accept the first good answer, as soon as possible (or less than 1-2 days after posting), or need some answer instantly, I think it should be considered an unusual use case and the posters should be encouraged to state the high-speed intended use of the question. Otherwise it can turn the site into a race instead of a forum that produces slower, more thoughtful answers.

Examples of very fast acceptance:

less than 1 minute delay: Prove that the numerator of $H_{p-1}$ in reduced form is a multiple of $p$ for $p$ an odd prime

less than 3 minute delay: Proving that $k[a,b,c,d,e,f]/(ab+cd+ef)$ and $k[x_1,x_2,x_3,x_4,x_5]$ are not isomorphic

This happens frequently; there is nothing unusual about those two links except that I knew them from having participated as an answerer.

EDIT: 24 hours seems like a canonical minimum waiting period, but the exact number is less important. It is not uncommon to see one correct answer, then a better answer an hour later, and a really excellent answer two hours after that, and I am primarily raising the question of whether acceptance of the first answer within minutes is something that should be discouraged. Whatever the reasons are for discouraging it, would probably also support a minimum waiting time measured in hours.

$\endgroup$
14
  • 1
    $\begingroup$ p.s. this is not quite a duplicate of the earlier "length of time before accepting answer" meta question. If you are voting the question up/down to register agreement/disagreement, I am proposing that anything below 24 hours be considered exceptional, and that posters be encouraged (but not required) to declare it in the question if they have a need for quick answers or an intention to quickly accept answers. $\endgroup$
    – zyx
    Commented Nov 6, 2011 at 21:11
  • 11
    $\begingroup$ I voted down to express my disagreement with a rule or encouragement to declare an intent to accept answers in less than 24 hours. This does not mean that I am against informing new posters that it might be a good idea to wait a bit. $\endgroup$
    – Phira
    Commented Nov 6, 2011 at 21:56
  • 1
    $\begingroup$ @Phira: thanks for spelling out your view in addition to the vote. I think it would be good to see the spectrum of what people think about this, especially about the extent to which any concrete "rules" or norms should be developed. I would oppose rules but not FAQs or (sufficiently loose and flexible) norms. $\endgroup$
    – zyx
    Commented Nov 6, 2011 at 22:00
  • 2
    $\begingroup$ I haven't voted on this question, but I'd like to add a few remarks. I don't think that as a rule (*) answers should be left unaccepted for 24 hours, but I support the idea. I made similar comments as you did in the threads you link to: Most recently here, where the timeline reveals quite a bit of indecision by the OP which answer to accept after my intervention. I was always a bit hesitant to leave a comment when another answer was accepted (I don't remember that I ever did that). $\endgroup$
    – t.b.
    Commented Nov 6, 2011 at 23:31
  • $\begingroup$ (*) I agree with Gerry's answer in the thread you mention in your first comment that it depends on the question. $\endgroup$
    – t.b.
    Commented Nov 6, 2011 at 23:33
  • 3
    $\begingroup$ @t.b. -- of course encouragement or discouragement of delayed acceptance could (and, I assume, would) depend on the question. In the case of questions as Gerry describes, that have a unique answer (such as reference-request for specific works) there is no sense in a delay. For questions that have multiple interpretations or multiple possible answers, which is the large majority, it's not clear to me that the better-to-wait principle is very question dependent. $\endgroup$
    – zyx
    Commented Nov 7, 2011 at 0:02
  • 8
    $\begingroup$ I am in favor this. Basically, because it discourages the single shot night before homework is due visits to the site. The lessons often take time to sink in (unless the whole point was to address somebody's homework woes). The time zone argument I also agree with. $\endgroup$ Commented Nov 7, 2011 at 8:16
  • 5
    $\begingroup$ Here's a modest proposal: let's ask them to tweak the software so that no answers appear until 24 hours after a question is posted. You can submit an answer anytime, but it will be "embargoed" until 24 hours are up: then, all of a sudden, dozens of answers will appear from all over the globe, leaving OP with plenty of choices. I'm surprised nobody thought of this simple idea before! $\endgroup$ Commented Nov 7, 2011 at 11:55
  • 6
    $\begingroup$ @Gerry I think this would lead to many people answering the same simple question, not knowing that it has been answered already more-or-less in the same way several times. Just notice that many questions get several similar answer during the first few minutes after posting. $\endgroup$ Commented Nov 7, 2011 at 12:53
  • $\begingroup$ This discussion seems to be somewhat related: meta.math.stackexchange.com/questions/2133/… $\endgroup$ Commented Nov 7, 2011 at 14:27
  • 1
    $\begingroup$ @Martin, perhaps you are not aware of the connotations of "a modest proposal". See, for example, art-bin.com/art/omodest.html $\endgroup$ Commented Nov 8, 2011 at 0:37
  • 1
    $\begingroup$ @Gerry: what is the logical connection between your modest proposal and the actual proposal in the question? In a successful Swiftian proposal there is some noticeable parallel to a real situation, which is then inverted to devastating effect. Whereas you have lampooned ideas that are foreign to the present discussion (such as software changes, rules and restrictions, and limits on the functionality of the site). Could you spell out the connection, if you think there is one? $\endgroup$
    – zyx
    Commented Nov 8, 2011 at 0:47
  • 1
    $\begingroup$ @zyx, the "logical connection" is that neither was put forward as a serious suggestion. $\endgroup$ Commented Nov 8, 2011 at 4:57
  • $\begingroup$ @Gerry You're right that I was not aware of this meaning of the phrase modest proposal. (I should have seen that you're not being serious about the embargo thing.) $\endgroup$ Commented Nov 8, 2011 at 7:38

