31

I'm calling it a Closehammer rather than an Onholdhammer because... well, just look at how odd Onholdhammer looks.

I'm proposing that trusted users be given the ability to instantly place a question On Hold unilaterally.


Now I'd like you to read the following and reconsider undoing your downvote...

A trusted user would be defined strictly, the emphasis of this trust being based not solely on the user's reputation but on their moderation activity. The following attributes could be used in this measure of trustworthiness:

  • A good track history of successful close and delete votes
  • A good track history of moderator flags
    • The user actively uses flags
    • A high percentage of these flags are marked helpful
  • User has a history of being active on the website
  • User has a history of being active on the website's meta

In addition, I would suggest that in order to use this power, the trusted user would have to add a comment (wait, let me explain before you go recast that downvote!) to explain why this extraordinary action was taken. This also will allow the user to cast a non-binding close vote--If you don't leave a reason, it is a regular vote. If you do, instaclose.

I can already hear the choruses of "asdfasdfasdfasdfsadfasdf" coming... But understand that these users are the least likely to pull this kind of BS move. And if you do, that's great! Because we can ban you from using the closehammer, as obviously you don't care about what you're doing.

Rationale for this proposal:

Let's not waste lots of time and effort closing stuff that is obviously unsalvageable.

Why this is not a duplicate:

Note the proposed duplicate is already . This proposal also doesn't consider reputation as an indication of trust, is about instant closure not increased weight... it's not a match at all.


After some consideration, I think it should be named the Haltenhammer.

Halt! Hammerzeit.

  • 1
    @psubsee2003 clearly on_hold_hammer is the right answer. – user213963 May 18 '15 at 13:16
  • 2
    +1 for idea, +1 for justification, -1 for my negative bias in the idea :D – Unihedron May 18 '15 at 13:17
  • 2
    Please don't require 10k for it kthxbye – bjb568 May 18 '15 at 13:19
  • 7
    +1 because "not solely on the user's reputation". Why not if they're deemed to be making good decisions. Things can always be amended or rolled back if they make the odd bad decision etc. And their rights revoked if too many bad decisions – James May 18 '15 at 13:19
  • 3
    Why do you believe there is a need for this? No rationale has been given in this proposal for why this change would be necessary and what problem it would solve. If you are unhappy with the speed at which a question you wanted closed was closed, and believe it urgently needs to be closed, you could raise a mod flag on it. – Jason C May 18 '15 at 13:50
  • 3
    Besides, don't we already do this every year or so in an election? – Jason C May 18 '15 at 13:54
  • 1
    Seeing as there are a lot of things that escape unnoticed on some sites with very high volume (Stack Overflow, for instance), I would be concerned with abuse not being discovered immediately. And if someone does begin to abuse it, it could take a while to be seen and to be cleaned up – psubsee2003 May 18 '15 at 13:55
  • 1
    @Undo but someone "abusing it" may not opt for speed. They might just have a misguided opinion of what makes a good question. – psubsee2003 May 18 '15 at 13:59
  • 2
    Yeah, what if @bjb568 gets it? – Jason C May 18 '15 at 14:01
  • 14
    @JasonC The justification is obvious on its face. Not only for the utter crap, off topic, unsalvageable questions that pour in on a regular basis, but also for the hundreds of thousands that have made it past the front page and wait, festering, for years to be closed. Why waste time with these? Shut them down immediately and move on. Saving everybody's time is never a bad thing. – Won't May 18 '15 at 14:14
  • 4
    @JasonC I get it, you don't think it's worth the bother. I can understand that viewpoint. I take the other view--let's get rid of crap as fast as possible with the least amount of effort. – Won't May 18 '15 at 14:29
  • 2
    @Won't I can understand that. There's a second more subtle issue, too: There is no rationale given. While a problem may seem "obvious" to you, this is a particularly sensitive topic in that generally calls for anything to increase power are met with blind agreement as most people want more power (not necessarily in a greedy or devious way). You should give a clear analysis that shows that the questions are a problem and that the suggested inability to currently handle them well is causing an issue, otherwise it has more of a bandwagony kind of vibe to it, if that makes sense. – Jason C May 18 '15 at 14:36
  • 2
    What's really being asked, @Won't , is for you to verify that this is an objectively bad problem that everyone is experiencing. Sure, we see that you are obviously perturbed by it, but what we don't see is that everyone is, and further that this would resolve the problem. It's not a bad idea in and of itself, but it has no foundation on which all can agree, and thus there's little to base agreement or disagreement on other than individual subjective perceptions. Bringing hard data to the table is difficult, but until someone does, all we can do is wave our hands and choose the bikeshed color. – Adam Davis May 18 '15 at 14:47
  • 4
    Can the commentators please post answers for additional suggestions or new view points? Comments really aren't well fit for extended discussions which extend this long. – Unihedron May 18 '15 at 14:55
  • 2
    @JasonC You need to bug the folks who use reputation as a measure of ability to moderate successfully. Moderation requires understanding the goals and standards of the website, not how to use tail recursion to solve euclid's sudoku or sommet. – Won't May 21 '15 at 13:37
19

