107
$\begingroup$

This is a follow-up for a recent questions of mine:

What are the merits, if there is any, of deleting (correctly) answered questions in the main site of MSE?

I have noticed recently that quite a few old questions with good answers are being closed and deleted (mostly initiated by a same user):

I do not see any merit for closing such old questions: who the heck do we really expect to come back and "provide additional context" for a question that was asked two or even six years ago? The further strong action of deletion mostly, if not completely, erases useful information that people created (with valuable efforts and time) together in the main site.

Can we please STOP closing and deleting old posts with answers?

A related post on meta:

$\endgroup$
  • 14
    $\begingroup$ "mostly initiated by a same user" Who is the user? $\endgroup$ – Isa Apr 27 at 15:20
  • 14
    $\begingroup$ @Isa Just copy the list of closers to a TXT file and take a look at it for 3 seconds. The guilty one shall emerge. $\endgroup$ – Rodrigo de Azevedo Apr 27 at 17:35
  • 7
    $\begingroup$ BTW, here's some unsolicited advice for those of you who have enough reputation points. If your user ID is integer ABC and you want to search for the string XYZ amongst your deleted posts, then search for user:ABC "XYZ" deleted:1. $\endgroup$ – Rodrigo de Azevedo Apr 27 at 18:45
  • 15
    $\begingroup$ The deletion is indeed initiated by the same user, but that might not be the biggest issue as others users who helped on deletion are active members of CRUDE (see also: math.meta.stackexchange.com/questions/28390/is-crude-healthy). So I think posting here won't help at all. $\endgroup$ – カカロット Apr 27 at 19:46
  • 8
    $\begingroup$ Lets not forget that often old posts get invoked as precedence in current debates. $\endgroup$ – quid Apr 27 at 22:07
  • 8
    $\begingroup$ "I do not see any merit for closing such old questions: who the heck do we really expect to come back and "provide additional context" for a question that was asked two or even six years ago? " Amusingly it seems several of your example did get improved. One could say the system works. $\endgroup$ – quid Apr 27 at 22:12
  • 9
    $\begingroup$ Stack Exchange and the almighty, privileged users... I know enough of this drama. $\endgroup$ – Ezequiel Barbosa Apr 28 at 2:07
  • 31
    $\begingroup$ Closures are fine to me; deletions may be debatable. $\endgroup$ – YuiTo Cheng Apr 28 at 4:12
  • 21
    $\begingroup$ @YuiToCheng In my opinion deleting good answers is highly debatable, and I dislike the practice. The downside of allowing good answers to remain seems minimal compared with the downsides of deletion, which are: wasting the time people spent writing the good answers, preventing future readers from finding and learning from the good answers, and discouraging people from writing future good answers (for fear their work will again be deleted). $\endgroup$ – littleO Apr 29 at 0:31
  • 12
    $\begingroup$ @YuiToCheng your "possible opposite view" works for closing and deleting questions today, but not old questions. It is dangerous to judge the past using the values of today. $\endgroup$ – user1729 Apr 29 at 6:29
  • 8
    $\begingroup$ I reopened. The question tries to make some effort to focus on a specific aspect and thus it is maybe not a clear dupe. But more crucially, I think it is important to avoid giving the impression that certain points of view are being silenced. $\endgroup$ – quid Apr 29 at 8:47
  • 9
    $\begingroup$ "who the heck do we really expect to come back and "provide additional context" for a question that was asked two or even six years ago? "- if the answer is "nobody", then if the questions don't meet site standards I think it's very reasonable to close and delete them. $\endgroup$ – Carl Mummert Apr 29 at 19:25
  • 11
    $\begingroup$ The deeper issue, I think, is that some users continue to answer, reopen, and/or undelete new no-context questions. I think it would be good to develop a compromise in which old questions can be preserved but answers to new no-context questions are strongly discouraged (not just in words), In other words, if "old" is the key point of this post, how can we arrange to handle "new" questions differently? $\endgroup$ – Carl Mummert Apr 29 at 19:33
  • 6
    $\begingroup$ As I said in my answer, props for digging up good examples. I like to think most deleted threads deserve their fate, but this collections does point at a problem. $\endgroup$ – Jyrki Lahtonen Apr 30 at 12:12
  • 5
    $\begingroup$ They should only be deleted when the answers are of low quality and do not have many upvotes. Usually a good answer after comments to the OP to clarify a bad question can be provide excellent information for future readers that would simply be lost after deleting. In short no if and only if the the post receives quality upvoted answers. $\endgroup$ – onepound May 9 at 8:41
31
$\begingroup$

