3
$\begingroup$

I recently came across an "answer" to a question which didn't actually answer it. The question is genuinely difficult, and the "answer" merely got rid of a large number of counter-examples, and so was really a comment not an answer. Also, the "answer" was posted within an hour of the question being posed.

I subsequently have two (related) questions.

Should an answer posted so soon after the posing actually answer the question?

I believe that they should. Otherwise, the answer leaves the "unanswered" queue, even although it does not actually have an answer. The leaving of this queue is less of an issue of a question has gone unanswered for a long time.

If so, what should be done about those answers which only provide partial answers?

I flagged the answer I mentioned about as "not an answer", but my flag was "disputed". Should I dispute this dispute? Or is the community sympathetic to quick-fire partial answers? I know that I could simply re-post the question and link back to the original, but them my question would simply be a re-posting of the original but saying "I actually want an answer, not a partial answer!". Which is silly! (Note: I don't actually want to discuss this specific answer, I am just using it as motivation and as a handy example.)

$\endgroup$
  • $\begingroup$ (This is related, but not entirely relevant.) $\endgroup$ – user1729 Jul 7 '14 at 11:21
  • 1
    $\begingroup$ The "not an answer" flag is not for partial answers. It is for things which should have been comments (and where this can be seen without necessarily knowing the math behind) or stuff that simply is not related to the question. $\endgroup$ – Tobias Kildetoft Jul 7 '14 at 12:22
  • 1
    $\begingroup$ @Tobias I don't see why the sentence "It is for things which should have been comments (and where this can be seen without necessarily knowing the math behind)" is not applicable here... $\endgroup$ – user1729 Jul 7 '14 at 12:24
  • 4
    $\begingroup$ Because it requires knowledge of the math behind to see that it does not fully answer the question but only removes some counter-examples. Also, if this answer is sufficiently long, then removing a bunch of counter-examples seems like a perfectly fine thing to do in an answer (as long as it states that this is what it does). $\endgroup$ – Tobias Kildetoft Jul 7 '14 at 12:28
  • $\begingroup$ @Tobias Your point seems to be "in practice, it only works for this". However, in theory it should hold for these questions, no? $\endgroup$ – user1729 Jul 7 '14 at 12:38
  • $\begingroup$ I am not sure where "in practice" came from. The official policy is that moderators should not be judges of mathematical content (except possibly the lack of it). $\endgroup$ – Tobias Kildetoft Jul 7 '14 at 12:42
  • $\begingroup$ @Tobias Okay, sure, I understand. So what if it was a short, two-line post which states that it is merely a reduction? (I think I agree with your comment on length - I suppose I am also imagining that the partial answers are not complicated.) $\endgroup$ – user1729 Jul 7 '14 at 12:47
  • $\begingroup$ I don't see anything wrong with giving partial answers as an 'answer'. Yes, they are philosophically hints (and so more akin to comments), but not space-restricted. As long as they point out that they are merely ideas/counter-examples/exceptions/etc., I honestly fail to see any reason to prevent the 'answer' section from occasionally serving as a scratch-pad. If you see a useless partial answer, I'd downvote or vote to close. $\endgroup$ – gnometorule Jul 7 '14 at 13:47
  • 3
    $\begingroup$ @gnometorule (You can't VTC an answer.) Here's one reason. My filter includes answers: 0, which means when someone posts an answer that isn't, the question becomes invisible to me (though I might have an actual answer). I don't think I'm the only user who pays more attention to unanswered questions. $\endgroup$ – user147263 Jul 7 '14 at 13:51
  • $\begingroup$ @Thisis: Point taken, but I still see use for leaving a partial solution which might form the basis for a full one. That's why in an ideal world acceptance of an answer should indicate it's closed. Of course OP mentions leaving such a partial answer too fast, and that's a good point; but I still don't see how this could or should be prevented by new rules, instead can be addressed by the means already available (downvotes and voting to close). $\endgroup$ – gnometorule Jul 7 '14 at 13:53
  • 6
    $\begingroup$ @gnometorule Only Chuck Norris can vote to close an answer. $\endgroup$ – user147263 Jul 7 '14 at 15:26
1
$\begingroup$

I am adding another answer, because it concerns a purely technical point raised in the post:

I flagged the answer I mentioned about as "not an answer", but my flag was "disputed".

