-6

This is a synthesis of two feature requests, one quite recent and the other quite old:


When a question is closed as a duplicate of another question, and it remains in the system, it contributes to searches and to aiding people asking a new question who might be unaware that theirs is a duplicate.

For instance, here is a basic scenario:

  • Original question title: Harvesting fruit
  • Duplicate question title: Picking apples [Closed.]

Both questions now contribute to checks against new-question duplication and to providing search results.

But if the duplicate question never received any upvotes or answers, or if everything was received negatively, then it will be marked as a duplicate and deleted automatically:

  • Original question title: Harvesting fruit
  • Duplicate question title: Picking apples [Deleted.]

I do believe that all types of questions should be automatically deleted if they don't meet the criteria that would let them remain. However, what's unfortunate about this is that synonymous question titles are lost along with the deletion. This means that if someone else tries searching for keywords being used by the duplicate, they may also go on to post a duplicate question.

What I propose is that, like tag synonyms, question titles also have so-called synonyms.

In this way, even though the question with the title Picking apples has been deleted, if somebody searches for "Picking apples", the system finds the question with the title Harvesting fruit,

The exact mechanism as to how this would be handled presents several issues, a few of which I will immediately present.

  1. I suspect what should occur is that upon deletion of a question marked as a duplicate, that question's title should be moved into a review queue, just as tag synonyms are put into a review queue. Not until a certain number of people have voted to allow the synonym would it become active.

    This would prevent the use of very poor, too general, or otherwise inappropriate question titles becoming synonymous and being used as a check by the system.

  2. It's also possible that there are simply so many questions that are marked as duplicates and then automatically deleted that such a list would quickly become overwhelming and unmanageable. If that's the case, then there should still be some way of allowing this. Perhaps a purely manual submission of question-title synonyms.

    If I mark Picking apples as a duplicate of Harvesting fruit—or even if I just become aware of it in some other way—and if I have the reputation, I could go to a question-title synonym screen—in the same way as I would for tag synonyms—and manually add the suggested synonym.

  3. What determines who can submit or validate question-title synonyms would follow the same principle of who can do the same with tag synonyms. So, it would be based on reputation. But it would likely also be tied to something else; however, I can't think of the specifics of that at the moment.

No doubt there are several more issues that would have to be addressed in any such implementation.

However, I think that the principle behind this is sound. Just because a question isn't worth preserving in general, if keeping its title is a valid check against future duplicate questions, then we shouldn't be discarding that information along with it.

