79

Currently, authors can't delete their own question

  • if it has multiple answers
  • if it has a single answer with one or more upvotes, even if it has a negative score
  • if it has an answer with an awarded bounty

I ask that another criterion be added, to prevent authors from deleting their question within 24 hours of an answer being posted.

It happened to me, and to others as well, that a relevant question of reasonable quality was asked, some time consuming research went into creating an answer, and as soon as that answer was posted, the OP immediately deleted the question. This happens often with questions asked for work purposes, and some companies seem to generate many such short lived questions every day.

This behavior is hard to provide feedback on, or to moderate for multiple reasons:

  • The community loses content (the question and the answer) without having a realistic time zone independent chance to notice or vote on it.
  • There is no natural way to contact the OP without spamming unrelated questions.
  • There is no way to flag the behavior as deleted questions are not flaggable by ordinary mortals
  • The current UI does not discourage this behavior in any way, so it may not even be intentional.

Also, per Tim Post at the Meta Super User question:

We have things in place specifically to stop people from potentially wasting their time by answering questions. You can't delete your own question if it has multiple answers, or a single answer with a score of one or higher. That generally stops it, what remains is the edge case of great answers not getting a chance to see votes before being removed.

Addressing the edge case is very simple, however.

I believe that even a relatively short protection period is sufficient to ensure an upvote on any high quality answer; but, more importantly, habitual or intentional answer snatching will become less practical, and honest question posters will better understand the public purpose of the site.

(While this does make it a little harder to delete one's own question after the OP genuinely changed their mind, there is a plenty of better ways of stopping to make a fool of oneself: improving one's question through edits, self-answering it, asking the answer owner for cooperation on deletion, waiting a day before deletion, flagging the question for closure, or requesting to be dissociated from the post.)

  • 1
    I have not experienced this "answer snatching" you speak of. In my experience, this kind of behavior occurs when the OP realized a silly mistake he has made. – Lix Jun 11 '12 at 17:37
  • Sounds good to me, though maybe I'd prefer answers with negative scores to be excluded. But then that would imply: with negative score from votes not cast by the question asker, making things just more complicated. – Arjan Jun 11 '12 at 17:43
  • @Lix, the deletion rules were changed to handle systematic self-deleting. But even with those rules, there's a few reports. But also: a recent one that was simply flagged for deletion by the OP, as self-deletion was not possible... – Arjan Jun 11 '12 at 17:56
  • While I'm strongly in favor of being able to easily see your deleted content regardless of rep, its more important that users be able to get rid of bad content, even if it allows for moderate abuse. Remember deleting posts can result in a question ban so it can't be used to continually abuse the community. – Some Helpful Commenter Jun 11 '12 at 21:38
  • 20
    I not only posted an answer, but did at least two additional substantial iterations of the answer as the user added additional constraints (as the question evolved from "how do I draw semi-circles" to "how do I do that in drawRect" to "how do I do that with CoreGraphics"). You've been bitten by a help vampire. Some of them do delete their questions when they're satisfied. Use garlic and stop providing assistance after the first iteration. – Frédéric Hamidi Nov 17 '13 at 23:38
  • (For your reference: help vampire and help vampire) – Aza Nov 17 '13 at 23:46
  • 2
    @FrédéricHamidi lol. But isn't the way to kill vampires with the light of day? Why let the vampire delete their questions that let them hide in the inky shadows? (ok, I've taken the metaphor too far.) My question is less about this vampire, than our ability to let good answers be deleted if they do it quickly. – Robert Ryan Nov 17 '13 at 23:52
  • If one answer has at least 1 upvote, the question can no longer be deleted. It's unfortunate that this didn't apply in this case (and the queston is on its way to undeletion) – Pëkka Nov 17 '13 at 23:58
  • 1
    +1 – Mac Nov 17 '13 at 23:58
  • 2
    And it's back... – Jon Clements Nov 18 '13 at 0:02
  • @Mac I was surprised when I didn't find that when I searched before posting, but I appreciate your reference to that link! Good to know this already has a little inertia behind it. – Robert Ryan Nov 18 '13 at 0:19
  • +1000000 - Now how do we get this to happen? – Dawood ibn Kareem Jan 2 '15 at 4:02
  • @DavidWallace - I'm pretty sure that +1000000 of genuine votes will make it happen. – Jirka Hanika Jan 2 '15 at 9:41
  • 2
7

If a question is truly valuable, there is a very limited window where it can be deleted. The current system is designed to allow for rubber duck instances, where the exercise of writing the question alone leads you to an epiphany that obviates the question. Some examples would be:

  • Duh, I missed a (brace, semicolon, chance to have my morning coffee)
  • I am totally solving the wrong problem here
  • Wait, this code actually works, the problem has to be somewhere else

... or other scenarios where someone would like to avoid wasting other people's time needlessly. And that's what it boils down to, wasting other people's time. If it was an honest mistake, and there's really no value in keeping the question and possibly one answer around, then we chalk it up to things happen from time to time.