1 Answer 1

18
$\begingroup$

I'm not convinced that there's a problem to solve here, and the proposed solution has the potential to lower the utility of the site for everyone.

Yes, sometimes an asker accepts an answer without taking time to think it through, but where's the real harm in that? There's an answerer who will get 15 perhaps undeserved rep points, but really, so what? It's only points.

On the other hand, blocking an asker from accepting until some arbitrary waiting period has passed will effectively disable the main function of the accept mark, namely to let everyone know that the asker considers his problem to be solved. That will make the site harder to use for answerers wanting to economize with their time.

And really, do you think a first-time asker will check back after 24 (or whatever) hours to accept an answer, perhaps long time after he's completed his proof or turned in his homework or whatever? He'll have no incentive to do so -- except for 2 rep, which will be irrelevant for a casual user. Thus, the likely result would be that almost every question would remain showing as "no accepted answer", until -- perhaps, eventually -- the asker has enough open questions get an accept rate calculated and start to be bullied about that.

That likely outcome isn't going to help anyone.

$\endgroup$
13
  • 2
    $\begingroup$ What does "blocking an asker from accepting" mean? There was no proposal of a rule or a change to the software. Adding a line to the FAQ, or a linkable item on meta, or encouraging some sort of informal social norm, was more the idea. And it is proposed that the instant-acceptance model is explicitly allowed, with the only change being that users should be encouraged to indicate if they just want a quick answer or will accept the first good answer. (Note: encouraged, not required or enforced.) The point being to convey more information to users and answerers, not less $\endgroup$
    – zyx
    Commented Nov 7, 2011 at 16:17
  • 2
    $\begingroup$ I read the proposal, perhaps wrongly, as a suggestion of technically preventing the checkmark button for showing up for the OP until a certain period of time had passed. As for encouraging to wait and see, don't we already do that? $\endgroup$ Commented Nov 7, 2011 at 17:00
  • 1
    $\begingroup$ To some extent it is already encouraged. The only novelty here is the idea of being more numerical in the encouragement, with a definite lower bound ("waiting at least $N$ hours is appreciated"), and also encouraging statements from the posters in cases where they intend to instantly accept or where they want an immediate answer more than a comprehensive set of answers at a later time. $\endgroup$
    – zyx
    Commented Nov 7, 2011 at 17:18
  • 2
    $\begingroup$ How would an asker know in advance whether he's going to accept instantly? He doesn't know yet how well the first answers will solve his problem. $\endgroup$ Commented Nov 7, 2011 at 17:24
  • 2
    $\begingroup$ Also, I disagree that waiting any particular period of time "is appreciated". I certainly cannot see why I ought to appreciate it. How is it going to benefit me -- or anyone else? I fail to see that there's a problem worth solving in the first place. $\endgroup$ Commented Nov 7, 2011 at 17:25
  • 1
    $\begingroup$ Many askers know before any answer is posted, that the minute they determine at least one answer is good enough, they will accept it. That is a completely different matter from forecasting whether any good answer will appear and requires only self-awareness, not prophecy. For an example of this, see one of the links in the question, where the OP describes his answer acceptance procedure. Also "accept instantly" is a short description of something that I spelled out more precisely in the question. $\endgroup$