  • That sounds like a kind of a (AI) machine learning job. I am not sure, but I have a feeling that a load of false positives would be produced doing it that way. – πάντα ῥεῖ Jul 27 at 16:22
  • @πάνταῥεῖ That's why there would need to be a validation step performed by users. Nothing would actually become active in the system unless approved. – Jason Bassford Jul 27 at 16:23
  • Thus every duplicate prone for automated deletion should go to a queue where the community decides? That sounds like an overcomplicatin for me. Also, all the XY problem questions won't be well addressed by that. – πάντα ῥεῖ Jul 27 at 16:28
  • @πάνταῥεῖ I address this in the second point in my answer. If that produces too much volume, I don't think that should rule out the ability to have manually-suggested question-tag synonyms. – Jason Bassford Jul 27 at 16:30
  • 2
    If the duplicate did not receive upvotes it wasn’t a great question to begin with. Why keep that around at all? – Martijn Pieters Jul 28 at 2:09
  • You might just as well preserve those questions that have synonymous terms as duplicates. Wouldn't it be easier? – Mari-Lou A Jul 28 at 16:28
  • @Mari-LouA I honestly don't believe that bad questions should ever be kept. I agree with the criteria for deleting them. If the only reason for keeping them is to provide synonymous question title consideration, then there should be a mechanism for that specifically rather than keeping the rest of the badly asked question. It's similar to babies and dirty bath water. I don't want to throw out the baby with the dirty bathwater—but I don't want to keep the dirty bathwater either. – Jason Bassford Jul 28 at 16:38
  • So how come dupes that have -25 votes are allowed to stay up but a question with +3 but -4 votes can be automatically deleted? – Mari-Lou A Jul 28 at 16:40
  • @Mari-LouA I can't speak to the algorithm used, as I don't know its exact details. As far as I know, it applies to every question, whether it's a duplicate or not. If it does treat duplicate questions differently, that would be new information to me. – Jason Bassford Jul 28 at 16:42
  • A duplicate follows the same rules as closed questions, closed questions are automatically deleted when they have no answers with positive scores and if their score is less than 0. They are termed as being abandoned. – Mari-Lou A Jul 28 at 16:46
  • @Mari-LouA Then it seems you already know the answer to your question about "dupes that have -25 votes," and I'm confused by the comment. Whatever the algorithm, it applies to everything equally. – Jason Bassford Jul 28 at 16:55
  • A closed (and even opened) unaswered question only needs -1 score and that unanswered post (dupe) gets nuked automatically by the system. It happened to someone's post, and it happened to me. I even wrote about it on EL&U meta. – Mari-Lou A Jul 28 at 17:04
  • @Mari-LouA By bad I mean those questions that don't meet the criteria of the algorithm for being kept. If a duplicate question doesn't meet those criteria, then it should be deleted. I see no reason to keep a question merely because it makes use of a useful alternate question title. However, the point of this post is that I don't think the useful question title should be wasted along with the question that should be deleted. – Jason Bassford Jul 28 at 17:14
  • @Mari-LouA In terms of duplicate questions and them having useful titles, I don't care what the mechanisms are for deletion. I'm concerned only with keeping the useful titles in some way. Debate over what should qualify for deletion is a separate question altogether. You've already made it clear that duplicate questions are treated identically to other questions. That's fine. (And that's what I had thought, until it seemed you were saying it wasn't the case.) Low-quality questions (however those are defined) are automatically deleted. Including duplicate questions. – Jason Bassford Jul 28 at 17:21
  • Low-quality questions (however those are defined) are automatically deleted. Including duplicate questions. No, there not. There are dupes with scores of -50 that are not deleted. Keep the dupe if its score is higher than -5 and/or its title is significantly different from the original, keep the dupe if it is older than the canonical question but delete it if its title and content is virtually identical to the original even if it has received answers. Cont'd. – Mari-Lou A Aug 1 at 7:56
1

I am not convinced about that the feature you propose would be a good idea and worth to be implemeted.

Some dupes will stay anyways since they are good signposts already. The others (e.g. lacking research efforts, or asking XY) would be deleted. I believe what's already in action is just fine

  1. I suspect what should occur is that upon deletion of a question marked as a duplicate, that question's title should be moved into a review queue, just as tag synonyms are put into a review queue. Not until a certain number of people have voted to allow the synonym would it become active.

    This would prevent the use of very poor, too general, or otherwise inappropriate question titles becoming synonymous and being used as a check by the system.

IMO that's a way too complicated process and no benefits would be gained for any researching OP to be pointed at possibly related or duplicate questions.

To pick up your arguments:

  • Duplicate question title: Picking apples [Closed.]

The next question titled Picking oranges typed in by an unexperienced user, won't see the dupe, because their title didn't match Picking apples.

The extrapolation when querying for such specific stuff is part of being the OP in charge.

If they're not able to grasp what's in common about Picking fruits from Picking apples I am afraid we can't do much about that, and the question would be probably downvored and deleted (automatically).

In case they are aware of the specifics of their question, and will react in editing their post, to make more clear which specific aspects of Picking apples aren't answered in the duplicate. The question would probably be upvoted over time, and serve as a signpost, or even be reopened and stands alone.


The apple / orange dichotomy I have introduced above also illustrates the problem about XY-problem quite well IMO.

Many cases of duplicates aren't coming with the duplicated questions title (at SO I've noticed these are even rare), but a behavior / code described in the questions body, which has nothing to do with the problem supposed in the title at all.


Thus I am afraid we will have another queue (to drown in) as you're saying yourself (emphasis mine):

  1. It's also possible that there are simply so many questions that are marked as duplicates and then automatically deleted that such a list would quickly become overwhelming and unmanageable. If that's the case, then there should still be some way of allowing this. Perhaps a purely manual submission of question-title synonyms.

    If I mark Picking apples as a duplicate of Harvesting fruit—or even if I just become aware of it in some other way—and if I have the reputation, I could go to a question-title synonym screen—in the same way as I would for tag synonyms—and manually add the suggested synonym.

I honestly believe that the system as working at the moment is sufficient to

  • Remove bad duplicate questions from search results
  • Keep upvoted or reopened duplicate questions

is sufficient and working well.

You must log in to answer this question.

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