feature-request status-completed

Don't forget also the trivial ability to reverse your own close-vote (e.g. in a case a question was improved significantly).

As the Mechanical snail points out, this has now been implemented, though not quite as-requested.

Of note:

• Only accessible from the review queue (so rather difficult to target a specific question). Yes, this is very much by-design.

• Voting against closing does not override anyone's close vote. However, a sufficient number of "Do Not Close" responses (currently 3) will kick the question out of the review queue and start aging the close votes - regardless of how many views the question has had.

• If the question is closed, Do Not Close votes do not translate into re-open votes. However, we may use them to prioritize items in the Reopen Queue.

The philosophy behind this is reflected in my response to Miles' bounty:

The development of new Stack Exchange sites has led to a disturbing tendency for on-topic questions to attract close/migration votes from a minority of users simply because they are arguably "more on topic" at a different Stack Exchange site. The "silent majority" cannot prevent such migrations. The suggested (and notably unpopular) approach of monitoring questions until they are closed, and then voting to reopen, does not work in migration scenarios. Let's revisit this feature request.

When you find a problem with the way in which folks are behaving on the site, try to correct it with a scalpel, not an axe. The "silent majority" can prevent such migrations:

• If you see a question being closed that shouldn't be, leave a comment expressing your rationale. Make it constructive - "I like this question, therefore it should stay, close-voters are stupid" accomplishes little.

• If you see a good, on-topic question in danger of being migrated, flag it - a moderator can always step in to prevent the migration. Note that we can and do review migration paths available to ordinary voters to address problems with migrations.

• Answer the question. Most of the sites on Stack Exchange are not available as migration targets for non-moderators. Moderators are encouraged to decline flags asking for clearly on-topic questions to be migrated. And nothing says "this question is on-topic" quite like a good on-topic answer.

We're also working on revamping the "review" tools to put questions on the path to being closed in front of those with the most expertise in their topics. I can't always tell if, say, an r question is better off on Cross Validated, but there are plenty of folks who can. This will also give us some better data for determining when the silent majority has actually reviewed a question, and the ability to then age close votes accordingly. Which is really what you want.

Requiring that "silent majority" to go around casting "unclose" votes wouldn't do anything but create more work within the system for the benefit of a few edge-cases and a rather larger increase in the ability for griefers to waste the time of those already going out of their way to review and moderate. It's a tool for creating gridlock, something Stack Overflow in particular doesn't need any more of.

this seems to come down to adding another dimention to the voting system.

        interesting
            ^
            |
            |
proper<------------>improper
            |
            |
        uninteresting

we already have buttons for voting on the intersting/unintersting axis
the buttons for voting on the open/close axis should simply be labeld
'close' and 'open'.

if close > x+open 
   then mark as closed

It is already "possible" to do this!

Step 1

Go to the new Review interface for close votes.

Step 2

Click through the ≈ 60,100 questions (on SO) with close votes until you see the one you want to vote not to close.

Step 3

Click Do Not Close.

Step 4

Profit!

And...problem solved!

But seriously

The functionality already exists, but is hard to use. If you agree with a close vote, you can always cast your vote to close directly from the question. Why should it be harder if you disagree?

(Also, it seems the queue interface is biased toward more recent questions, so you wouldn't have to click that many times. And voting not to close doesn't literally cancel out a close vote, but if several people do it the question is removed from the queue.)

More discussion

Shog pointed out that close votes come with reasons to close, and so contain more useful information. (Here we're not concerned about obvious duplicates and the like; people aren't going to be disputing those: only the more complex cases.) Shog is afraid that allowing no-close votes will lead to people canceling out close votes without providing a useful rationale. I'll grant you that's a problem, but the problem is symmetric: close vote reasons are vague and often require additional explanation for why they apply, and most of the time close voters don't explain exactly why they think it's e.g. a duplicate. Fundamentally it would be no harder to comment along with your no-close vote from a question, than to add a useful argument along with a your close vote for a non-clear-cut question. I therefore see the no-close-vote proposal as orthogonal to the problem of people not providing rationales.

For example, I have seen a fair number of duplicate votes for which the questions look subtly different; it may require a minute or two of reading to determine whether they really are dupes. In these I don't recall ever seeing comments like "This poster's underlying problem is xxx which is the same as the other question".

Because it's easier to get questions that should be closed out of the queue, the current system encourages questions getting closed and reopened repeatedly. Lack of explanation is a different problem; perhaps adding a comment field to the (no-)close dialog would help.

Would the same work for reopening too?

Would it reopen the moment the number of close votes dropped to 4, or would it have to gain a net score of 5 reopen votes to be reopened?

This is a splendid idea. But the whole concept must be as simple as possible: "Vote not to close" is in fact the same as open, so I think it could be simplified. The whole concept could be as in the case of score votes. There would be two buttons, available all the time, regardless the state (open/closed/on-hold) that would operate on current number of close votes (0-5):

• close (+1) - raise number of close votes. If it reaches 5, the question goes to on-hold/close state and no more close votes can be cast.

• open (-1) - lower the number of close votes. If it reaches 0, the question is goes to open state. It cannot go below zero.

This concept is very simple and easy to understand, and also has all the benefits we starve for:

• vote not to close (294 votes now),
• undo close vote (691 votes now).

Sounds like you'll get a closing war: close it, no, yes, no, yes,... with a long comment thread why and why not. Than SO is the discussion group and not Meta. I don't like this scenario.

I strongly agree with the need for amending the current completely non-closing process democratic - which is by the way illustrated very well by the voting in this particular discussion (as of 25.11.2015, +421 votes supporting the idea, -85 votes disagreeing with the Jeff Atwood's rejecting answer).

I suggest some improvements to the original proposal:

1. If implementing the new counter is too complicated, we may re-use the already existing upvote counter. If 10 users upvote a question, this ipso facto most probably means that at least 10 users do not want to close that question.

By the way, I came MUCH TOO OFTEN via Google search to questions which I found very useful and on-topic, and based on their upvote score together with me tens of other users also did, but the question was closed because a small [sarcasm]elite group of 5 überusers[/sarcasm] decided that the question is off-topic.

2. I would replace the rule

Only when the vote to close outnumbers the vote not to close by 5, that's when a question is indeed closed.

with statistically more relevant

Only when the vote to close outnumbers the vote not to close by 150% (with minimum absolute difference of 5), that's when a question is indeed closed.

So acceptable scores for closing are e.g.: 5:0, 15:0, 15:10, 150:100, but not 16:11 or 150:145.

Please leave the closing/reopening system alone. It works fine as is. There is a bias towards reopening, but most of the noise is getting purged properly.