Thanks for saving me the time to post the exact same question. Let me just add my own examples to the list. My inclusion criterion is "contains a neat proof", which may be more liberal than yours, but I am pretty sure that you will find something worthy of preservation in the list (temp mirrors included if you are not a 10K+ user):

$\endgroup$
  • 10
    $\begingroup$ I think this list is tangential to the question. Of the three links of yours I clicked on, one was closed as a duplicate and the other was closed within days of being posted. This question is about questions which were closed and deleted years after they were asked. $\endgroup$ – user1729 Apr 27 at 18:23
  • 2
    $\begingroup$ @user1729: If it is tangential, then the point is rather singular. The dupe still had at least one valid answer that could not be merged because it doesn't generalize to the dupe target. I don't think "within days" vs. "within years" is such a great distinction; to me, what matters is the existence of good answers. $\endgroup$ – darij grinberg Apr 27 at 22:12
  • $\begingroup$ @MartinSleziak The link (the first one) is still pointing to the answer... Is it a mistake? $\endgroup$ – YuiTo Cheng Apr 28 at 2:51
  • 10
    $\begingroup$ I understand your point with the questions you list, but they are still tangential to the question. "within days" vs. "within years" is a great distinction. To quote the question: who the heck do we really expect to come back and "provide additional context" for a question that was asked two or even six years ago? $\endgroup$ – user1729 Apr 28 at 7:40
  • 3
    $\begingroup$ @darijgrinberg Could you please delete this answer? The question was closed for being a duplicate of this question. This answer and the close reason go hand-in-hand, but as I said about it is based on a miss-interpretation of this question. Possibly you could post your answer there? $\endgroup$ – user1729 Apr 29 at 9:03
  • 3
    $\begingroup$ @user1729: Not sure if this is worth the hassle. Most of the links here work fine in either thread, and the old one isn't particularly active. $\endgroup$ – darij grinberg Apr 30 at 1:11
14
$\begingroup$

As quid already remarked on the closure aspect of this meta question, let me address deletion (being aware that this might not make the answer popular):

The further strong action of deletion mostly, if not completely, erases useful information that people created (with valuable efforts and time) together in the main site.

The key word of dispute in the above sentence is "useful". Let's make some statements:

  • Every question is useful.
  • Only questions fitting the good question FAQ are useful.
  • Questions with good answers are useful.
  • Sufficiently old questions are useful.
  • Popular, highly voted questions are useful.
  • Questions consisting of only an image are useful.
  • An eager OP makes a question useful.
  • Only questions with pristine Markdown and $\LaTeX$ formatting are useful.

Everyone, depending on the assessment of these statements, has a different understanding about which questions are "useful".

Therefore it does not surprise that some users are of the opinion that questions they deem sufficiently lacking in "usefulness" should be deleted. The fact that this meta thread arises (and many before it) proves that there is a wide variety of interpretations of this word "useful".

And since neither side of the spectrum is going to convince the other, our job as a community is to find a proper balance on the spectrum. (Closing-not-deleting might be a part of this.) Just like we managed to create the good question FAQ.

As long as the self-righteousness of both sides continues to prevent a compromise, we are not going to get somewhere. Talking about "valuable efforts and time"...

