3

Is there anything that Stack Exchange does to avoid "race conditions" in answers?

That is, when several people see the same question, and try to answer it at the same time, with the result that several people post redundant information.

I'm not worried about it in Stack Exchange, but there's another web site that I use (lang-8.com) where there's a similar "race condition" problem, and I'd like to know how Stack Exchange handles it, to see if it can be implemented on that web site as well.

Just to be clear, I'm not talking about the Fastest Gun in the West problem, where people decide to write low-quality answers quickly rather than take more time and write high-quality answers.

6
  • 2
    No. Should there be done anything about it? – juergen d May 12 '13 at 9:59
  • One might say that the FGITW (to be more precise: the ability to easily edit answers after posting, with or without a grace period) is actually handling this? People might be discouraged to post their similar answer, seeing the someone already posted a few lines. – Arjan May 12 '13 at 10:06
  • @juergend arguably. Redundant answers are clutter we keep forever and ever. They're certainly not "useful" and if you believe the +/- overtext, redundant answers should be downvoted, but we don't do that currently. – djechlin May 12 '13 at 10:23
  • 2
    @djechlin How is it bad/a problem though? Okay, one answer might not add anything over another answer. But if they are both perfectly good answers, the duplication of them does not matter a whole lot, I'd say. Unless one of them is posted a considerably long time after the first one, which I would downvote. But that's not the premise of this question. – Bart May 12 '13 at 10:43
  • You get noticed that there is a new answer. That's it. You can either click on it to see it or go ahead. – Johannes Kuhn May 12 '13 at 12:12
  • @Bart if they're different explanations it's useful to keep. User may not understand first, may understand second. If they are very close to identical then it's extra information we're making perusers of this site waste their time determining they already just learned it. I'm not saying this should be a high priority problem, just that yes it's not perfect. – djechlin May 12 '13 at 13:08
2

Is there anything that Stack Exchange does to avoid "race conditions" in answers?

No.

Redundancy in answers is pretty common across a lot of tags. See this as an example. Take a look at the timestamps and revisions as well. Every person ended up with the same answer and suggestions. Is it a problem ? Not really.

Although, this was a very basic question, so there isn't much scope for diversity, on tougher questions, different approaches evolve in different answers.

Also, this does relate to FGITW issue.

5
  • The only person affected by FGITW is the first answerer. I'm not sure how that explains why people would post an identical answer later (or wouldn't delete their own if it really was a race condition). AFAIK you can't get the FGITW badges if you were the second or later answerer, even if the first answerer later deletes theirs. Sadly, many of the badges promote one or more good behaviors while simultaneously promoting other, less desirable behaviors. In the past I've suggested removing the FGITW badges altogether. – Aaron Bertrand May 12 '13 at 12:35
  • 1
    @Aaron, as for "The only person affected by FGITW is the first answerer", I'd say that the second answerer, who posted a much more complete answer before the first answerer extended theirs in the grace period, is affected much more. Not for badges, but for (not) getting the upvotes that the very first answer somehow tends to get. – Arjan May 12 '13 at 13:55
  • 1
    @Arjan sorry, I meant the only person who gains. I've suggested before that the initial state of the post should be in the permanent history (status-declined even though I don't see how it harms anything) so that people don't get in line first with crap and then improve it. We should also just get rid of the badges that promote this behavior. – Aaron Bertrand May 12 '13 at 13:58
  • @Arjan and the reason I stated that is that subsequent answerers have very little motivation to post an identical answer in the first place. Now, if someone is gaming FGITW, the system currently has no way to show users that the "first" answerer posted crap and edited within the grace period. In some cases second+ answerers can actually lose in this scenario, as many people advocate down-voting identical later answers - even though it can't be certain they were identical at the time they were posted. There are easy ways to fix all this but Stack doesn't seem to want to try... – Aaron Bertrand May 12 '13 at 14:12
  • (Agreed, @Aaron. And I very much support that feature request.) – Arjan May 12 '13 at 14:13
6

Etiquette would suggest that later, identical answers should be self-deleted, and I've seen this happen quite often (even when the difference is only minutes or seconds). This doesn't always happen, of course, for at least two reasons: (1) rep greed (2) the answers aren't actually identical.

I don't really see a problem with multiple answers that essentially say the same thing, as long as it's not verbatim. Sometimes a different way of explaining the same thing can be more helpful to the OP (or to a future reader).

I would probably argue that if a question garners multiple identical or similar answers very quickly, it is probably a close candidate - either as a duplicate the user didn't bother to locate, or as RTFM they didn't bother doing before coming to the site in the first place. :-) I think folks tend to find it easier to just answer the question rather than find a duplicate.

I agree with Bart that an identical answer posted significantly later is undesirable behavior, but for me it has less to do with time and more to do with similarity of content - an identical answer posted a minute later or an hour later is still an identical answer, and time stamps make it easy for the later poster to decide if they want to delete their answer or improve it to make it unique. At least ideally. What happens in practice...

8
  • 2
    Third reason: deleted answers count towards the answer ban. – Antony May 12 '13 at 12:06
  • @Antony self-deleted answers? Really? I've deleted many of my answers for this and other reasons, but I suspect it would be ratio-based rather than absolute. Seems counter-productive and not really helpful re: discipline. In any case, in a round-about way, that could fall under (1). :-) – Aaron Bertrand May 12 '13 at 12:09
  • Really. Some people delete their own questions (because they want to clean up some of their not so stellar questions) and result in a ban, and required mod attention to undelete some of them. I assume this happens to answers as well. I agree this is counter-productive. But probably the ratio may come into play. No one knows the exact formula of the ban. – Antony May 12 '13 at 12:12
  • @Antony I believe the post ban look only on downvoted deleted posts, deleting zero scored post should not have any effect. – Shadow The Vaccinated Wizard May 12 '13 at 12:13
  • @Antony: When one has got quite a number of answers, I don't think there is any need to worry about answer ban if some identical answer is deleted now and then. – nhahtdh May 12 '13 at 12:18
  • Right, I can understand that for answers with down-votes, but not for answers with a score of 0 or higher... which is likely the case when you genuinely post a duplicate at around the same time as others (which is also more likely for softball, non-constructive questions). If that gets someone penalized I'd suggest it should be fixed. – Aaron Bertrand May 12 '13 at 12:24
  • 1
    "I think folks tend to find it easier to just answer the question rather than find a duplicate." Truth. Unfortunately, it is way easier (and faster) to just answer the common questions than finding a good dupe. By the time one has found a dupe, there are already about 3-5 answers with 5-10 upvotes most of the time. – Daniel Fischer May 12 '13 at 15:16
  • @TinSoldiersAndNixonsComin' As the number of unique questions goes down over time, I think maybe there is a case for making the cost of asking a duplicate question higher. There seems to be a conflict though - the site seems to love duplicate questions because it generates more traffic and SEO. But in the tags I frequent, more and more questions are questions that have already been asked - and answered - over and over. They keep getting answered because of exactly what you state - it is easier (and more profitable) to answer than to close as a dupe. And there is no disincentive to the asker. – Aaron Bertrand May 12 '13 at 15:42

You must log in to answer this question.

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