44
$\begingroup$

It does feel to me that the moderators here go too far when it comes to deleting questions here. These are often questions that were quite interesting in their own right and received some good answers with lots of upvotes,....,but the person asking did not do a thorough enough job documenting the context of the question and so technically the stackexchange guidelines were not met [and frankly, EVEN THAT is debatable]. But why go so far to delete a question that has received a suite of good answers.

I have no problem closing a question that is poorly written or is clearly a simple exercise copied out of a textbook no context. The 'Do My Homework!' questions are the worst, and so are the imprecisely written posts. But deleting a question that appeared to be well-received on here and that sparked a suite of good answers is just a hole nother kettle of fish I think.

In any event, deleting a question with the accompanying upvoted answers feels extremely heavy-handed to me on the part of the moderators. They are deleting content that the person writing put effort into, and that others on the site appeared to have found valuable. I cannot imagine that most people on this site would be in favor of this.

Please advise.


ETA: Well, I did want to respond to this thread that I wrote. As the comments have gotten a bit long and have seen to go off-topic, I figured I respond here.

First note that, as of 13:01 EDT Pi Day 2022 my original post got 36 upvotes and 10 downvotes, and the one answer favoring deletion got 14 downvotes and only 4 upvotes. So many many people here seem to agree with me. They do NOT seem to agree with those who want to delete questions with well-received answers. Moderators, are you paying attention to this. Maybe the way you want to enforce the EoQS isn't the way to go about doing so.

Something to think about.

$\endgroup$
54
  • 16
    $\begingroup$ A point of clarification: most post deletions are accomplished by regular users (i.e., not diamond moderators) with "vote to delete" privileges and can be reversed by other users with the same privileges. There's even a dedicated thread for reopen and undeletion requests, so there is some method for undoing this. $\endgroup$
    – KReiser
    Mar 8 at 3:42
  • 10
    $\begingroup$ Why delete them? Because it has become clear that closure is not enough, when the users who keep answering them get rewarded regardless and ignore the community's established preference for all content to be good and not just serve as a seed for other good content, deletion takes away the encouragement to keep answering them, which takes away the encouragement to keep asking. This discussion has been held multiple times, several on the recent past, and it appears the explanations are never accepted, so why bother with yet another version? $\endgroup$
    – Nij
    Mar 8 at 3:57
  • 16
    $\begingroup$ Come up with specific examples of deleted questions with good answers, Mike, and I'll vote to undelete them. $\endgroup$ Mar 8 at 5:46
  • 9
    $\begingroup$ @Nij Previous meta posts like this one: math.meta.stackexchange.com/questions/30163/… suggest that the consensus on deleting question with good answers is not as clear-cut as some like to present it. $\endgroup$
    – MaoWao
    Mar 8 at 9:36
  • 12
    $\begingroup$ These deletions mostly are by users, not by moderators. But I agree with the sentiment: closed questions are much too often deleted. Often, I agree to the closure, but not the deletion. So I disagree with the comment of Nij, but I also disagree with the comment of GerryMeyerson. Many questions should be closed but left that way: not re-opened and not deleted. I have an answer (in another forum) to a closed question which still occasionally receives upvotes and even badges. $\endgroup$
    – GEdgar
    Mar 8 at 14:09
  • 14
    $\begingroup$ The moral of the story is that users should not put their time and energy into answering low quality questions. It would be better to put your time and energy into either (a) answering high-quality questions or (b) improving low-quality questions until they become high-quality questions. If you choose to answer low-quality questions, then the deletion of your work is a likely outcome. Good answers don't save bad questions. $\endgroup$
    – Xander Henderson Mod
    Mar 8 at 20:05
  • 9
    $\begingroup$ "Many questions should be closed but left that way: not re-opened and not deleted." @GEdgar, that would be good, but the problem is that closure is very often just the prelude to deletion. $\endgroup$ Mar 8 at 22:54
  • 24
    $\begingroup$ "It would be better to put your time and energy into either (a) answering high-quality questions or (b) improving low-quality questions until they become high-quality questions." But when anyone other than the user posting the low-quality question tries to improve it by editing to add context, etc., the edit gets rolled back on the grounds that edits by others shouldn't go beyond the intentions of the original poster. $\endgroup$ Mar 8 at 22:59
  • 17
    $\begingroup$ "...we need to help askers in improving their question. Involvement of asker is important." @Para, why? If the goal is to make the site a repository of good mathematical questions and answers, why does it matter whether the asker is involved in improving a question? $\endgroup$ Mar 9 at 1:50
  • 15
    $\begingroup$ "how would you feel if someone edited one of your posts substantially?" Straw man, @Para. We're talking about questions that get closed (and then deleted) because OP doesn't edit them up to our standards. I suspect that those posters would be only too happy to see their questions get edited to the point where they can be answered without risk of deletion. But if you can find a question I've posted that you can improve vastly, then, by all means, go for it. $\endgroup$ Mar 9 at 5:35
  • 16
    $\begingroup$ @ParamanandSingh If people getting offended when their posts are altered is a valid argument against editing posts, then surely we shouldn't close and delete questions at all anymore. People get upset and complain about it all the time. $\endgroup$
    – MaoWao
    Mar 9 at 14:03
  • 10
    $\begingroup$ @XanderHenderson: Can you give a link to "the last time that this came up" that you refer to? I have never seen any discussion anywhere on meta that indicates a consensus that a question should not be rewritten to include any context which was not provided by the author. The guidelines that you link to appear to directly contradict this--they say explicitly that certain kinds of context are OK to add, and other kinds are not, without reference to whether that context was provided by the author. $\endgroup$ Mar 9 at 17:02
  • 11
    $\begingroup$ @XanderHenderson: Sure, the votes are ambiguous, but it certainly seems as though the assertion that such context edits were appropriate (perhaps when accompanied by polite comments) was largely uncontroversial at the time (for instance, there are no comments taking issue with that assertion). You are the one that is claiming there was a consensus in the opposite direction, and there is little evidence for that consensus. $\endgroup$ Mar 9 at 17:28
  • 15
    $\begingroup$ @Para, if the "policy" is in conflict with the overarching aim of having a repository of good mathematical exposition, then policy shmolicy. $\endgroup$ Mar 10 at 3:53
  • 11
    $\begingroup$ @XanderHenderson I believe that deleting a ill-answered question with good answers is going against the spirit of teaching people: as someone that is not a mathematician, I have learned a lot with questions here, and sometimes because of my lack of background, questions are not fulfilling all the requirements: simple because are "chasing something" by intuition, ignorance related to the topic, translation issues, and its a "catch 22" situation to over-elaborate because long questions are less probable to be read. I have bookmarked questions as reference, that are 3 lines short (see [big-list]) $\endgroup$
    – Joako
    Mar 14 at 2:09