We also strongly encourage self-answers in lieu of closing or deleting if there's even a chance that the post could be beneficial to someone in the future. Many users opt for this, even in oh, duh! instances.

On the flip side of this, we will not tolerate someone asking a perfectly valid question, getting a great answer, then trying to remove all traces of it. It's understandable that some folks work under rather draconian policies either at work or school, but it's up to them to work around those suitably. Should this happen to you again, here is what you do:

  • Flag it for moderator attention, select 'other' and let them know what happened
  • Moderators will contact the community team and request that we disassociate the post from the original author's account
  • The community team will do this, and restore the post, and possibly have a private chat with the question author if the moderators haven't.

We don't put up with toxic behavior anywhere, and this is no different.

I think that the real answer to this is making the disassociation of content a more self-service sort of feature, but there are numerous obstacles in front of making that a reality.

Tightening up deletions as proposed solves the narrower / rarer 'hit and run' problem, but could lead to a bit more wasted time in the benign cases where there really wasn't a problem to begin with, or it's nowhere in the scope of the original question.

I'm not declining this because I want to chew on it a bit more, but I think the key here is making it much easier for users to get their name off of certain contributions in a hurry, which seems to be the impetus for the rapid deletion. That just gets a little complicated.

The goal of course being minimizing wasted time as a whole.

  • 7
    I get it. But I still can't help getting caught up in the weird cognitive dissonance of "you cannot delete a question for which there is an up-voted answer, but its ok if you do it so quickly that no one has a chance to vote." And the proposed remedy process suffers a bit of a cognitive dissonance of its own, whereby "if a question is deleted there is a process for remedying this, but we won't tell you it was deleted and we'll make it almost impossible for you to find the deleted question later." Don't get me wrong, I hear you, but it seems internally inconsistent. – Robert Ryan Nov 18 '13 at 4:58
  • 2
    It seems quite logical to provide "delete" button that says "there is answer, so you have to wait 24 hours." Or perhaps in the first 24 hours, users cannot delete outright, but can only "vote" to delete their own question if there are any answers and add their reason to delete in the comments (so future reviewers have basis for affirming or rejecting delete vote). – Robert Ryan Nov 18 '13 at 4:58
  • 2
    I appreciate a new answer to an old open question. As it stands now, your advice to "flag it" isn't an action I am allowed to take. I can't even see it anymore! There are many alternative approaches to fix this problem, like cutting a small and natural hole into the 10K restrictions. (As a minimum: give everyone notification, profile listing, read only access for 30 days, and the ability to flag, as long as they are a co-author of the deleted thread - meaning, they answered or edited (etc.) a question before it was deleted. The same privilege already exists for the question owner.) – Jirka Hanika Nov 18 '13 at 13:14
  • 5
    -1 for the penultimate paragraph, which seems to be going off in entirely the wrong direction. You're very likely correct that one major reason why some askers do this is because they're doing something naughty (e.g. cheating on a take-home exam) and want to hide any evidence of it, but surely the right answer can't be to make hiding the evidence easier. In fact, the fix proposed by Jirka above might actually discourage such users from using SE as a free cheating service, which I'd consider a net win. – Ilmari Karonen Apr 15 '16 at 16:30
  • 1
    a compromise approach is to allow 24-hour access window allowing non-10K answerer see the answer to self-deleted question (possibly along with the question itself but I'm not 100% sure about that). That way, if they believe that question and answer make a valuable content, they could recover it as self-answer (as proposed here) or save it off-site for future use etc – gnat Apr 15 '16 at 17:53
  • 1
    If anyone wonders why I finally accepted after 4 years of keeping this open, this comment is here for you. 1. I stopped answering substandard questions of the "help me" type. So I get burned less often. 2. I reached 10K privileges meanwhile - and so the problem is solved for me. I can now actually use Tim's offered solution "Flag it for moderator attention". If you are a <10K user, I suppose that Tim is advising you to flag an unrelated post, and describe the invisible post as well as you can and hope that a moderator will be able to find it. – Jirka Hanika Aug 27 '17 at 14:24
  • 5
    I upvoted this years ago, but now realize that because deleted questions are practically invisible, the only people who are likely to raise those flags are people who were involved with the question during its brief lifetime. The main one of those is the person who wrote the answer. An experienced user knows to flag, but this is a terrible experience for a newcomer -- you come to the site, contribute your knowledge, and it all disappears with no trace. Is that person going to return? – Monica Cellio Jul 8 '18 at 15:17
11
+50

Analysis

The problem here is that not everyone has the super-cow powers to see deleted posts. As a result, I recommend that you don't trust SO as the sole repository of your posts.

My Personal Solution

I edit all responses in Tomboy, and paste them into SO when I'm ready. That means I often lose out to "fastest gun in the west" responders, but it also means that I ultimately don't have to care what SO does with the data I post because I always have my own copy.

Answer Your Own Questions

If someone deletes a question to which you have a truly unique answer, there's nothing stopping you from asking (and answering) your own question. Just make sure that both the question and answer aren't duplicated elsewhere, or your posts will probably be flagged and/or deleted.

