128

When a close vote expires, it can no longer be re-cast.

This didn't use to be the case: back when close votes always expired in 4 days, you could re-cast your close vote after it had expired. This behavior was unintended, and now you only get a single chance at voting to close.

This is no good outside the fast-paced world of high-volume tags on the Trilogy. On a typical SE2.0 site, here's what happens:

  1. A poor question is posted. Comments tell the OP to improve the question because it's unanswerable in its present form.
  2. A couple of people cast close votes. Less than 5, so the question remains open.
  3. The OP never improves the question.
  4. The question gets more than 100 views over time, because it has a promising title.
  5. At some point the question resurfaces. The people who cast the close votes in #2 want to close it for good, but they can't cast a close vote now.

Please allow the closers to re-cast their votes at #5. This may have been unintended behavior before, but it was the right behavior.

(Note that this is not about close/reopen wars. The question has not been closed at any point.)

2
  • 4
    There's always moderator flags.
    – user102937
    Jul 21, 2011 at 20:31
  • 12
    @RobertHarvey Or, as we call it on U&L, the @MichaelMrozek method. I wish we had a better solution. Jul 21, 2011 at 20:59

7 Answers 7

14

Votes to close and reopen, as well as flags to close, which age away on the post can now be re-cast 14 days after they expire. This was implemented along with some other changes to the way these votes age away, which includes starting the aging process after 14 days regardless of view count.

Do note that votes cast before 2012, at the time you asked this question, cannot be recast because the fact that the votes were invalidated due to age (and not some other reason) wasn't saved.

40
+700

The main problem with the current system is that it ties a user vote to a random variable - whether other people happen to see it and agree with it. They are gambling that right now is the right time to make their vote effective, and whether it's a good gamble or not, it's still non-deterministic from their point of view.

If we are going to expire votes, they should be refunded so that they are up to the user's continued action on that post rather than up to a random process depending on others seeing and agreeing with the issue.

I don't think juggling is a problem - a user can juggle it if they want to spend the time and effort. We do the same thing for area51 proposals - they slowly dissipate unless a user visits them occasionally and refreshes them.

Keep in mind that the main reason we expire votes is that if we don't then ultimately everything will be closed. Expiration, but refunding them, will still solve the problem of "entropy heat death" of a question. Only questions where several users continually work to close it will be closed. And once it's closed, if it's wrong it will be quickly reopened, and the original closers will have cast their one effective vote so they won't be able to participate again.

The system would still be self-limiting, and meet the original goals of the expiration while giving users the ability to continue to express their views until they are taken into account at least once.

8
  • 4
    But then they're just stringing along their personal belief that the vote should stick. If there aren't enough eyes period on the post to make something happen, it's probably indicative of a more general problem with the community being able to moderate itself (due to size, or the high barrier to seeing the vote queue, etc). Otherwise, if there is plenty of eyes and people just aren't agreeing with the vote in a timely manner, then the outcome seems correct.
    – Tim Stone
    Mar 1, 2012 at 21:51
  • 1
    @TimStone But the key difference is they have to continue to take action. Keep in mind the primary reason the expiration is in place is to avoid the fact that left unchecked everything would ultimately be closed. The expiration takes care of that, but there's no need to make the expiration permanent. Letting people keep re-activating their votes won't lead to the heat death of every question - only those that continue to garner negative interest.
    – Pollyanna
    Mar 1, 2012 at 22:28
  • 1
    There is a difference, but I'm not sure how significant it is. Instead of counting on the votes trickling in, you're basically just spinning the wheel until you get lucky. I guess I'm just not convinced that the problem (to the degree that there actually is or isn't one) is actually anything to do with the vote expiration.
    – Tim Stone
    Mar 1, 2012 at 22:53
  • 3
    @Tim: The problem is the system is very complex and doesn't map to the users' expectations. As a user, I expect my vote-to-close to count equally whether I do it today or in, say, three weeks. But the system sometimes makes my vote today more effective and sometimes makes my vote in the future more effective. There's a complicated calculus required to estimate which situation you might be in. Until yesterday, I had no idea (or had forgotten) how the system works. Shouldn't the system just encourage immediate voting on questions that ought to be closed or reopened? Mar 2, 2012 at 18:56
  • @JonEricson Well, ideally if you vote to close bad content as soon as you come across it, there's a good chance that that content is still in its "high visibility" period (on the front page). If that's the case, the vote expiration shouldn't matter since the community should adequately handle it during that time. If it survives that period, the community either doesn't have enough people who can deal with the problem, or not enough people who agree that there's actually is a problem. I feel the biggest problem with the expiration is really that people know it exists to worry about it.
    – Tim Stone
    Mar 2, 2012 at 19:10
  • Granted, I don't assume that things always work out like they should by any means, but on the whole I'm not sure there's a significant issue that's directly solved by removing the expiration.
    – Tim Stone
    Mar 2, 2012 at 19:12
  • 4
    @Tim: We agree that it's an edge case and won't solve the social issues. But I think this edge case is worth smoothing over somehow so that people aren't even tempted to game the system. Note that most users won't see the issue, but the users who do see the issue will probably feel cheated out of their vote. Whether they are or not is a side issue by the way. Preventing people from feeling like the system cheated them is worth making a change even if it probably will not have any other major effect on the community. Mar 2, 2012 at 23:02
  • I've been burned by the expiration too many times on a small site where only a few active users had VtC privs and the mods wouldn't intervene unless it was egregious. We should not be encouraging users to guess and wait for the right time, or frustrating them because they come across 6-month-old junk a second time and can't recast a vote that didn't take the first time. If this doesn't happen often then great -- most people won't notice. But where it does matter, it makes a big difference. Mar 3, 2014 at 21:39
27