$\endgroup$
  • 4
    $\begingroup$ So... has the time come? Should we attempt to draft a compromise? Should we attempt to draft an approach to drafting a compromise? $\endgroup$ – Lord_Farin Apr 29 at 20:05
  • 8
    $\begingroup$ I would like to see a compromise, if the moderators will be willing to stand behind one. Nothing is ever going to find unanimous support. We need a way for the site to make a decision which is binding even on those who might disagree with it. $\endgroup$ – Carl Mummert Apr 29 at 20:09
  • 3
    $\begingroup$ I've been reading not only the recent posts, but also some older ones to get a good idea of how it used to be here. There's been quite a few disagreements, sometimes even quite acrimonious ones, with a lot of it apparently due to not having a clear, generally agreed policy. In a community as relatively large & diverse as this one, you're never going to get complete acceptance of any document. However, drafting some sort of fairly detailed compromise, then possibly have a vote/poll on it, & enforcing it, is something I strongly agree with, even if I won't fully agree with what it might state. $\endgroup$ – John Omielan Apr 30 at 6:15
  • 4
    $\begingroup$ I think the issue of "usefulness" indeed deserves more attention as it's more subtle than the discussions sometimes give it credit for. A problem is that often posts are discussed in isolation and it is said something that amounts to "look, clearly this is or could be useful" and this would be often true if the post was the only content available. However, in a situation where some information is available multiple times removing lesser versions of it actually can be helpful to those searching for it as they are more likely to find a better version. $\endgroup$ – quid Apr 30 at 11:59
  • $\begingroup$ I think you have missed a way of addressing usefulness "Will this question answer combination lead a person to get the answer to a question they would be asking". $\endgroup$ – Q the Platypus May 1 at 4:38
  • 1
    $\begingroup$ @quid if there is redundant examples of the question isn't it better to mark as duplicates and move the answers? $\endgroup$ – Q the Platypus May 1 at 4:40
  • 2
    $\begingroup$ @QthePlatypus depends. For the same reason I gave a higher number of answers are not necessarily an improvement. As a matter of fact we have questions with near identical answers from the same user. Serves little purpose in my opinion. Besides, and often more relevant the redundancy is not one of exact duplication, and even in that cases merging is not always an opition. I agree that the point you raised in the first comment is important. The point is many of the so-called useful posts score very low on that if you ask me. $\endgroup$ – quid May 1 at 11:42
  • 5
    $\begingroup$ Beyond that and as explained earlier I prefer some things to be delete for the sake of educating users mostly those that answered questions that should not have been answered. There is a trade-off though between that goal and removing valuable content. Thus, in a case where things can be delete without valuable content being removed I might find it preferable to do that even if one could also preserve it, especially if the latter creates extra work. @QthePlatypus $\endgroup$ – quid May 1 at 11:46
  • 6
    $\begingroup$ Pedagogical research shows that negative feedback is a poor teaching tool. In my mind it weighs against removing content and towards improving (via editing) that content. In my mind deletion of content is a least best option and we should ask ourselves “What can we do in order to best use this content”. $\endgroup$ – Q the Platypus May 2 at 23:50
  • 2
    $\begingroup$ @QthePlatypus If negative feedback is a poor teaching tool, you seem to imply positive feedback is a good teaching tool. So what effect does it have to reward answers to poor questions with upvotes and other positive feedback? (Admittedly this applies more to recent bad questions.) $\endgroup$ – Lord_Farin May 3 at 6:31
  • 3
    $\begingroup$ @Lord_Farin this question is about old answers and questions so any reward has already had its effect. Deletion doesn’t undo it. When applied inappropriately deletion just removes useful content. $\endgroup$ – Q the Platypus May 3 at 7:38
  • 2
    $\begingroup$ @QthePlatypus I have to agree with you. We established that timing influences the desirability of deletion, in line with best practice to not retrofit today's norms on the past by means of deletion (for closure separate assessment is required). This should factor in any compromise to be reached. $\endgroup$ – Lord_Farin May 4 at 14:57
  • 3
    $\begingroup$ @QthePlatypus: You're misusing pedagogical research. Negative feedback is a poor teaching tool for the general population. So it does not apply to the minority of users who persistently regularly answer PSQs. Secondly, rep gained by the asker is taken away if the question is deleted, as well as rep gained by the accepted answer, and this has been the case since the beginning of SE, which implies that indeed the designers of the SE system thought the penalty should be there regardless of the age of the question. $\endgroup$ – user21820 May 5 at 8:39
  • 3
    $\begingroup$ I don't disagree with this answer, by the way, but if a small minority insists on restoring bad questions then obviously there is going to be a problem. And in many cases it is precisely the users who got their rep from PSQs who are the most insistent on this even if there is no contention that the question is a bad one. $\endgroup$ – user21820 May 5 at 8:47
  • 1
    $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$ – quid May 11 at 17:56
10
$\begingroup$

Attempting to outline my thinking. Not sure how mainstream my views are, so feel free to vote/comment as you see fit. As always, I may have overlooked something.

