7
$\begingroup$

The moderators have been cleaning up and deleting closed questions. While in principle, this is good to keep the site clean, it has the huge negative effect of reducing our hard earned reputation. I answered the questions in good faith, but now I've lost more than 100 reputation points. I know that in CSTheory.SE, we only ever deleted spam, but almost never deleted questions.

In my opinion, closed questions (with up-voted answers) should not be deleted. It is not a bad idea to keep closed questions around, for the following reasons:

  • Closed questions quickly fall down the list and do not appear on the front page.
  • Closed questions provide a record of what questions are not on topic.
  • Closed questions can also be referred to if the same question appears again, thereby avoiding any debate or disgruntled feelings.

Can someone point me to where in the SE general policy or in the CS.SE policy it states that closed questions will be deleted? Should (non-spam) closed questions be deleted by moderators?

After doing a little investigation, I found the following on this page https://cs.stackexchange.com/privileges/moderator-tools:

Closed questions that are of no lasting value whatsoever should be flagged and deleted.

One of the deleted questions was Why does Knuth persist with MIX/MMIX?. This is the kind of question that could reappear. It's a natural question, even though it was judged to be off topic. Leaving the question around would be useful for other people who wish to ask a similar question, or even those who ask questions about what is in the mind of some author when they wrote some paper or book.

$\endgroup$
  • 1
    $\begingroup$ I agree. @Gilles, why did you not discuss this on Meta first? If retagging is to be discussed, deleting questions should be, too. $\endgroup$ – Raphael May 6 '12 at 8:38
  • $\begingroup$ Concur. I had a mind to delete some questions about a week ago, but decided to leave them alone. If they're not clearly worthless, I'd feel better about the community eventually deleting them. $\endgroup$ – Patrick87 May 6 '12 at 11:35
  • $\begingroup$ Hopefully, there are not many closed questions with upvoted answers. So you are talking about some exceptional cases. $\endgroup$ – Tsuyoshi Ito May 6 '12 at 19:18
  • 2
    $\begingroup$ Actually there were three or four questions deleted overnight that had up-voted answers. I only noticed them because of the negative reputation points I earned. $\endgroup$ – Dave Clarke May 6 '12 at 19:20
4
$\begingroup$

Let me quote a very good explanation of the rationale for deleting closed questions:

If a question is valuable enough that it should stay on the site, it should be reopened or fixed into a state that permits reopening. Otherwise, excepting duplicates, it should eventually be deleted if no one expects that it'll be reopened.

If you take the "head on a pike" approach of showing "This is what happens to people who ask what we don't want", it looks rather unsightly. If the problem is because the question is incomplete or broad or something else that can be fixed, keeping it around when someone does not fix it just means you keep what is essentially a mockery of the asker. If the problem is because the question falls outside the scope of the site, then keeping it around invites the problem of users discovering the site because they found that subject matter on the site. The majority of traffic to sites come from searches, so keeping things you do not want people to ask out of sight through deletion helps prevent more people from coming and asking the same thing. "Head on a pike" only really serves as a gruesome reactionary measure than a prevention of the incident.

As such, I recommend that if a question can't be fixed or does not wish to be fixed, that it probably serves the community better if it is deleted. You can give equal guidance to those who run afoul of the rules by pointing at established guidelines - things you write out on the above pages as well as in Meta discussions.

There is a general Stack Exchange policy, expressed in two of Jeff Atwood's blog posts:

But a closed question is no longer alive in any meaningful way, and certainly well on its way to the bit-bucket of /dev/null.

Let me be clear: we do not seek out deletion, by any means. But we believe not having the guts to cull some of your worst content is much, much more dangerous to your community than letting it sit around forever in the vague hope that it will magically get better over time.


Why would you delete a question? Isn’t closing it enough?
(…)

  • If users see a lot of closed questions, they’ll note that we don’t enforce the guidelines, so why should they? Without any final resolution, asking questions that get closed becomes something we are implicitly encouraging — a broken windows problem. If this goes on for long enough, we’re no longer a community of programmers who ask and answer programming questions, we’re a community of random people discussing.. whatever. That’s toxic.
  • If enough of these closed questions are allowed to hang around, they become clutter that reduces the overall signal to noise ratio — which further reduces confidence in the system.