I agree, please implement this

If you are concerned about a user being able to "juggle" a question with pending votes infinitely, than prevent users who have unsuccessfully voted in the past from initiating the first close/reopen vote. Or implement a timer which begins after the last vote expires and clears all past unsuccessful close/re-open votes so they can be re-cast. Perhaps 14 or 30 days.

I am asking for this because I do not like seeing someone making an effort to close or re-open a question, only to find that I can't assist them because I already voted to close/ re-open it at some point in the past. I find this especially irritating on smaller sites, where the active user base is much smaller than Stack Overflow.

1
  • When I've encountered this problem in the past (casting the fifth vote, counting expired votes), I've just flagged the post for moderator attention and explained the situation. The post was promptly closed by the mod.
    – JDB
    Mar 3, 2014 at 20:33
7

Given the recent fuzzy-ing of close vote review it would seem especially important to allow users to recast expired close votes.

With the close vote review focusing on reviews with multiple close votes (3 at the moment), It looks like an awful lot of close votes are going to expire without being reviewed. While I agree that clearing the backlog is a worthy cause, we shouldn't be throwing countless close votes away to do it. If this is completely unavoidable, users should at least be able to recast close votes until the close vote review is a little less "fuzzy".

3

The real issue is that the question isn't getting enough views for the required number of close or reopen votes for its state to be changed. (Well maybe it is, but people don't agree that it should be closed/reopened — but that's a different issue.)

There are already mechanisms for dealing with this:

  1. Ask in the site's chat room for others to vote on the question.
  2. Ask on the site's meta for others to vote on the question.
  3. Flag the question for moderator attention using the "other" option.

You can then explain why you think that the question should be closed or reopened (and hopefully improve the question in the process). The first two options will hopefully get more views on the post and people cast their vote as appropriate. The last option is for low traffic sites or (somewhat perversely) high traffic sites where a question can easily get lost.

Which you choose is up to you. But speaking as a moderator I'd like to see more flags on improved questions that allow them to be reopened.

5
  • The question is a feature-request for a change to the system, and your answer is simply a workaround using the existing system. Every one of your options relies on finding someone else to take an action because the system does not allow you to take action yourself. I'd rather see a change to the system instead of having to get someone else involved all the time.
    – Rachel
    Sep 20, 2012 at 15:37
  • @Rachel - it was presented as a workaround.
    – ChrisF Mod
    Sep 20, 2012 at 15:38
  • 1
    Also as an example, I know of at least one case where Yannis said he'd reopen a question in chat if it got more reopen votes, but I wasn't able to vote to reopen because I had already cast an unsuccessful reopen vote in the past. I'm not sure how often that comes up, but if moderators are on the fence and consider reopening based on the number of reopen votes already there, I'd like to be able to cast my vote again.
    – Rachel
    Sep 20, 2012 at 15:38
  • @Rachel - if you are able to recast your vote then the question may well remain in it's current state forever. What you want to happen is for it to be closed/reopened and that requires more people seeing the question. What I suggest are ways for that to happen.
    – ChrisF Mod
    Sep 20, 2012 at 15:40
  • I think reopens are rare enough that the chat/meta/flag approach is sustainable on non-SO-activity-level sites. It's heavyweight for closures, I don't relish the vote-to-close chatrooms that some sites have. Fortunately the new close vote queue (open to all closers, not just 10kers) should take care of that. For reopening, there's also the different problem of getting attention to the improvement. (Cc @Rachel) Sep 20, 2012 at 19:39
1

Counterpoint: This would make expiring close votes all but superfluous. Say that this is implemented on SO; even if every vote expires after four days, during that interval I would acquire 200 close votes. Every day, I could easily review my browser history, re-vote on any that expired, then use any leftover votes on new questions.

Glancing at a couple niche tags I'm interested in on SO, it looks like it could easily take as much as two weeks for many questions to hit 100 views. That would allow me to have 700 or more close votes in effect simultaneously and all but permanently. Low-traffic sites would have a similar situation.

Now, to be honest the idea is a bit appealing, but it seems to run somewhat counter to the whole idea behind how close votes work. In particular, I'd be suspicious of a significant overlap between people who would misuse their close votes and people who would go to sufficient lengths to keep reviving their votes as I described.

I agree with you about it being hard to get old, crappy questions closed, but I'm not sure this would be the way to do it.