  1. I downvoted Jack's question because the title is too sweeping. If adopted as a policy, and taken literally, this would be turned into a tool prohibiting all community control on content quality of past posts, and I find that unacceptable. It is possible that I misinterpreted his goals. But given the recent history of his posts and the literal way some users want to apply "rules" like this I cannot support the general proposal.
  2. However, this time Jack, IMO, collected good examples that certainly make a strong case for criticizing the large scale deletions. With the exception of "How to prove this logic" thread, my first instinct would be to keep the question visible (not taking a position whether they should be closed, mainly discussing deletion here).
  3. The reason is that, in my opinion, the undesirable content mostly consists of A) homework, B) duplicates. These I will ever tolerate only grudgingly, and seek to close/delete at the slightest excuse. Yes, there is room for compromises (which is why the rules on PSQs/context were formulated).
  4. I acknowledge and support quid's "broken window" -argument. Therefore I won't argue in favor of reopening most of these threads. However, I don't whole-heartedly support deletions unless the question is an uninteresting piece of homework and/or covered many times on the site already. Actually, I would even welcome a new answer of the "$p-2$ and $p+2$" question that does away with the need to call upon Dirichlet's theorem. I may exercise my option to post that as a new question, but IMO such an answer would organically belong to that thread :-)
  5. A reason for me not to vote to delete an older post is that the deterrence effect is simply not there. The threat of deletion never stops users not interested in rep points from answering, it is only a stronger signal of disproval to them (which is the chief reason for deletions anyway). In those cases when a user is playing the rep game, it does serve as a deterrent also. But remember, deletion does not cancel the rep points, if the post earning them has stayed undeleted for 60 days. Because deletion is different from closure, in the case of an older question it should not (in my opinion) be the nearly automatic follow-up to a closure as is the case with low quality new questions.
  6. This (lack of need for deterrence) has also affected my voting when checking out review queues. I used to think that review queues mainly function as a way of collecting questions that need to be closed urgently. I can usually fill my daily quota of votes-to-close by reviewing new questions alone. Therefore I felt that I would get more bang for my votes by skipping old questions while reviewing. I would only come back to them, if I find the time (using CRUDE makes that easy). But, closing an old question is never urgent. Nevertheless, I am re-evaluating my position here while trying to absorb quid's recently posted guideline.

Don't know if I reached any kind of a conclusion? May be just another call to judge things case-by-case?

$\endgroup$
  • 1
    $\begingroup$ I also don't know how mainstream my views are, but what you wrote seems very reasonable to me. Although each case needs to be judged on its own merits, or lack thereof, having a relatively detailed set of guidelines should, I trust, reduce the amount of conflict & disagreement. $\endgroup$ – John Omielan Apr 30 at 14:04
  • 1
    $\begingroup$ FWIW I followed up on that whim here. $\endgroup$ – Jyrki Lahtonen May 5 at 13:23
8
$\begingroup$

To almost equate closing and deleting for old questions is not a good idea for this discussion.

Regarding closing the following aspects are important to keep in mind, in my opinion.

  • New users might often get to know the site via stumbling over existing (old) content when searching the internet for information. If there is no indication on the old content that currently it could face push-back, it is not unreasonable that they simply mimic they style of the content they saw when asking a new question. Even if the old questions is never improved, the question being closed might guide new users behavior.

  • The meta-question here asks:

    I do not see any merit for closing such old questions: who the heck do we really expect to come back and "provide additional context" for a question that was asked two or even six years ago?

    Let's turn it around. What's the problem with the question being closed? It only prevents new answers. Now, let us be realistic, how often does it happen that a new answer appears much later that adds something substantive (especially on the more routine type of questions that are very common on the site)? Even if it is given, how much attention does it get?

Granted, for deletion the balance might be different.

