14
$\begingroup$

Suppose there is an old question on mathOverflow, asked by someone else, which has been voted up a quite decent number of times, and received several answers which have also been voted up, but no complete answer was given (and none was accepted). Suppose I would like very much to know the answer to that question, and that I think there should be specialists out there who know that answer, but don't see the question as it is drown deep in the archives.

What would be the ethical way to bump up this question and make up visible to all, in the hope that it finally receives a complete answer?

I can think of several ways in doing so, none of them completely satisfying: (a) make a trivial edit to the question the make it bump up. (b) make an edit to the question explaining why it has not been answered completely. (c) re-ask the part of the question that has not been answered as my own question.

What do you suggest?

$\endgroup$
  • 28
    $\begingroup$ You could place a bounty on the question. $\endgroup$ – Dilaton Oct 30 '13 at 0:33
  • 3
    $\begingroup$ I'll just mention that bounty not only adds the question to the featured tab, but it also bumps the question. $\endgroup$ – Martin Sleziak Oct 30 '13 at 6:57
14
$\begingroup$

Bumping an old question with an edit is fine, even if the edit is trivial, so long as it isn't done too often! The system already does this by itself but the algorithm is imperfect and some good oldies escape it and would stay dormant without user intervention.

In the event it's not clear why the question needs more attention, an edit explaining that is very appropriate. I would discourage reasking the question unless there is a completely different variant of the question that might attract users with different interests.

$\endgroup$
  • 7
    $\begingroup$ I think a bounty is more appropriate in these cases than an edit. And in the bounty comment the user who is interested in further answers can explain the reason for the bounty. $\endgroup$ – Kaveh Oct 30 '13 at 3:39
  • 4
    $\begingroup$ This is tangential but "The system already does this by itself but the algorithm is imperfect" seems an unfortunate formulation to me in particular given the fact that there is quite a bit of confusion around what bumps the comunity user does. In any case the community user would never bump any question with answers with positive score (and I think it is even more restrictive than just that); now one could consider this as an imperfection of the algorithm but rather I would say it is designed for something quite different. $\endgroup$ – user9072 Oct 30 '13 at 10:26
  • 3
    $\begingroup$ I disagree a bit about re-asking. The OP says the question has many views and several highly voted answers none of which completely answer it. I imagine that many would notice the question on the front page, realize it was bumped by an edit rather than a new answer, consider it already answered, and not bother. But if he asks a new question, linking to the old one and pointing out that the answers there show x and y but not z, then people might take a crack at it. Am I completely off base here? Note also that the software won't bump this question if it has an answer with 2 or more votes $\endgroup$ – David White Oct 30 '13 at 13:11
  • 2
    $\begingroup$ @DavidWhite: The new question would get closed as a duplicate, unless it is clearly different from the original (as I recommended). $\endgroup$ – François G. Dorais Oct 30 '13 at 13:14
  • $\begingroup$ @quid: The bumping algorithm used to be better tuned for that but I think it changed after the new bounty system. Unfortunately, the new bounty system hasn't addressed all the issues we had with bounties. $\endgroup$ – François G. Dorais Oct 30 '13 at 13:18
  • 1
    $\begingroup$ I for one would not vote to close a new question which clearly stated it was part of a previous question but was never answered. Especially with motivation from research, which the OP seems to have. Sometimes one question contains too many parts and they don't all get answered. Regarding bounties, if there were no new answers, wouldn't it just go to one of the current (incomplete) answers? $\endgroup$ – David White Oct 30 '13 at 13:31
  • $\begingroup$ @DavidWhite: Unfortunately, votes to reopen don't count until the question is closed. However, asking just part of a question seems like a substantial change to me. In that case, I recommend editing the original question to redirect that part to the new question. $\endgroup$ – François G. Dorais Oct 30 '13 at 15:46
  • 1
    $\begingroup$ @DavidWhite according to the documention, no, it would not be awarded at all (automatically) as there it is mentioned only question created after the bounty was started are eligible. (However this was also said like this on MO 1.0 while it wasn't like this, so I don't know what would happen in practice.) For convenience "If you do not award your bounty within 7 days (plus the grace period), the highest voted answer created after the bounty started with a minimum score of 2 will be awarded half the bounty amount. If there's no answer meeting those criteria, the bounty is not awarded to anyone." $\endgroup$ – user9072 Oct 30 '13 at 22:56
  • 1
    $\begingroup$ On the bumping: All the time I followed this, which is quite a bit before the change of the bounty system, the only questions I ever saw getting bumpred are those that are in 'unanswered' while having an answer, ie, have answer(s) but none with positve score. And the following SO FAQ at least only says that unanswered questions will be bumped. $\endgroup$ – user9072 Oct 30 '13 at 23:03
12
$\begingroup$

This seems like a perfect case for a bounty, which since MO 2.0 can also be placed on other user's questions.

  1. It will "activate" the question just like an edit.

  2. It will give it extra visibility via the feature tab.

  3. It will make a lot more clear that somebody activated the question for a purpose and not just as a consequence of making some actual small edit. (In my observation old questions bumped by edits get relatively little additional views.)

The only drawback seems to be that it 'costs' 50 points, but then this is not that much. (I think a small one is perfectly sufficient.)

While I do not see it as a problem to do this via an edit, I believe it will be a lot less effective. In particular, I would advice against doing a trivial edit without further explication. Chances are the question will be mainly ignored and somebody will complain about the trivial edit here on meta.

$\endgroup$
  • 2
    $\begingroup$ I summarized the bounty advice given in various comments. In my opinion it really deserves to be also mentioned in an answer as this seems like the use case for new bounty system. $\endgroup$ – user9072 Oct 30 '13 at 10:19
  • 1
    $\begingroup$ Apart from the above, as far as I recall, the question will even be bumped a few times during the active bounty (I don't recall the frequency of this). $\endgroup$ – Tobias Kildetoft Oct 30 '13 at 11:57
  • $\begingroup$ I agree that trivial edits wouldn't help here. A good edit would say EDIT (10/30/13): the current answers still leave open ........, which I am very interested in knowing because........ I still think asking a new question has a better chance of getting answers, though. $\endgroup$ – David White Oct 30 '13 at 13:32
  • $\begingroup$ What happens if I set a bounty on an old question that is already marked as "answered"? Not knowing this meta discussion, I reasked a question that already appeared in an old question. The main question in that old MO post had been answered and so it was marked as "answered". Would putting a bounty on it have worked? $\endgroup$ – Helene Sigloch Dec 22 '14 at 9:38
  • $\begingroup$ @HeleneSigloch I couldn't think of anything that would be particularly different. (There might be a detail in case of auto-awarding in case you are the original poster, but even this is not clear.) Technically it would have worked, that's for sure, to place a bounty on somebodies else alray answered question. See an older answered question with an active bounty Note this is on another site but using the same software. $\endgroup$ – user9072 Dec 22 '14 at 10:12

You must log in to answer this question.

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