2
  • 6
    Expiring close votes isn't good for low-traffic sites/tags in the first place, that's why the behavior changed recently. As far as I've heard, no SE 2.0 site has any problem with users abusing close votes, but most have trouble gathering 5 close votes. Now the requirement for closers to synchronize have gotten even worse, which ruins the chances of closing without moderator intervention. Jul 21, 2011 at 20:49
  • Fair enough. I'm mostly active on SO, as you may have gathered, and I definitely sympathize with difficulty in getting low-visibility questions closed. I probably agree with you overall, despite the "devil's advocate" argument above.
    – McCannot
    Jul 21, 2011 at 21:27
-4

Currently close votes expire in a 4 day sliding window, once per day, and we limit users to one close and one reopen vote per question.

(as you noted, close votes no longer expire ever on questions with < 100 views.)

Removing the "can only cast one close or reopen vote on a question" rule would mean a single user could effectively "juggle" a question with 1 pending close vote forever.

Can't say I am a fan of that.

13
  • 9
    You're still thinking in terms of high traffic sites. The counter case which we face on low traffic sites is where a low quality question gets one or two down votes but other voters hold off until they see if the OP comes back to improve the question. If they don't, they will chip in to vote, but by that time the early voters can no longer lend their weight and in a site like U&L where really only 5 or 6 people are voting at all, we struggle to keep the place clean without making the mods do everything. This is distinctly different than "juggling".
    – Caleb
    Jul 21, 2011 at 21:01
  • 2
    You're discussing one isolated example of dozens on a specific SE 2.0 site -- why can't the answer be "in these rare cases, use moderator flags?" Jul 21, 2011 at 21:02
  • 10
    Because it's hardly rare and we feel like we're abusing our precious moderators as it is. As Gilles points out we have named that action after a moderator. The problem is there are actually lots of potential close candidates lurking around, but right now it's not even worth voting on them. Either they are bad enough to make a moderator take care of it or they sit around because we can't do much. Having a moderator do things defeats the point of voting and seeing if other people are for or against closing or have better solutions. It makes our community ineffective at leveraging consensus.
    – Caleb
    Jul 21, 2011 at 21:09
  • 3
    I guess I could support increasing the no expiration threshold from 100 views to 200 views, beyond that, it is as designed. Jul 21, 2011 at 21:11
  • 4
    Is the problem with questions not getting closed an insufficient number of views, or is it the lack of a critical mass of users who can cast close votes? Perhaps a better metric to tweak might be the number of votes required to close, say 3 or 4 close votes instead of 5 on the low-velocity sites.
    – user102937
    Jul 21, 2011 at 21:15
  • 2
    @JeffAtwood On U&L, going from 100 views to 200 views almost doubles the number of applicable questions. Better, but still not ideal (this kind of questions typically gets a bunch of initial views, then gets forgotten for a while). Is “100 views since the close vote was cast” implementable? Even better, I think, but harder to implement: “2 days plus 50 views” (not sure about the exact numbers, but the point is to exclude the early views from the count)? Jul 21, 2011 at 21:20
  • 3
    @RobertHarvey Lack of a critical mass of users who can and do cast close votes. Most of the 3k users will see the question once and then forget about it. I'd rather keep 5 votes, but lower the requirements on /tools?tab=close (on U&L, until a few weeks ago, there was a single non-♦ 10k user). Jul 21, 2011 at 21:22
  • 2
    @Robert: I think for us it has much more to do with the fact that our voting user base is small and active close-voters is ~6. When you make two or three of those totally ineffective, you rob the community of it's ability to be a community. Why can't the number of votes and the vote timeouts be somewhat proportional to the user base? Why is U&L required to jump through as many hoops as SO when it doesn't have 1000th the people to send jumping?
    – Caleb
    Jul 21, 2011 at 21:23
  • @Gilles: The only problem with lower requirements is that you'll be raising the bar later (when the site gets larger), and disqualifying some people who already have the privileges. We already do that in beta, and it sucks when it happens to you.
    – user102937
    Jul 21, 2011 at 21:28
  • 2
    @gilles you realize we get 600+ moderator flags PER DAY on Stack Overflow, right? And you're worried about sending, what, 2-3 flags for closure a day to a community moderator (and you have THREE OF THEM)? I'm sorry, can you explain to me again what the issue is here, because I'm surely not seeing it. Jul 21, 2011 at 21:29
  • 8
    ...would it help to have "Don't hesitate to dump work on your community moderators" as official policy somewhere? Because that's what Jeff seems to be saying to do, and feeling "like we're abusing our precious moderators" seems to be the main reason given for not flagging with extreme prejudice.
    – McCannot
    Jul 21, 2011 at 22:06
  • I feel allowing users to recast their vote would be better than increasing the view threshold. Although the majority of close votes are cast on questions with under 100 views, the majority of re-open votes are cast on questions with over 100 views, and this makes it particularly difficult to re-open older questions with a lot of views/votes because many users who would have voted to reopen the question have already done so in the past.
    – Rachel
    Mar 1, 2012 at 21:16
  • 2
    To address this concern, why not just prevent users who have unsuccessfully voted in the past from casting the first vote?
    – Rachel
    Mar 20, 2013 at 12:58

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