12

On a given question a comment is posted:

please refrain from voting to close/reopen this question until a consensus has been reached on Meta.

And the question is locked to enforce this opinion - that the question remain open until a subjective question on Meta is answered.

Why are moderators disabling the community consensus on Stack Overflow with the built-in voting mechanism and forcing the community to go here?

Isn't that the purpose of the voting mechanism? Why not allow both communities to deal with it, and once Meta reached a decision a moderator can enforce it?

Preventing normal Stack Overflow participants from voting via the normal mechanisms and favoring one side over the other seems to be an exceptionally poor choice, and I'd like to understand the reasoning behind it.

|improve this question
  • Do you have a link to the question? – juergen d Jul 24 '13 at 16:56
  • 1
    I'd like a general discussion, rather than one focused on the particulars in this situation. However, here are the questions which made me wonder if this was policy, or if this is exceptional: stackoverflow.com/questions/17807531/… and the meta discussion: meta.stackexchange.com/questions/190106/… – Adam Davis Jul 24 '13 at 16:58
  • 6
    @AdamDavis: It is standard policy for controversial questions that jojo between states / edits / attract loads of comments. – Martijn Pieters Jul 24 '13 at 17:06
  • 1
    Consensus does not necessarily need to be reached on Meta, but it's not a bad idea to take such significantly disputed questions to a somewhat more neutral setting. At least arguments for or against have a somewhat better chance of being listened to there than under the question itself (which is not really the place for it to begin with). That doesn't mean you have to agree with the "consensus" though. You're still free to vote as you see fit once the dust settles. – Bart Jul 24 '13 at 17:12
  • @Bart But then why disable voting? – Adam Davis Jul 24 '13 at 17:15
  • 3
    Voting isn't disabled, it's moved to the meta question. Close/open seesaw isn't an efficient voting mechanism (in particular, it doesn't allow MORE than five votes at a time in one direction, then five in the other, etc.) Locking the thread forces that voting effort to move to meta, which is for the better. – Joe Jul 24 '13 at 17:17
  • 4
    @AdamDavis OK, seriously? You think that question (and its answers) needs more votes than it already has? – Robert Harvey Jul 24 '13 at 17:17
  • 2
    To enable informed voting afterwards @AdamDavis? That's at least an advantage I see. You can ultimately still vote as you see fit, but at least you are more likely to listen to the arguments made. – Bart Jul 24 '13 at 17:17
17

It doesn't need to. But like it or not, Mark's meta post brought a lot of additional attention to that question - which intensified the voting even while arguments were being made and refuted here on meta.

It's practical in these situations to make sure that there is time for discussion to be had. Locking provides an opportunity for this.

The post shouldn't have been locked indefinitely though; this is exactly why we have timed locks. I've reduced the period to 24 hours; that should be enough time for folks to weigh in on the discussion; at that point, voting can resume. To be clear, the final decision on the fate of the question should still be determined by voting on Stack Overflow, but ideally folks are voting informed by prior discussion.

|improve this answer
  • You are just pushing off the issue and the vote war will be had after the lock is gone - there will still be those, particularly around questions such as this, which will not be swayed to the other side of the issue. Locking the question doesn't necessarily prevent the close war. But my question is more basic - what are the bad effects of the normal process in this case? I can understand the point about giving people a time to reflect before casting their one and only vote - but I don't think that's a good enough reason to stop voting. Once the link is in the comments, they can wait or not. – Adam Davis Jul 24 '13 at 17:28
  • 3
    I'm not particularly concerned about a close war - no one can vote more than once anymore, so as far as I'm concerned close-warring is a mostly-historical oddity. Focusing on making good arguments in the active discussion (rather than just voting and remaining unheard) is important though, and I think encouraging folks to focus on that is helpful. Tangentially-related: meta.stackexchange.com/questions/166329/… – Shog9♦ Jul 24 '13 at 17:32
  • "encouraging folks to focus on that" by removing their vote entirely and forcing them to visit meta (when the vast majority of SO users don't visit meta) seems quite excessive for "encouragement." – Adam Davis Jul 24 '13 at 17:35
  • 4
    There's a fairly prominent link to the specific discussion in the comments there, @Adam. If someone has a specific interest in the question, I tend to think it will be more helpful than just seeing the question closed/reopened/edited coupled with the fairly bare-bones discussion that was happening in comments. – Shog9♦ Jul 24 '13 at 17:41
  • 4
    Seems to me you're telling us that we're actually hitting two birds with one stone @AdamDavis. Informed voting, and an introduction of more people to Meta, which is an essential part of the network. Win-win I'd say. Heck, add to that potential future relevance of the discussion to similar situations and we're up to 3 positives. – Bart Jul 24 '13 at 17:42
  • @Shog9 Right - the link was there, and those interested can go interact with others similarly interested if they so desire. But if they don't want to, and they simply want to cast their vote - they can't. Why are we removing that choice from them? Why not allow both the discussion on meta to continue, and allow people to continue to vote? What is the specific harm to Stack Overflow or its community that is being prevented by disabling voting? – Adam Davis Jul 24 '13 at 17:51
  • 3
    The specific harm in this case would be knee-jerk voting from folks finding the question via meta without bothering to read or participate in the actual discussion, @Adam. – Shog9♦ Jul 24 '13 at 18:37
  • @Shog9 Wait, wait, wait - so the problem was actually that the question was raised on Meta, and because of that we have to switch off voting on Stack Overflow? How bass ackwards is that... – Adam Davis Jul 24 '13 at 18:39
  • 3
    I did kinda say that in the first line of my answer, @Adam. And it is what it is - questions that get a lot of traffic end up having different problems than average questions, and those problems generally depend on the source of the traffic. – Shog9♦ Jul 24 '13 at 18:45
  • While I disagree with the policy, this does appear to represent policy so is the best answer for the question. – Adam Davis Jul 25 '13 at 21:49
18