Even senior/trusted users can disagree, so this will need some additional safeguards to avoid close/reopen loops.

On the "what privileges should we give to 30k users?" question I proposed expedited closures with the following limitations:

  • Instantly put a question on hold for any reason other than duplicate, provided that:

    • The question has not been previously closed and reopened.
    • The question, if in the review queue, does not have pending "leave open" votes.
    • (Maybe - open for discussion): the question is on the front page or, failing the ability to implement that, "new" (definition TBD).
  • Instantly reopen a question that was put on hold by a 30k user, provided that:

    • The question has been edited since it was put on hold.

This would exclude duplicates (already covered) and migrations (more disruptive, so should involve a mod or community vote).

That was in the context of high-rep users, but I can see your argument that an actual track record of good moderation is a good basis. We all know high-rep users who rarely vote, flag, or review, and lower-rep users who do a lot here. I think the following (in some combination) would be reasonable factors:

  • Enough reputation to cast close votes. (I can see this being a 10k privilege, but even if it's not you still need to be able to cast, and have a record of casting, close votes.)

  • Has voted to close some minimum number of questions (100?) and maintains some level of ongoing activity (some number of close reviews in the last month, for example).

  • Has raised helpful flags on at least 50 (?) posts (comments excluded) and no more than 10% (?) of resolved flags in the last month were declined. (More than 10% declined in 7 days is where you start to get warnings; I kept the percentage and lengthened the period so privs wouldn't oscillate as much.)

  • Convention badge: meta is important.

The specific criteria can and should be tweaked, but something in that general ballpark is a good baseline for "we trust this user to do this".

I didn't include a measure of "successful" close votes because that's hard. That a question got reopened doesn't mean it shouldn't have been closed; maybe it needed and got an edit, or there was a scope discussion on meta resulting in a change. I'd like to capture some measure of "doesn't screw up closures too much", but I don't know what to do there.

