18

The title may seem a bit weird, let me explain.

While going through some questions, I found a spammer network about 5 users with each of the users posting answers on others' questions.

Besides these users, some of the community users had also posted their answers.

Now, I proceeded with destroying the spam users. While the questions and answers of spam users were gone, curiously, the answers of community users still remain, even through the question was deleted.

To cite an example: The question remove kaspersky internet security 2011, I mean I need a clean way (10k+ only) had answers from 3 spam users + 2 by community users. I destroyed the spam user accounts, while the question was deleted, the answer by Daniel & Paul remain.

  • 3
    Interesting edge case! – ChrisF Jun 7 '11 at 10:58
  • 1
    it looks to me like they were all deleted... – nhinkle Jun 7 '11 at 17:14
  • 1
    @nhinkle they were deleted later. ( check the timestamps, and discussion in teacher's lounge) – Sathyajith Bhat Jun 7 '11 at 17:16
  • So that's what happens to all my answers... – Daniel Beck Jun 11 '11 at 9:26
  • @DanielBeck heh ;) – Sathyajith Bhat Jun 13 '11 at 4:34
9

There are some scenarios where answers are not deleted when the question is.

(it's also possible for users to undelete their answers to a deleted question in some cases)

Never fear, however, because a scheduled daily task cleans this up -- it goes through and automatically deletes any non-deleted answers to deleted questions.

  • The scheduled task clean up explains why the remaining answers were deleted later. But why defer the deletion when the question is deleted? – Sathyajith Bhat Jun 8 '11 at 4:13
  • @Sathya: I believe it's part of the global reputation recalculation. Given the other reputation differences that can occur due to up and down voting, the remaining answers are probably not worth fixing directly. Other than just deleting the answer you also have to correct the reputation of the user and vote counts of other users, which is better done by a clean up task meant for this purpose... – Tamara Wijsman Jun 8 '11 at 11:30
1

I don't really see the problem here? This looks like to me.

You are destroying an user with the intent that his questions, answers and comments are deleted.

This action should not affect other users or their reputation; think of an edge case where the bad user also asked a very good question which received a magnificent answer, or think of an edge case where you would want to revive a specific question that will be deleted in the progress which means you would have to revert the actions on the answers too ...

In your case, they have spent time on looking for a removal tool so keeping their reputation is worth it.

Changing this behavior would be unnecessary work, as reverting will be more complex (or impossible)...

  • In my opinion a deleted question should always have the same behaviour wrt to not changing the rep of anyone involved, and so questions that need saving (as per your edge case example) should just be undeleted. Note that undeleting a question will undelete the answers that were "deleted" by the question's deletion (and leave those manually deleted by the OP, ♦ or community), so I don't think there's a concern about time taken to revert - it's just a simple question undelete. – DMA57361 Jun 7 '11 at 12:18
  • 2
    Few things (a) Deleting a question deletes all the answers posted (b)undeleting a question will undelete all the answers on the question. (c) When you destroy a user, then you are essentially deleting all their questions - so why should it keep the answers in this case? It's not status-bydesign - it is a bug. – Sathyajith Bhat Jun 7 '11 at 12:30
  • @Sathya: Do you have a reference for your first two statements? Because that's simply not how it currently is implemented as you have shown, and as I explained there is really no useful reason to implement it that way. As for your last sentence, think about why you should delete the answers in this case? ;) – Tamara Wijsman Jun 7 '11 at 18:08
  • @DMA57361: That's what I said with the first edge case. Your note doesn't apply, as shown the answers remain to be there and deleting them isn't really useful. Why implement extra logic that determines by who it has been deleted if the answers can just remain to be there? If you were to delete this meta question, I would consider the deletion of my answer to be a bad thing as I can't get it back through my profile anymore. However, as a user that has moderation privileges I should be able to find my answer... – Tamara Wijsman Jun 7 '11 at 18:14
  • Of course it could be implemented the other way by deleting them too, but it really rather looks like it's an OCD to just get the answers deleted rather than there is something useful to it. I agree with your suggestion as I have nothing against it, but just changing something that seems by design feels rather unnecessary... – Tamara Wijsman Jun 7 '11 at 18:27
  • @Tom, the info that I didn't give out in my first comment (not sure why :/) was that the non-deleted answers on the deleted question couldn't be accessed by any user with under 10k, and they didn't appear on the users' profile pages for non-♦s (we checked) - so these non-deleted answers could have been essentially been giving the users "phantom" reputation (which would be bad). – DMA57361 Jun 7 '11 at 18:29
  • 1
    @DMA57361: You would consider it bad because it's doesn't result in right calculations given the answers visible to an user. But I could consider it good as they got reputation for their good answer... – Tamara Wijsman Jun 7 '11 at 18:34
  • 1
    @TomWijsman it's inconsistent, which is bad. Users shouldn't get rep for answers that don't exist to the public any more. Phantom rep is confusing, and should be eliminated. – nhinkle Jun 7 '11 at 18:56
  • @nhinkle: On the other hand, it's also confusing to suddenly have a drop of 100 to 200 reputation in such edge cases without being able to figure out why. It does happen, unless you were to check all the question when you destroy an user... – Tamara Wijsman Jun 7 '11 at 21:10
  • @TomWijsman that's how the system works. Yes, it's confusing when some of your rep goes away, but that's just part of how it goes. The reputation is only lost when a user's rep is recalculated, either by the user, a moderator, or as an automated process. If you want to make a feature-request to show in the rep report what changed when doing the recalc, you can do that, though I doubt they'd approve it since that would be too CPU intensive. – nhinkle Jun 7 '11 at 21:31
  • er @TomWijsman reference would be the behaviour that I see every day when I delete/undelete questions. I'm not sure how you can claim that's the design without having seen how either one works. The current implementation that you claim is not standard behaviour for a deleted question, but one that pertains specifically when you destroy a user. – Sathyajith Bhat Jun 8 '11 at 4:11
  • @Sathya: I'm not claiming anything, I'm just repeating your example. You could as well interpret the behavior that you see every day as a bug given that you don't have a reference which explains why it is implemented that way, which allows for other ways to look upon this. So it isn't necessarily non-standard behavior... – Tamara Wijsman Jun 8 '11 at 11:34

You must log in to answer this question.

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