8

According the the comment reply FAQ, it's possible to @-notify close voters in a very particular set of circumstances:

For questions: The moderator or gold badge holder who closed or reopened the question, provided they were the only one to do so. Users who have closed or reopened a question without a binding close vote (i.e. without a gold tag badge or a moderator vote), and those who bindingly voted, but other users were also involved (e.g. closed by User1, User2, and Moderator) cannot be notified. [emphasis added]

However, here is a question which was closed by two regular users and a moderator, where the moderator involved (me) did receive an @-notification, without having left any comment.

Note that two years ago, a bug report about a missing notification was marked . The answer to that report was used as the basis for the FAQ text.

System working differently from designed? A documentation error? An intentional but undocumented change? The new normal? Inquiring minds, etc.

To clarify, this does work as per the above specification for binding duplicate closures by non-mod gold badge holders (i.e., there's no notification in that case); the notification is only raised for closures by a moderator (including cases where a moderator also happens to possess a gold badge and closes as a duplicate).

10
  • Note: at this moment, the linked question has three reopen votes, so it may be not-closed in the future.
    – rob
    Mar 27, 2020 at 1:03
  • 1
    @Rob This contradicts the highlighted part in the quote. It says when one with a binding vote weren't the only one to cast a close vote, then that one isn't pingable. If SE meant that the other two users aren't pingable, then the highlighted sentence isn't able to explain the clearly.
    – Tom
    Mar 27, 2020 at 11:39
  • @Tom, Sonic made the same observation; to which Shog9 replied, I have no additional information than what an employee has already written. Unrelated additional information: it has been Gold Hammered open.
    – Rob
    Mar 27, 2020 at 11:50
  • @Rob Mhh, Sonic just noted that the FAQ didn't include the limit on pinging a close voter like Shog mention in his answer. That FAQ was then updated to include that (the quote in this question is from that updated FAQ). So, also those descriptions say the same thing, "rob" shouldn't be pingable.
    – Tom
    Mar 27, 2020 at 12:07
  • @Tom, Shog9's advice was ignored; the FAQ was not changed to what he suggested. Definition: Unilateral.
    – Rob
    Mar 27, 2020 at 12:56
  • @Rob I read your linked 2018 answer by Shog9 (and the change to the FAQ made at that time) exactly opposite to your comment here. Also at this time the question has not been reopened; your link goes to a "leave closed" review. In conclusion, there are four lights.
    – rob
    Mar 27, 2020 at 14:24
  • 2
    @AdamLear Are there any news on this topic? Is this a bug or working as intended?
    – Tom
    Apr 3, 2020 at 9:31
  • 2
    @Tom AFAIK nothing has changed there... but I haven't had a chance to dig in. [status-review] means that this is at least somewhere on our radar, though.
    – Adam Lear StaffMod
    Apr 16, 2020 at 23:22
  • @AdamLear Thanks for the update. I thought "status review" actually means a staff member is currently working on it, thanks for the clarification.
    – Tom
    Apr 16, 2020 at 23:30
  • @Tom, see JNat's announcement for details.
    – Rob
    Apr 17, 2020 at 0:47

1 Answer 1

3

After a semantically unrelated issue came up, I ended up figuring out why this works the way it does.

Basically, the description in Shog's answer which is used as the source for the FAQ in question isn't quite correct. The actual way it works, according to this later comment from Shog in response to my query asking why it was designed the way it is, is that it looks at the one user ID associated with the "revision" that closed the question.

Therein lies the "bug": per my first link above and the question it links to, the user ID is filled in whenever a moderator closes a question or when there are no pending close votes from others - otherwise, user ID -1 (the Community user's ID) is filled in. This means that when a gold-badge non-moderator user bindingly closes a question when there are pending votes from others, the Community user is noted with the revision.

As to why it's implemented that way? Most likely, it was not updated when the ability for gold-badge users to bindingly close questions as duplicates was added - it's a relatively recent feature addition to the site. On the other hand, the check for a single non-moderator voter is probably there because in the earliest months of the ability to close questions, every vote was binding, and then when it was expanded to multiple-vote closure, it was then updated to take moderator binding votes into account. It seems that the goal of the algorithm was so a specific user wouldn't have to be chosen in the case of "community" closures, but again, this algorithm was likely not updated when more users gained binding votes.

This fact does lead to some edge cases, though. In particular, since the user ID is saved permanently, this means that whether notifications work is based on whether the user was a moderator at the time of closing. This may seem benign, but may lead to some weird cases. For example, when a moderator who would otherwise qualify to bindingly close a question through a gold badge does so, they're shown as having the gold badge in the post notice; however, internally, their moderator diamond takes precedence, so their ID is still noted in the revision history. If they then cease to be a moderator, then the notification will then appear to still work in a case where it shouldn't have. Also, if the final close voter where multiple were case wasn't a moderator at the time but later became one, they won't be eligible to be notified.

You must log in to answer this question.

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