The point of deleting questions is in fine to make the Internet a better place. Closed questions (throughout this post, this does not include duplicates) are questions that have been considered unwelcome here. Off-topic questions are questions we do not have the expertise to answer; our answers have not been reviewed by competent people, so their validity is put into question. Non-constructive or “not real” questions are questions that cannot reasonably be answered; answers to such questions are usually partial and often miselading. Too localized questions do not add value due to highly limited interest.


Responding to your arguments:

  • Reputation points are not a consideration when taking moderation actions (whether acting as a user with reputation-granted privileges, or as a diamond moderator). The purpose of moderation is to clean up the site.
  • Closed questions are still found in searches. They have exactly the same visibility as open questions.
  • I have never seen it useful to have a record of rejected questions.
  • When a borderline question is asked, it is best if there is no closed, non-deleted similar question already on the site. This lets the new question be evaluated on its own merits, rather than being summarily lumped with the old question and pressured into sharing its fate.

I do not see why Computer Science should have a different policy from other sites.

Closed questions are not deleted while there is still debate as to whether they should be reopened. After two weeks, I did not see any hint of a debate as to whether Why does Knuth persist with MIX/MMIX? might be reopened. If you think this question can be improved and reopened, answer here or flag it and a moderator will undelete it.

$\endgroup$
  • 1
    $\begingroup$ I think answers to closed question can contain valuable information, there is a reason why such questions are not deleted automatically nor can be deleted by the OP alone. If I rephrase Dave's point in my own language, I would say that it is not nice to people who spend time and effort to write answers to delete their posts just because the question is closed, reputation is not the real issue. If I spend an hour to write an answer and it gets deleted because the question get closed later although my answer is up-voted I would be less eager to answer questions. $\endgroup$ – Kaveh May 6 '12 at 22:04
  • 2
    $\begingroup$ I also feel two weeks is too short, it might make sense on a site of the scale like Stack Overflow but on smaller sites I think it is different. Having policies similar to sites like Mathematics is more reasonable IMHO, at least for the moment. Of course deleting closed questions which are clearly off-topic,... and have negative total vote is more acceptable. $\endgroup$ – Kaveh May 6 '12 at 22:09
  • $\begingroup$ @Kaveh Of course answers to closed questions contain information in the information theory sense. The problem is that this information is not reliable (off-topic → no competent peer review; NC/NaRQ → answers are biased or incomplete). Is two weeks with no activity (no edit, no comment, no issue raised on meta) really too short? Do you really think some of these questions might yet have been edited? $\endgroup$ – Gilles May 6 '12 at 22:17
  • 1
    $\begingroup$ It is more than information theoretic sense, the OP might not be able to phrase the question in the right form so it is answerable but if there is real question behind it then it is not really a worthless question. The answers might address the issue behind the post and I think the effort people put in answering question is valuable, particularly when the question and answers have several up-votes. $\endgroup$ – Kaveh May 6 '12 at 22:21
  • $\begingroup$ I sincerely think two weeks might be too short, it is not usual for new users and non-regular to drop by the site infrequently, I have seen a few questions that were closed and the OP improved it several months later and it was reopened (though it happens really seldom). Of course the OP can repost a new question but I think you can guess the feeling of OP when the question is deleted. My personal opinion is that unless a question is harmful or we start to have an issue (e.g. too many closed questions) we should be relaxed about deleting closed questions which have up-voted answers. $\endgroup$ – Kaveh May 6 '12 at 22:25
  • 1
    $\begingroup$ @Gilles: Just because the question did not result in any additional activity after it was closed is not a reason for deleting it. If the question serves no purpose at all or is toxic for the site, then it should be deleted. For example, this question from stackoverflow has been very useful as a reference for sites not wanting certain big list questions: stackoverflow.com/questions/194812/…. Closed questions do have value. $\endgroup$ – Dave Clarke May 7 '12 at 6:37
  • 2
    $\begingroup$ Also, I disagree that the purpose of moderation is to clean up the site. The purpose is to ensure that the site runs smoothly. Taking a less aggressive, more hands-off approach, only intervening when necessary is, in my opinion, a better approach. It's less likely to annoy people. $\endgroup$ – Dave Clarke May 7 '12 at 6:41

You must log in to answer this question.

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