  • 8
    I think asking your own question to re-post the answer after it was deleted is a great idea, but I don't believe the OP has their post, and a lot of people type their answers directly in the site. – animuson Jun 11 '12 at 18:00
  • So you you save the question too, or would you just try to re-write the question based on the answer. (I don't know about you, but I suck at Jepordy.) – Servy Jun 11 '12 at 18:16
  • Maybe it would be useful to allow user's access to answers they have posted on deleted content? It might be complicated to do, relatively, but this shows a good use case. – Latty Jun 12 '12 at 0:13
5

Questions can be undeleted, but somebody needs to have a handle to the question in order to vote, flag, or appeal on meta. Deleted questions aren't very discoverable, so if the person who answered doesn't know what to do about it, it'll probably stay deleted. It would be better to add a little friction to the deletion instead of cleaning up after a subset of these abuses.

A delay on delete votes after a new answer is posted would allow time for the community to vote. If the answer is good, one vote stops deletion. If the answer is not good, there's no harm in a little delay before removing the question and answer -- unless you're trying to hide your tracks from your professor, in which case that's your problem not ours.

How long should the delay be? Good question. Somewhere between a couple hours to a day or two, probably. The ideal answer might depend on the site, but even a fixed limit across the network is probably better than the loss of goodwill from people who contributed valuable content only to have it wiped out unilaterally. New users in particular won't know how to get it reversed, or even what happened. They'll just see their work...gone. Should we be surprised if they don't give us a second chance?

At a time when we're concerned about retention, I think it's worth trying a little harder to prevent this kind of answer deletion.

  • I originally posted most of this on this related question before being reminded of this question. – Monica Cellio Aug 24 '18 at 22:44
  • @TimPost - This is the answer I'd obviously accept instead of yours if I was still a new user (i.e. any user who does not yet have the 10K permissions). – Jirka Hanika Aug 25 '18 at 11:43
  • You raise good arguments here. I proposed a variant on your idea that tries to address the arguments on both sides; I'm curious to see if you have any thoughts on it. – D.W. Sep 3 '18 at 2:09
1

Based on the other answers here, I'll propose a slightly different version of what you mentioned:

  • Allow the owner of the question to click 'delete', but if it is within 24 hours after an answer was posted, delay the actual deletion until 24 hours after the answer was posted. At that point, check the vote count on the answers. If any answer has any upvotes at that point, the deletion doesn't take effect; otherwise, it takes effect 24 hours after the answer was posted if the question is still eligible for deletion.

The benefits of this proposal are that it still allows question-owners to delete questions that they later realized are bad (e.g., a typo or something else where the question will never be useful again in the future); but it prevents askers who delete to prevent anyone else from seeing the answer.

This was inspired by Monica Cellio's answer, and is an attempt to balance the competing concerns and achieve (as much as possible) the best of both worlds.

  • Do you mean "it takes effect 24 hours after the answer was posted if it is still eligible"? Or do you mean just a delay, no further check? – Monica Cellio Sep 3 '18 at 2:11
  • @MonicaCellio, the former -- and that's an excellent way to describe it, thank you! edited! – D.W. Sep 3 '18 at 2:13
  • 1
    I disagree with this idea. If users want to hide things from their teachers, they can request that the post be dissociated from their account. – Sonic the Bracketed Hedgehog Sep 3 '18 at 2:36
  • 1
    If it doesn't get an answer that even one person upvotes, does it matter? The problem here is people giving good answers and having them yanked away immediately. – Monica Cellio Sep 3 '18 at 3:42
  • The delayed nature of the operation would have to be made clear to the deleter via a dialog; and the dialog would have to point them to dissociation option (if their reason to delete is such that they cannot wait for 24 hours). So at the point of the actual deletion they'd often no longer be listed as authors of the post; and they couldn't withdraw the delayed deletion from proceeding even if they realized that dissociation was sufficient for their needs, other than by upvoting the answer. It would end up fairly complicated to implement and use. – Jirka Hanika Sep 21 '18 at 5:49
  • @JirkaHanika, as far as I know the interface doesn't currently point users to disassociation if they try to delete and cannot, so I don't understand why the interface would need to point them to disassociation if deletion will be delayed. I do agree that the interface would need to indicate that deletion will be delayed. Currently disassociation is extremely rare. The situation you mention (where a post is deleted where disassociation would suffice, and where no answer is upvoted, and yet for some reason we want to keep the question) seems so rare that I think we could tolerate it. – D.W. Sep 21 '18 at 6:25
  • @D.W. - I was trying to confront your idea with Tim Post's tenet that "the system should allow for rubber duck instances". I'm in no way opposed to your idea as a potential solution. – Jirka Hanika Sep 21 '18 at 7:24
  • @JirkaHanika, OK, thanks. To me, disassociation seems orthogonal and not terribly relevant to the "rubber duck" situation (and it's probably not something we want to be recommending or encouraging in the ordinary course of events). – D.W. Sep 21 '18 at 19:08

You must log in to answer this question.

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