  • 1
    The tone of this comment is neutral. Do you believe there is a "slippery slope" (either in a bad or a good way) component to this type of criteria? This would become the only site privilege that isn't based on a simple rep threshold or single badge. It is a significant outlier, and opens the door to a similar discussion wrt other privileges by setting a precedent. I'm positive such discussions have taken place before but I can't find them, if I do I'll post links. – Jason C May 18 '15 at 16:00
  • 3
    I don't believe badges are a good matrix, out of all the options there are, especially that they aren't retracted. – Unihedron May 18 '15 at 16:06
  • 1
    @JasonC I see the "slippery slope" argument. There are a couple privileges that are not rep-based, but they're rare. The dupe-hammer is tag-based, and standing for election on some sites requires certain badges. So in a sense we're already on the slope, but I agree we need to be careful. I originally proposed this as a 30k privilege and think that idea still works, but I also see value in using criteria that are closely related to the privilege in question, if we think we can get it right. – Monica Cellio May 18 '15 at 16:09
  • @Unihedron Deputy is a proxy for "has a minimum number of helpful flags", and goes with a "not too many bad ones" that's needed to keep this from being gamed. I'd like some measure of meta contribution but it doesn't need to be that one; any suggestions? – Monica Cellio May 18 '15 at 16:11
  • How about top 40% in terms of amount of helpful flags - declined flags without taking into account <3k (or Beta equivalent rep for closing) users when sorting by the net flag count of every site? – Unihedron May 18 '15 at 16:14
  • 1
    @Unihedron comment and post flags, or just the latter? (I don't really want to create an incentive for people to flag every single comment in a thread instead of flagging the post to tell us the whole thread is obsolete.) – Monica Cellio May 18 '15 at 16:17
  • I don't know. Flag weight relative to site, close votes history relative to performance, etc seems more trusty figures than helpful flags that scores a badge (and it's not even that hard, considering I have Marshal on Drupal where I don't participate on, and Deputy on various sites network wide). – Unihedron May 18 '15 at 16:20
  • That's a good point. Some notion of time-based flag weight would give a more accurate picture. Let me think about how we could do that. – Monica Cellio May 18 '15 at 16:22
  • It's been two days - how's your thought experiment? – Unihedron May 20 '15 at 17:46
  • @Unihedron see my edit. I changed it to a minimum number of flags on posts only (no gaming it by searching for "thanks" or "+1" comments) plus a recent track record of helpful flags. The specific numbers can be tweaked, of course, but I think I'm in the right ballpark. – Monica Cellio May 20 '15 at 19:06
  • To add to the "slippery slope" concerns, it strikes me that this would be the only privilege I'm aware of that a regular user can't tell if someone has from info on that person's profile page. It's not just that it's somewhat complicated, it's that the relevant info is currently only visible to the user themselves (and maybe diamond mods, not sure what you guys see). With other privileges, it's either rep-based or based on preexisting badges, while this is not. – cpast May 21 '15 at 2:16
  • @cpast interesting point. I hadn't considered that, though I also don't know how much it matters because anybody who uses the power leaves a public record (the closures). Is it important to know who has the privilege? (Any privilege, for that matter?) We also don't know who's currently flag-banned or review-banned, to approach it from the other direction, so you don't actually know that someone has a privilege just because he has the indicators. – Monica Cellio May 21 '15 at 3:26
  • My concern remains that these requirements aren't a better matrix over the other options to be used. To run on, point 1 is moot because point 2 directly covers that, while point 2 is moot for the same reason in my answer - that "no arbitrary matrix is capable of measuring effectiveness of their ability to close." Point three covers generous values (relative to SO, anyway) though the (?)s kinds of signifies how hard it is to set the marker. Then again if you can't come up with where they are set, I don't believe anyone else could either. ;) – Unihedron May 21 '15 at 3:40
  • 1
    @MonicaCellio I already voted on that last month when you posted it. :) – Jason C May 21 '15 at 14:13
  • 1
    @MonicaCellio If a site is too small to have any users with close hammers, doesn't that mean its small enough for the existing close queue and/or moderators to manage? The whole premise (I assume, although I may be being generous) behind close hammers is giving more users more power to make up for the normal system being overwhelmed on larger sites (which, incidentally, probably warrants a discussion of how to fix an allegedly broken close system rather than giving users extra power as a kludge to make up for it). Do you think a "close hammer" is needed on smaller sites? – Jason C May 21 '15 at 14:18
5

I like this idea.

However, there are problems with the suggestion.

Making closing incredibly easy for a sector of users means that they are treated as stewards for closing.

(In case you didn't know, stewards are those with a heavy responsibility to look after someone's property.)

As a result, they tend to close more than to downvote like anyone else could do. Especially in the case of Stack Overflow, each close voter gets 50 close votes a day, but only 40 up/down votes. Try to convince anyone how they will still vote fairly and not overuse their ability where they should be downvoting instead.

Additionally, no arbitrary matrix is capable of measuring effectiveness of their ability to close. Unlike providing the Close Votes Review queue, users don't pass through "are you still sane?" checks in form of audits. It is incredibly difficult to correct mistakes if they don't get noticed soon enough.

As my last point, closing is for questions with serious problems that are not likely to be improved into answerable state whatsoever. Implementing this feature means sites like Stack Overflow will become a battleground where trusted users battle low quality questions. This is not healthy, for all of these parties:

  • Audience (guests who look for answers)
  • Askers (They might get shot down any moment)
  • Answerers (Accused for not helping to shoot down more questions; Have a narrower source of reputation points)
  • Moderation parties, caretakers (labelled into "stewards" and "non-stewards", divert in moderation decisions causes hammer misuse and frequent problems)
  • The pathetic people shaking their heads (seeing the site officially become battlegrounds isn't fun)
  • 9
    All good points of concern. However, I'd argue one point--that StackOverflow is already such a battleground, but one where opposing forces of good and crap are more evenly matched than I think they should be. – Won't May 18 '15 at 14:37
  • 7
    "closing is for questions with serious problems that are not likely to be improved into answerable state whatsoever" No. That's the criteria for deletion. Closing is a marker that the question either is off-topic or doesn't meet quality standards and can easily be reversed provided OP puts in the effort to improve the question. With closevotes, you should vote early and often, there is no need for wait or making sure it's unsalvageable. – bjb568 May 18 '15 at 14:38
  • 2
    @Won't I'd like statistics or realistic entries showing exactly how Stack Overflow is such a battleground that it won't get worse after implementing this or how this feature would mitigate / reduce / solve the problem you mentioned. – Unihedron May 18 '15 at 14:40
  • 1
    ^ I strongly concur with the above comment, I would also like to see that analysis. I'm not passionately for or against this idea, I just think it would make for a much stronger case, and would also be really interesting. Right now it seems to just appeal too much to a reasonable general desire for more power without a concrete rationale. – Jason C May 18 '15 at 14:42
  • @bjb568: While I wholeheartedly agree with you as one who votes to close with your exact same criteria, the help center states differently. I quote: "[...]. If a question is edited by the original poster when it is marked [on hold], it will automatically be placed in a review queue to be considered for reopening. If it is not reopened within five days, the [on hold] notice automatically changes to [closed].There is functionally no difference between an [on hold] question and a [closed] one; neither can be answered until it is re-opened, but they both allow comments, votes and edits." – Unihedron May 18 '15 at 14:44
  • @bjb568: [continued] Closed questions are quickly split into "Improved" and "Idle" groups. If a post isn't improved in time (5 days), they will not go back into the reopen queue again (and given the attention they get as a closed question, they seldom do). Even if a post is improved, they have to be "approved" for reopening by community users, each with their own judgement. Under this closing system, the "on hold" mark is no longer only a marker for temporal problems, but instead a seal for a high potential for the question to be banished. Forever. – Unihedron May 18 '15 at 14:48
  • Makes sense given that OP almost never improves the question and it's impossible for other community members to salvage it. These questions should die without OP intervention. – bjb568 May 18 '15 at 14:50
  • 1
    I appreciate your judgement but don't think that rationale is what makes this feature worthwhile. – Unihedron May 18 '15 at 14:50
  • I'd definitely agree with having to have a way of tracking the impact and effectiveness of this change prior to implementation. It's a big stretch, no doubt. If the impact is negative, it should be known. – Won't May 18 '15 at 16:03
  • You can prevent battles (via this privilege, anyway) by limiting which questions can be closed in this way, as I laid out in my answer. Don't enable further oscillation; that's no good for anybody. A close-hammer should apply once per question, not every time a privileged user comes along and disagrees with the state of the question. – Monica Cellio May 21 '15 at 12:53
  • @Monica: That doesn't prevent any battles at all, only mistakes in the highly visited spectrum. "sites like Stack Overflow will become a battleground where trusted users battle low quality questions" <- Your answer mitigates none of this problem that arises from this "feature", neither resolves any of the problems I listed here. – Unihedron May 21 '15 at 13:11
  • You seem to be talking about the "close! no, open! no close!" tugs of war that currently require five users on each site (and no repeats IIRC; you only get to VtC once). I'm saying that by limiting which questions a unilateral close (or reopen) can apply to you mitigate that type of tug of war. The broader one is beyond the scope of this feature request; we just want to avoid making it worse. You seem to be saying "can't make it perfect so don't bother making it better"; I disagree with that. But we both have answers, so we should make our cases there, not in comments back and forth. – Monica Cellio May 21 '15 at 14:00
  • @MonicaCellio: "You seem to be talking about the 'close! no, open! no close!' tugs of war that currently require five users on each site (and no repeats IIRC; you only get to VtC once)." No. I'm talking about how Stack Overflow caretakers takes the front page as a waterfall of zombies from Plants vs Zombies. – Unihedron May 21 '15 at 14:08
4

I'm not sure if I like this idea. I may like this idea if a more concrete rationale with evidence is given. However, I dislike the idea of giving more users unilateral powers, as I am personally a big fan of community handling of tasks. I also am not sure if old low-view poor content is necessarily a problem to be urgently solved -- if nobody really views it then it's not really watering down information quality on the site (which is the fundamental problem being identified here).

That said, the dupe-hammer seems to be working well; but that does not cause (or lead to, by opening up the possibility of delete) loss of content.

In any case, presuming that a) the community truly can't handle bad content effectively and b) said bad content is problematic and has a negative impact on the site's quality, I would rather see other ideas for calling attention to such content besides giving more unilateral super powers all around.

The surface problem being identified here is that old content sticks around a long time and not enough people notice it to deal with.

