42

There is one mechanism currently that discourages too many answers on one question, and that is the autocommunitywikification after 30 (on some sites only 15) answers. This is an ugly and ineffective hack that abuses the CW feature solely for its reputation denial.

I propose a more effective and radical replacement, it should simply not be possible to add more than 30 answers to a question. There are several reasons why limiting the number of answers would be a good idea:

  • A question with more than 30 answers is likely not constructive anyway, so no harm done
  • Few people read beyond the first page, which means the later answers don't get as much review by the community as they should
  • Because many users won't read all the answers if there are so many, duplicate answers are very likely to occur

Only undeleted answers should be counted for this, to prevent non-answers and spam from taking up spots.

To prevent bad answers from blocking a spot, it could be also designed in a way that negatively voted answers would be pushed out by new answers, or other similar mechanisms.

Improve this question
16
  • 7
    What about edge cases, like the Townhall chat digest?
    – Raghav Sood
    Mar 9, 2013 at 9:44
  • 7
    This would not need to be enabled on metas, only on the main sites.
    – Mad Scientist
    Mar 9, 2013 at 9:45
  • 2
    I see what you did with the title, there! </CaptainObvious> I think I like the general idea of this. Hmmm... is there a list somewhere of questions with many answers? (say, more than 10 or something...) that seems like a decent possible tool/review/something. Or not...
    – Andrew Barber
    Mar 9, 2013 at 9:50
  • 1
    @AndrewBarber stackoverflow.com/search?q=answers%3A31
    – Mad Scientist
    Mar 9, 2013 at 10:01
  • 1
    Well, now; that works! hehe
    – Andrew Barber
    Mar 9, 2013 at 10:01
  • 5
    It's not just the non-constructive questions. Many of the popular questions attract a steady stream of low-quality and duplicate answers that pile on. Many of these are "not poor enough" to warrant moderation deletion, but get upvoted enough to where they can't be community deleted.
    – Mysticial
    Mar 9, 2013 at 10:05
  • To prevent answer blocking, what about having Cummunity auto flag a question for mod attention once a question gets at least 25 answers? That way a mod can go in an manually attempt to clean it up. It would also draw the question to their attention in case it is worthy of closing.
    – psubsee2003
    Mar 9, 2013 at 10:07
  • 6
    One of the biggest fears that I have is the branch predictor question accumulating enough "answers" to where it gets auto-wikied. Currently it seems to pick one up every 2 weeks, half of which stick (don't get deleted). I take a lot of pride in that answer and I would hate to see my name and flair removed from it because of wiki.
    – Mysticial
    Mar 9, 2013 at 10:12
  • 2
    Or another idea to prevent blocking or automatic deletion is maybe a special review queue in the moderator tools for 10K users and mods for questions with a high volume of answers so we rely on the community to help keep the answer count low? I'm not a fan of an automatic mechanism since it is difficult to make it perfect. I'd rather have human eyes on it.
    – psubsee2003
    Mar 9, 2013 at 10:14
  • 9
    I'm thinking why not get rid of the entire feature, and have the community and mods take care of it through protecting and deleting, possibly combined with @pbubsee's auto-flag suggestion above? Does this (questions with > 30 answers) happen on a frequent enough basis to be a real world issue at all?
    – Pekka
    Mar 9, 2013 at 10:27
  • 3
    @Pekka웃 Moderators don't delete bad answers, so relying on them to keep this in check would be problematic. This would be something the community has to do.
    – Mad Scientist
    Mar 9, 2013 at 10:39
  • @MadScientist It's not necessarily bad answers as it is NAN and duplicates. A good question worthy of keeping open that gets 30 answers is almost guaranteed to have non-answers and duplicate answers.
    – psubsee2003
    Mar 9, 2013 at 10:47
  • 4
    Questions that gain > 30 answers accrue at approximately 5 a month. In comparison, more then 200,000 questions get asked on Stack Overflow each month. I'm skeptical whether this is a) a problem, and if it is, b) ... worth solving. :/
    – Matt
    Mar 9, 2013 at 13:01
  • 1
    @Matt questions with many answers tend to be highly visible, sticking to the top of collider. Due to their high visibility, quality of answers in these sort of sets the general level people (especially newcomers) tend to follow
    – gnat
    Mar 9, 2013 at 14:16
  • @Mad Scientist: We do accept flags for and delete bad new answers if they add nothing to what's already there, i.e. repeating what has already been said in other answers.
    – BoltClock's a Unicorn
    Mar 31, 2013 at 5:32

2 Answers 2

Reset to default
11

I agree with this, with a few caveats:

  • This feature should never reach the shores of meta. CW and questions with millions of answers actually make sense to be allowed on meta (MSO and child metas alike).
  • If a question is made CW via mod powers, then it should allowed to accumulate as many answers as it likes. Firstly, it provides an escape hatch for edge cases which I haven't thought of yet. Secondly, for sites like Mathematics, which allow questions1, this allows them to continue without radical changes.
  • This should only take into consideration non-deleted answers.

As a side note, I don't really see the point of CW for answers either (except on metas where it makes editing painless). I have yet to see a (new) example of an answer to a non-NC question which really would benefit from CW. And the no-rep part of CW doesn't make much sense either, given that polls are now disallowed.

1. Yes, they probably shouldn't allow these, but that's a different matter which probably will lead to tons of meta debate that I don't want to get into

Improve this answer
2
  • -1 for "they should not allow big-list questions". These are important, and not only on math.SE
    – yo'
    Mar 9, 2013 at 14:11
  • 3
    @tohecz: Like I said, that's a different matter. There's a reason why I put it into a footnote, so that I don't have to debate it. If I wanted voting to be decided by that I would have kept it in the main post. :/
    – Manishearth
    Mar 9, 2013 at 15:04
7

Doesn't the escape hatch for this circumstance already exist, in the form of Closure for a question? In essence you seem to be arguing for a form of "Auto-Closure", but I don't yet see that the problem is so extensive as to require that. The list of questions with over 30 answers is only 500 questions long, which is not a lot given the size of SO.

Improve this answer
4
  • 1
    I don't know how to do that query, but I'm curious as to what percentage of these were "protected" by a moderator along the way, which I understood to be the minimalist mechanism (to prevent 'thanks' and 'me too' responses).
    – hardmath
    Mar 9, 2013 at 13:57
  • 1
    Take a look at the StackExchange Multicollider (where your inbox is). Every day, a few of the top qs has hundreds of answers. That's a problem.
    – Manishearth
    Mar 10, 2013 at 13:54
  • @Manishearth: You say "a few of the top qs has hundreds of answers"; But is that one or two or three NEW answers each day for a FEW questions? That seems easily handled by the exisitng manual procedures.
    – Pieter Geerkens
    Mar 10, 2013 at 15:30
  • @GangDownvoted: Doesn't matter, their high visibility sets a bad precedent meta.stackexchange.com/questions/171063/…
    – Manishearth
    Mar 11, 2013 at 3:14

You must log in to answer this question.

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