1 Answer 1

-9
$\begingroup$

My thoughts on the question

The only reason I can see for an existing answer to affect a question's MSE on-topic nature with respect to the guidelines is that it actively refutes a particular close reason. For example, an answer citing facts and studies works against a question subsequently closed as opinion-based, and an answer that uses an overarching concept or a universal trick to tackle two or three apparently separate questions in one fell swoop works against the argument that the question is broad.

All votes do is attract my attention to some answers and make me analyse, more carefully, whether they refute the applied close reason. So votes matter, but more importantly the nature of the answer as compared with the applied close reason matters.

Closure does enough to perhaps redirect people elsewhere and send the need-for-improvement message. Deletion, on the other hand, is a measure that removes questions which haven't been active for a long time (via the community user) and nullifies the visibility of a closed question, preventing it from receiving any attention from the majority of the community and leading to loss of reputation, especially if it's a recently created post.

Mere closure doesn't seem to have been enough to deter the answering of low-quality questions. Deletion, a stronger measure, has a better effect in curbing this habit, per empirical observation, because of reputation loss and the accumulation of such posts leading to moderator action ref. EoQS. Deletion is therefore a deterrent to the answering of low-quality questions.


It is clear that I placed an application of the guidelines over the well-received nature of the post and the efforts of all participants on that thread. So why did I do that?


Write a new question?

The first is that I am not sure why people are not using the best method to rescue content : a new question. The insistence of the methods of improvement-prompting and context-rewrites are historically unsuccessful, the former because of the asker being a participant and possibly not having enough interest, and the latter because of the balance between "words in asker's mouth" and "too little context" being very fine and open to interpretation as I see it. These methods should only be used if one is very patient or very confident.

The only problem I see in asking a new question is that the question could already be off-topic for another reason (which would likely be the case for the earlier question as well) or the original question could have more than two answers , which are difficult to copy across to a new thread. Given that most of the answered deleted questions I observed had precisely one answer, this captures the majority of such questions. Furthermore, such a process of rescue scales in numbers, as I explain below.

There were roughly ~$400$ questions asked yesterday. Of these, about $85$ were closed, giving a rate of roughly $20$%, say. The deletion rate is slightly lower, around $19$%. Of a randomly chosen list of $30$ deleted questions (note : continuously chosen from a list of questions deleted about six hours ago) , $23$ were unanswered, $7$ were answered with $5$ having single answers (of which all had vote either $0$ or $-1$) and one had two answers, each with a $-1$ and the last one was actually well-received with $2$ answers, one with a $+2$. Most were community-user deleted. When scaled, there are likely four questions in a day which are deleted with an upvoted answer : easy enough for individual-question rescue. Finding them is perhaps a bit more difficult, but the daily number is quite small.