In the current system if somebody does notice the content and CVs it, it will go to the close review queue, with the hopes that ultimately others see it. The OP's proposal does not show that the current close review queue system is inadequate, but let's assume that it is for the sake of discussion.

To that end I think a better solution to the underlying problem would be to prioritize questions with close votes cast by gold badge holders in the close review queue. This is a balance. No new unilateral powers are given, but trusted users essentially gain the ability to call the community to more quickly look at specific questions that they deem are bad. There is no real possibility for abuse here. Does this negatively impact the review queue as far as questions with no gold badge votes on them? I don't know. I don't have the ability to perform that kind of analysis.

Another possible way to address the more fundamental issue of information quality could be to auto flag old low scoring questions that exceed a certain views per time threshold. This draws attention to old, potentially bad questions that may actually be having an impact on the world due to views. What should these thresholds be? I don't know.

But mostly, gold badge close review queue priority is my choice here as a balance. I realize it may not satisfy many people's reasonable desire for more power but I am not necessarily sure that that is a bad thing. Also the success of it could potentially be a foundation for implementing a "close hammer" in the future, or at the very least may provide some interesting data to back up such a proposal.

  • 1
    "... could be to auto flag old low scoring questions that exceed a certain views per time threshold." With the test runs of the Triage queue on Stack Overflow, I think you got it. – Unihedron May 18 '15 at 15:01
  • 1
    If you don't like unilaterality of closing for non-mods, how about increasing the weight of some CVs, possibly only requiring 2 or 3 to close some questions? – bjb568 May 18 '15 at 15:11
  • 1
    @bjb568 I could stand by that as well. I feel like I've seen that suggested before. I was actually looking for it earlier but couldn't find it. – Jason C May 18 '15 at 15:14
  • @bjb568 But I wonder if it would have the desired effect. If it's not drawing any more attention to the question than the current system, does it help? We'd want to gather some data on how many CVs old unclosed questions actually get before the votes decay (assuming that happens) to see if it would be effective. Most of the thoughts in my answer revolved specifically around drawing attention to unnoticed content, but maybe that is a good different approach. – Jason C May 18 '15 at 15:17
  • Sounds like what I proposed here except I was answering a question about what to give to silver badge holders. – Louis May 18 '15 at 16:49
  • Having a tag badge makes you better able to judge if something is a dupe, but it seems completely orthogonal to evaluating whether a question is unclear, opinion-based, etc. – Monica Cellio May 21 '15 at 12:55
  • @MonicaCellio Having a tag badge doesn't make you better able to judge anything. Having a tag badge is currently used as a rough approximation of trust. The dupehammer is given based on that trust. Close vote queue priority could be given based on that trust as well. Gold tag badges aren't a measure of your ability to do anything, they're a rough trust metric, a simplified, already-existing approximation (at least that's their intent) of the set of criteria you proposed. In the end, gold badges and more complex criteria are equally inaccurate. – Jason C May 21 '15 at 13:13
  • @MonicaCellio I don't mind using gold badges to indicate that a specific user is trusted to perform more than the usual moderation tasks on a specific sub-area of a site, fwiw. But a plain old reputation threshold with site-wide privileges works just as well as all the other suggestions here, too. – Jason C May 21 '15 at 13:15
  • Only allowing this for gold badge-holders wouldn't help most sites, but if you're saying do that in addition to something broader, that works. – Monica Cellio May 21 '15 at 14:01
  • @MonicaCellio Reputation would be the appropriate choice then. If improvements to the reputation system are required before it is appropriate for closehammers, then the world is not ready for closehammers yet and reputation system improvements should be discussed first. That said, if smaller sites are lacking gold badges this probably means that they just don't need the users with extra privileges due to their small manageable size. So I actually don't see much issue with that; it seems to be acceptable re: lack of dupe hammers at least. – Jason C May 21 '15 at 14:08
  • I like the auto-flag idea. We have auto-flags for many comments and many answers, so many views in a short time window fits into that pattern. It'd be nice to find a way to make that information available to more than just the moderators, though; while only mods can clean up long comment threads, lots of people can take appropriate actions on suddenly-popular questions that might have issues. – Monica Cellio May 21 '15 at 14:20
1

Instead of giving trusted users nuclear weapons, why not give them bigger guns? Instead of being able to unilaterally close a question, just have their close vote count as 2 or 3 or 4 votes. Then you still get some community oversight.

You must log in to answer this question.

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