Some questions attract close wars. 5 people vote to close. The minute they succeed, 5 different people vote to reopen. The minute they succeed, another 5 people vote to close. None of these people can vote again later. Rather than "using up" all these close votes, people should hold on until a somewhat larger group of people can weigh in.

Locking an open question does not favour the side that wants it open. (For example, new answers can't be added.) A closed question and its answers can still be voted on and gain rep for those who got in before the close; a locked question and its answers cannot.

Finally, something drastic like locking is one of the only ways to stop an ocean of comments that typically don't say anything that was not said before. Those with actual content to contribute on the meta-issue of whether the question belongs are typically willing to come to Meta and contribute it. The driveby commenters "oh doodz Y U hate fun?" are typically not.

|improve this answer
  • Since only one vote per person is allowed, the close war is nothing more than a voting booth which ultimately does terminate all by itself. I can see pushing people to comment onto meta, but locking them out of voting removes their ability to vote on the matter, thus removing consensus. – Adam Davis Jul 24 '13 at 17:09
  • 2
    Sure, except it a) is irritating and b) takes a really long time to resolve since it's 5 people at a time. Further, there's no real opportunity for actually discussing; we don't usually have elections without first having debates, after all. That's what meta is. – Joe Jul 24 '13 at 17:13
  • @Joe It only takes a long time to resolve when one side is losing, yet persistent. If you look at the voting record it is being closed far faster than open - one side is clearly winning, and eventually the interested parties will cease to have votes left, and it will stay closed. Why are we stopping this process? – Adam Davis Jul 24 '13 at 17:20
  • 2
    Because the see-saw voting process is inefficient, while actually discussing things first is much more efficient. Meta's voting options are far superior than this annoying see-saw bit. – Joe Jul 24 '13 at 17:24
  • 2
    @Joe But meta is not binding - unless a moderator chooses to make it so. Further, there are many, many, many stack overflow users that do not visit or care about meta. Why are we disenfranchising them? – Adam Davis Jul 24 '13 at 17:31
  • 3
    @AdamDavis: "Why are we disenfranchising them?" Because people who refuse to participate in MSO even when provided with a specific link to an active discussion... are disenfranchising themselves. This isn't the same thing as people who don't regularly visit MSO; you're talking about people who choose not to use MSO. Meta is part of Stack Overflow. You can choose to participate in it or not. But if you choose not to, it's just like choosing not to vote. – Nicol Bolas Jul 24 '13 at 19:20
  • "But if you choose not to, it's just like choosing not to vote." No, it's not at all like choosing not to vote. They have a vote, the vote is being withheld. Even if they visit meta they still can't vote. – Adam Davis Jul 24 '13 at 19:48
  • 6
    This is a reason for locking a post, but I think your last paragraph nails the real problem: if discussion is necessary and is causing comments to go off the rails, then diverting this discussion to meta gives it some breathing room. – Shog9♦ Jul 24 '13 at 20:08
12

Oscillating randomly between two states does not achieve any consensus beyond the simple fact that it is contested.

If the community is split close to equally on an issue then it would seem to suggest that the problems are more complex and deserve more detailed inspection in order to reach an outcome that is satisfactory and reasoned to a point where both sides can agree.

Escalation to Meta makes sense and avoids the flip flop madness - it's the only practical public escalation method available.

|improve this answer
  • 6
    I detect some oscillation... – Andrew Barber Jul 24 '13 at 17:05
  • 2
    Since everyone can only cast one vote then ultimately a decision is reached - it won't oscillate forever, and the side with more votes "wins". The system is designed to operate this way. – Adam Davis Jul 24 '13 at 17:07
  • 3
    But if the number of voters is high and the split very close then the difference will be tiny and the result won't be statistically significant. Then we're forced to resort to reasoned arguments instead of raw numbers in a first past the post voting system. – Flexo Jul 24 '13 at 17:11
  • 4
    There are over FIFTY SEVEN THOUSAND users who have earned a badge for casting their first flag. If just those (never mind those who never flagged or voted to close before) weighed in on this, 5 at a time, how long would consensus take to achieve? – Kate Gregory Jul 24 '13 at 17:14
  • 3
    @KateGregory If that is truly the problem, then we should significantly overhaul the system. But that isn't the issue - they aren't all going to weigh in, and the vast majority will never see the question anyway. – Adam Davis Jul 24 '13 at 17:22
6

The normal mechanisms only work for normal situations. Sometimes there are exacerbating conditions that render the normal tools ineffective.