Part of the reason I favour deletion is because I want to encourage people to rescue content from these questions by writing their own questions with context added, thereby raising the question's standard and allowing the rescued answer to fit more hand-in-glove with its rewritten question. I would like participants and viewers to redirect their efforts to this form of rescue : generally uncontroversial and successful.


On votes

The other has to do with what I think of votes. What exactly does one vote for? I think one should base their votes on how much hand-in-glove the answer is with the entirety of the question.

For example, let's say there's an answer that refers heavily to the user's context but is somewhat long and labouring, while the other is a slick one-line answer but doesn't comment on why the OP's effort is misdirected or not enough. Which would you vote up? Perhaps both : but I would definitely vote up the answer that appreciated the context provided by the user and commented on it in their answer.

Context is, after all, supposed to help someone address the question to the best of their ability, so to see someone not so much as go near it but get $10$ upvotes, while the other one gets just $2$? It's natural because we gravitate towards beautiful and slick math, but it isn't appropriate for MSE if we are choosing to have people add context. Which sounds weird, but true.

If votes were directed at answers that worked the hardest at harnessing the user's context and attempted to nullify any potential reasons to close the question, then we'd have an immediate reason to see why well-voted answers can easily deter deletion and closure. For example, if a question is broad, an idea that somehow brings the entire question under one umbrella should get credit over three distinct tricks. For questions which may be opinion based, you would vote for the answer that has enough facts and citations, and not for the post that you relate with.

Therefore I prefer deletion over preservation of highly-voted answers : because the way people seem to vote is natural, but often doesn't favour and promote answers that actively work to use tools from the question and work on the potential site-based pitfalls of the question. That means I'm in a minority, but I can see how well this notion of voting yields a more positive enforcement of the guidelines, and would therefore defend it all day and night.

$\endgroup$
7
  • 4
    $\begingroup$ Similar to how I had said in the chat previously, you assume that people are voting intelligently for questions most aligned with site policy. However time and time again, it has been shown that is not the case. Refer. There are more recent examples but those ones are removed swiftly $\endgroup$ Mar 9 at 21:11
  • 24
    $\begingroup$ Rescuing deleted questions is great in theory, but in practice it's hard to do. To rescue a deleted question, you have to know it's deleted. No one notifies us every time a question gets deleted. I'm looking at the "recent deletions" listing on my "tools" page, and I see 45 questions deleted in the last 8 hours (40 of them in the last hour). There's no easy way to see questions deleted more than eight hours ago, and no easy way to sift through them even if I did see them, to see which ones were worth rescuing. There may be dozens of potentially-good questions gone that I'll never see. $\endgroup$ Mar 10 at 4:04
  • 1
    $\begingroup$ @GerryMyerson You're right, the scale argument is only great in theory and not in practice. Some SEDE scripts can help retrieve IDs of deleted questions, yet you'll have to visit these questions(and be able to view them) to figure out which ones are worth rescuing. Having said that, there are some questions above which we could try to rescue : no one has done it yet. I forfeit some parts of my argument, but continue to believe that there should be an organized round-the-clock effort to track these questions and restore them, similar to the organized effort of removing low-quality content. $\endgroup$ Mar 16 at 8:33
  • 4
    $\begingroup$ @GerryMyerson Since you've mentioned looking for deleted questions, I'll add a link to this: Are there some tools to follow deletions on this site? (I certainly agree that there is no easy way to find the deleted questions, but SEDE gives at least some information about the deleted posts. Perhaps the only exception is that every user can see their own deleted posts - those ones are easily accessible.) $\endgroup$ Mar 17 at 4:25
  • $\begingroup$ @MartinSleziak Thank you for the very informative links. $\endgroup$ Mar 17 at 4:44
  • $\begingroup$ Actually, is it easy to find how many have been rewritten? I have seen personally 2, but its not like we tag them. I tried searching for links to the guidelines and found nothing. I guess the mods would know? $\endgroup$ Mar 17 at 12:33
  • 1
    $\begingroup$ @CalvinKhor I do not know of a method of measurement, and unfortunately I'm not very optimistic the mods have a method either. It's good to know you've seen 2, I myself handheld one person across a rewrite in a meta post maybe a year back and felt good about it. To me, even if the guidelines on context rewrites are quite clear, what seems to be plaguing it is this notion of what comes between "enough to improve the question" and "doesn't change the author's intentions". Sometimes, it feels like adding definitions takes away the lack-of-clarity in the OP's definition, etc. $\endgroup$ Mar 17 at 15:22

You must log in to answer this question.

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