$\endgroup$
  • 21
    $\begingroup$ -1: It is reasonable to "almost equation closing and deleting", because that's precisely what the CRUDE people do. They close to delete, not to prevent new answers. Further, your first bullet sounds like a corner case; and your second one, at the very least, goes against the design of the site, where we have badges like "Reversal", "Revival", and "Necromancer". $\endgroup$ – Martin Argerami Apr 28 at 14:52
  • 5
    $\begingroup$ Given that 90 percent of the sites traffic comes through search engines, I'd contest it's that much of a coner cases. But of course it's hard to know for sure. What we can and do know is that not that rarely precedence does get invoked. "But this other question etc." Thus, I stand by it there being value in current standard being also (at least in part) being applied to old content. (Of course somebody might disagree with current standards but that's a separate concern then.) $\endgroup$ – quid Apr 28 at 19:26
  • 3
    $\begingroup$ It's interesting to note that that you seem convinced that the old content does have a lot of impact as otherwise its deletion would not be an issue. I feel that to have it both ways is not possible. Obviously the impact of having old stuff closed will be limited, but I find it hard to believe that the influence is as negligible as you claim (especially relative to other effects we are discussing). $\endgroup$ – quid Apr 28 at 19:28
  • 4
    $\begingroup$ On the other point, of course the site also appreciates contributions to old threads. Yet, two things. First this would seem to suggest that we should strive to maintain the old threads and not be as dismissive as OP in mocking the idea that somebody might improve an old questions. (Since you seem to like badges there is archelogist for instance). Second, the discussion here is specifically about old threads that already have good answers. Sure somebody might add another good answer, but honestly that seems like a corner case to me and in any case not an ideal allocation of resources.. $\endgroup$ – quid Apr 28 at 19:33
  • 11
    $\begingroup$ I fail to see your point. This thread is motivated by the CRUDE people deleting old questions that have good answers. They don't stop at closing, so I don't see where you are going with this. $\endgroup$ – Martin Argerami Apr 28 at 21:31
  • 4
    $\begingroup$ @MartinArgerami well, if we want to find some sort of compromise it would seem relevant to know if for example for you the problem would be resolved if they did stop at closing. I mean to simply post STOP, which is basically what OP does, won't get you very far. The thread has some relevance as a document of a feeling of exasperation, but it's not really a contribution to a discussion that has the potential to lead to anything constructive; $\endgroup$ – quid Apr 28 at 21:50
  • 6
    $\begingroup$ Further, and maybe more to the point, OP mocked the activity of closing. Therefore I explained why I think it has some relevance and not all that much drawback (contrary to deletion where sometimes there is a significant drawback). This resulted in you contesting the arguments with not so great arguments, and dodging any further discussion with some non-argument. I'll also note that earlier explanations on why deletions do have some merit went basically unacknowledge. One cannot help put feel that there is no willingness at all to actually discuss anything. It's just thinly veiled ranting. $\endgroup$ – quid Apr 28 at 21:58
  • 20
    $\begingroup$ Of course it's a rant. What kind of compromise is there, and by whom? The CRUDE people vote to close 20 times a day, every day, and nothing happens. I vote to keep open 20 times, once, and I get admonished and banned from reviewing. These people are destroying content, and the authors of said content are not even notified; how can you feel that's fair, at all? And I don't know what you think I ignored about deletions; your answer said that this is not about deletions, and I disagreed. $\endgroup$ – Martin Argerami Apr 28 at 22:47
  • 8
    $\begingroup$ I have fundamental difficulty to understand your comment. I said "if we want to find some sort of compromise it would seem relevant[...]" You completely ignore to the part I said I would find relevant and proceed to rant that there is no compromise. Well, yes, there is none. And it seems that you are not interested in finding one. Since you did not answer a simple question related to this, and instead continued to complain. Where did I say that it is not about deletions? I said I think it is relevant to distinguish the two Then I discussed closing and made a remark on deletion. $\endgroup$ – quid Apr 29 at 8:54
  • 3
    $\begingroup$ @Martin Argerami I agree there is a delete downvote fetish on here $\endgroup$ – onepound May 11 at 10:05
  • 3
    $\begingroup$ @onepound relative to some other SE sites few questions are closed on this site (even just restricted to new ones, yet of course in a relative sense). $\endgroup$ – quid May 11 at 11:31
  • 2
    $\begingroup$ @quid: Please give explicit official statistics supporting your claim. $\endgroup$ – user 170039 May 11 at 14:11
  • 6
    $\begingroup$ @user170039 it's not clear why you can't look this up yourself but there you go: meta.stackexchange.com/questions/321318/2018-a-year-in-closing For 2018, among the about 170 SE this one does barely make the Top 100 in closing percentage, 98th I think. Among the about 30 sites with more than 10k question during that year, it is about 20th and among the 18 sites under science it is 16th. Relative to other sites, we certainly do not close a lot rather the contrary. $\endgroup$ – quid May 11 at 16:05
  • 4
    $\begingroup$ @quid But compared to the only other site with comparable volume (StackOverflow) this site closes at about a 44% higher rate, viz. 16.6% vs. 11.5%. All the other sites are far too small volume to compare too. $\endgroup$ – Bill Dubuque May 14 at 21:46
  • 3
    $\begingroup$ @BillDubuque what is your criterion for volumes being "comparable"? I ask as the ration of questions SO to math.se is significantly above 10 (that is quite far away), while the ration of math.se with several other sites is significantly below 10. Still far away but closer. In absolute numbers SO has 2889k Q asked we have 216k Q asked then AU has 62k, SU has 53k, and there are several above 20k which is closer to us than we are to SO. $\endgroup$ – quid May 14 at 22:25

You must log in to answer this question.

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