One of those conditions is a bike shed style question that hits the multi side collider.

The normal tools such as VTC are calculated to work well under normal loads. 5 people close a questions. 5 more re-open. Normally you might not have more than 5 or 10 people with those kind of privs hitting a question anyway so closing takes some doing. The problem is those tools fall apart when you have hundreds of people with both privileges and opinions on a matter. 5 close. 5 reopen. 5 close. 5 more reopen.

That's just stupid. The normal tools aren't suited for this case, they don't help resolve the issue and using them in a game of repetitive tug of war does not do anybody any favors.

Enter a tool for the job. Lock + meta. Because meta can actually handle a log of people throwing up an opinion and sorting out which way the community really wants to take it.

|improve this answer
  • 1
    So you contend that the normal voting mechanism isn't working for this question. In what way has it failed? – Adam Davis Jul 24 '13 at 17:28
  • 6
    @AdamDavis Yes, I contend that the normal voting mechanism isn't working for this question. It's been closed/reopened 7 times! A close war does not determine community consensus. That's not to mention the little edit war. When that stuff starts happening it's a sure sign that you need to escalate to the next level of tool for determining community consensus. Meta. – Caleb Jul 24 '13 at 17:36
  • Of course a "close war" determines community consensus. The system only allows one vote per person. Eventually the side with more users "wins". That's absolute. – Adam Davis Jul 24 '13 at 17:38
  • 3
    @AdamDavis No actually it doesn't. Flexo articulated that bit well in his answer: "Oscillating randomly between two states does not achieve any consensus beyond the simple fact that it is contested." Even if you went through every VTC capable voter on SO, the final state of the question would depend on how the die rolled on who the last set of 5 was. As soon as it's obvious that an oscillation is starting, that's your clear symptom to break out a the tool made for that job. And lock it so people don't bash their heads using the wrong tool. – Caleb Jul 24 '13 at 17:40
  • It's hardly random, and it's not even 50/50 in this case. If you look at the timing pattern in successful close/open status you find that the closing is happening more quickly than re-opening. It's quite clear closing is winning - there are more people interested in closing the question, so closing happens more quickly. It's not at all random. And the vast majority of users will never see this question. The "roll of the die" would only matter if there truly was an exact 50.000% split, which is not the case here. – Adam Davis Jul 24 '13 at 17:46
  • 1
    @AdamDavis It might be apparent to you that closing is winning -- and I'd even agree with you that you are right, esp with the mods being on the close side repeatedly. So if it's winning why let it keep oscillating? That's what's broken. Even if it was only a couple more cycles, that isn't serving anybody well. If you can look at it and made the judgement call that close is the right option AND community support, then close + lock resolves this until people give it a rest. Meanwhile meta has your back, figuring out if that actually is community consensus + why. – Caleb Jul 24 '13 at 17:49
  • That's actually what I'm asking, but in reverse. Why not let it keep oscillating? What is the specific harm or damage being done to Stack Overflow that disabling voting has or is preventing? Why not allow both voting and community discussion? – Adam Davis Jul 24 '13 at 17:53
  • 1
    @AdamDavis: For the exact reasons you've mentioned before: given a choice, a lot of people simply won't participate on MSO. By forcing them to come to a place designed to hash out these issues, we get the issue actually hashed out, rather than playing the oscillation game. – Nicol Bolas Jul 24 '13 at 19:24
  • @NicolBolas Again, what harm occurs from oscillating? It is naturally damped. It isn't going to shake Stack Overflow apart. Why is that a reason to take away someone's vote? – Adam Davis Jul 24 '13 at 19:49
  • @AdamDavis: "It isn't going to shake Stack Overflow apart." I could turn the question around: it's not going to shake SO apart to stop it from oscillating. – Nicol Bolas Jul 24 '13 at 19:58
2

I would guess that it's an effort to prevent more controversial posts or posts that are kind of "on the fence" from swapping between opened and closed, and to make people aware that there is a discussion on Meta where they can bring up their concerns and have a more open discussion on the topic.

Part of the function of Meta, at least from what I've seen, is to provide a space for these kinds of debates so that posts on the main site don't fill up with noise.

On the other hand:

In many cases I question what constitutes "a consensus" on Meta. It seems that oftentimes posts that are controversial on the main site are just as controversial here, hence reaching any real consensus can be a bit murky.

The example you sited in the comments seems like a great example of the sort of murky consensus I'm referring to. The decision still looks a little bit split...

Referring to Shog9's answer on the consensus issue perhaps the real problem is that we have an awful lot of sadomasochists who want to enjoy free ice-cream while they disembowel people.

Democracy can be fickle and slow, but it sure beats letting the moderators make all the calls. When the call gets made on Stack Overflow, it's generally just an anonymous vote for option A or option B. When the call gets made on Meta, users have the chance to voice their opinions, be heard by their community, and possibly even mention option C which no one even considered.

|improve this answer

You must log in to answer this question.

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