    – zyx
    Commented Nov 7, 2011 at 17:31
  • 2
    $\begingroup$ Regarding what is appreciated and why, I can't speak for you, but some of the reasons I would appreciate a generally longer wait are: that people other than the OP (that is, the majority of the site users) benefit from having a generally more comprehensive population of answers under each question, so that what satisfies OP is not the only issue; that accepting a "first good enough" answer reduces the probability that the accepted answer is the best answer (which reduces the information content of the checkmark as a signal about answer quality); and discouraging a race mentality on the site. $\endgroup$
    – zyx
    Commented Nov 7, 2011 at 17:37
  • 1
    $\begingroup$ On the other hand, if you ask askers to pretend that their problem hasn't been solved yet when in fact it has, answerers will waste time writing more answers to them, which could have been used more productively (for all of us) by constructing answers to problems that don't yet have solutions. $\endgroup$ Commented Nov 7, 2011 at 17:43
  • 2
    $\begingroup$ Pretense is not a relevant concept here. It is a matter of making clearer how the checkmark is interpreted. The checkmark sends two signals, one about answer quality and the other about whether the OP is satisfied. The latter can be fully (and more informatively) expressed by a "thank you, that solves my problem" in a comment under the first good enough answer, which is another behavior I would encourage. Maybe you prefer giving priority to the "OK, problem solved!" interpretation of the checkmark, but doing so dilutes it as a mark of answer quality, and comments make this unnecessary. $\endgroup$
    – zyx
    Commented Nov 7, 2011 at 17:54
  • 4
    $\begingroup$ The mark of answer quality is the upvote score. As far as I'm aware, the checkmark is exclusively for signaling OP satisfaction. That's not dilution; it's all it was ever meant to be, and trying to make it more than that will just create confusion. Comments do not satisfy that role because a "problem solved" comment is not going to show up in the question list and search results. $\endgroup$ Commented Nov 7, 2011 at 18:01
  • $\begingroup$ Answer quality is signalled by: upvotes, acceptance of answer, comments and comment votes under the answer, and indirectly by answerer reputation (in a broad sense as well as the user statistic). That the checkmark is not "exclusively for signaling OP satisfaction" can be seen by the ability to move the checkmark to other answers. $\endgroup$
    – zyx
    Commented Nov 7, 2011 at 18:14
  • $\begingroup$ This year(2013) I noticed that the time between the question appeared and the answer was accepted, was less than 30 seconds(!!). I'm not native English and I type slowly, but this behavior should prevent. And it is not 15 points, because others seeing that, maybe go on to other questions. And the answer is correct, not trivial, it has the chance that later will be earn other points. $\endgroup$
    – vesszabo
    Commented Aug 8, 2013 at 13:55
  • 1
    $\begingroup$ (I wanted to reedit it and continue, but typing took time more than 5 minutes.) So here is: And the answer is correct, not trivial, it has the chance that later will earn more points. I think 24 hours is much. I suggest 10 minutes. What is the reason that polling about it hasn't been made yet? $\endgroup$
    – vesszabo
    Commented Aug 8, 2013 at 14:29

You must log in to answer this question.

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