Flags used to be disputed in the 10K flag queue but it does not exist anymore. So, who disputed the flag? Here is a (slightly edited) quote from the linked post:

Not an Answer flags go into /review/low-quality ... until one of the following conditions is met:

  1. Post is edited from within review.
    Outcome: flags are marked "helpful"
  2. Post accumulates $3$ Delete votes (can only happen when post scores $\le 0$ and reviewers have $\ge 20$K rep).
    Outcome: post is deleted, flags are marked "helpful".
  3. Task accumulates enough "Looks OK" reviews.
    Outcome: if the number of (Recommend)Delete reviews is $\ge$ the number of Looks Good reviews, mark flags "disputed" and raise DisputedLowQuality mod flag. Otherwise, mark flags "declined".
  4. Task accumulates $6$ RecommendDelete + Delete reviews.
    Outcome: mark flags "helpful". If the post scores > 0 then raise DisputedLowQuality mod flag, else just delete post (current behavior).

It seems that in your case, at least half of the jury voted for execution, but others insisted on the pardon. The case went to the Supreme Court of Diamonds, but you won't see the outcome of that because they will handle a new, automatic flag. Your flag stays "disputed" regardless of whether the post lives or dies.

$\endgroup$
  • $\begingroup$ Hmmm. Compare this crap from yesterday, after a nice answer by Hagen where he forgot to carry a constant along: math.stackexchange.com/questions/857774/… $\endgroup$ – Will Jagy Jul 7 '14 at 17:55
  • $\begingroup$ @WillJagy Hmmmm? I don't see how it's relevant to my post. $\endgroup$ – user147263 Jul 7 '14 at 17:57
  • 1
    $\begingroup$ May not be. General idea being that there may be some answerers who are not doing it right, but there are more ungrateful and lazy people asking questions. $\endgroup$ – Will Jagy Jul 7 '14 at 18:06
6
$\begingroup$

I flagged the answer I mentioned about as "not an answer", but my flag was "disputed".

If a post takes a stab at answering the question (e.g., solves a special case), I'd expect an NAA flag to be declined. A moderator cannot be expected to judge the significance of the advance made.

Some flagging guidance from SE, with answers replaced by apples:

A, NAA, A, NAA, VLQ

In order: no flag, NAA, no flag, NAA, VLQ


Should an answer posted so soon after the posing actually answer the question?

I think it should, but this is something that can't be enforced with a flag. Although I am on the anti-comment side in general, this is a situation where an explanation is called for.

I'll take a stab at it. :)

Quick posts that do not answer the question have a side effect: they may deprive the question of the extra attention that other unanswered questions get. Consider addressing the question more fully. You can temporarily delete the draft of an answer, and undelete when it is more complete.

$\endgroup$
  • $\begingroup$ The graphic says it nicely :) $\endgroup$ – rschwieb Jul 7 '14 at 16:32
  • 1
    $\begingroup$ +1, if only for the picture. :) $\endgroup$ – apnorton Jul 9 '14 at 3:28
  • $\begingroup$ The picture is nice but also loaded with extra assumptions. For example I think of my own partial answer as being a clean cut half-an-apple - with the caveat that the matching other half may not exist. I somewhat like pic #4, but many think that posting a link to the lecture notes published at the homepage of professor X (together with a couple lines) is an answer. Or a link to a related question at MO. I'm undecided myself. $\endgroup$ – Jyrki Lahtonen Jul 9 '14 at 11:22

You must log in to answer this question.

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