43

I noticed this morning that posts that were closed with a gold badge dupe hammer have a new banner.

A couple of examples of the new look are below.

Is this a new change or just testing as I can't find a post on meta post about this?

What happens if the user doing the closing has a gold badge for multiple tags on the question?

migrated from meta.stackoverflow.com Oct 22 '15 at 9:17

This question came from our discussion, support, and feature requests site for professional and enthusiast programmers.

  • Yea, I noticed that too. Curious about the multiple badges case :-) – Cerbrus Oct 7 '15 at 12:41
  • 20
    Side note: as a moderator, I kinda want the badge qualification to take precedence over my diamond wherever applicable. I hardly ever mark questions as duplicates in tags where I don't have dupehammer abilities unless I'm responding to flags. – BoltClock's a Unicorn Oct 7 '15 at 12:56
  • @BoltClock unless I am mistaken I thought if a mod closed a question it always attaches the mod diamond as that supersede everything. – NathanOliver Oct 7 '15 at 13:05
  • 6
    Exactly - I'm saying I'd prefer if the badge qualification took precedence as I usually only dupehammer those questions anyway. – BoltClock's a Unicorn Oct 7 '15 at 13:08
  • 15
    @NathanOliver: What BoltClock is requesting: if(hasHammer()){ showBadge(); } else { showDiamond(); } – Cerbrus Oct 7 '15 at 13:32
  • 4
    I can't see the difference: those two images look the same! What is the difference? Where are the freehand red circles, people!?!? – Yakk Oct 9 '15 at 18:01
  • "If those answers do not fully address your question, please ask a new question." - and have it closed as a dupe again? – immibis Oct 10 '15 at 2:54
  • 3
    @Yakk It took me a while to figure this out, but I think those are both examples of the new banner. So there isn't any difference. – The Guy with The Hat Oct 10 '15 at 3:10
  • 1
    @Yakk Both images are the same as the older banners changed as well. I was just showing older closures and new closures have the same banner – NathanOliver Oct 12 '15 at 15:39
58

We have made two changes to closure by users that own a gold badge:

First, we now look at the question's current tags instead of its original tags, unless you were the one that edited them (either by doing it yourself, or by approving an edit from someone else).

This is done to address the request made in: Only prohibit those who edited the tags from using the dupe hammer.

Second, we made some UI changes to make this behavior more discoverable.

The old behavior could be confusing, both for users who didn't realize they had this newly-acquired power, and for users who didn't understand why it didn't work now and then (because the tags had been changed).

So, to make the dupehammer easier to understand, we made making two UI changes:

  • If you cast a close vote and that vote is binding because of a tag badge, this new styling gives you an explanation of why that vote was binding (it mentions the actual tag badge we used). The message also mentions reopening.
  • If you cast a close vote that would have been binding had you not changed the tags yourself, there'll be a message explaining that as well (instead of simply telling you "your vote has been recorded").

The new notice (in case the vote was binding) looks like this (the popup shows on hover):

new dupehammer notice

Note that this banner was not added on the first few uses of the hammer, which did not record any indication of the user having immediate closing powers. However, it was added to later uses which just had the old gold circle.

  • 4
    A: "Hey, B, can you add <tag> to <question>?" B: "Sure thing!" A: *Closes question* "Okay, you can remove it again!" B: "Sure thing!" <-- How do you guys prevent that from happening? :-) – Cerbrus Oct 8 '15 at 7:50
  • 3
    Also, if you edit tags on a question, without touching the one you have a gold badge on, can you still hammer it? – Cerbrus Oct 8 '15 at 7:53
  • 24
    @Cerbrus we don't prevent it. We don't think we need to, either, for two reasons. First, gold badge users have proven responsible so far (there's a lot more context in that meta post the request came from) and that abuse is pretty easy to detect anyway. Second, users can already band to close a question (with votes), we're not sure "one gold badger + one user with full edit privileges" is that much easier to come by than "five users with close vote privileges". – Thomas Orozco Oct 8 '15 at 7:55
  • 2
    @Cerbrus On that second point, we're not sure yet, but this will be clarified in the meta post. Thanks! – Thomas Orozco Oct 8 '15 at 7:56
  • 2
    If you edit out the tag(s) that would give you hammer privileges on a question, can you still hammer it? q.v. meta.stackoverflow.com/questions/255936/… – zwol Oct 9 '15 at 5:55
  • @zwol, if the question is worth closing, why would you edit it? – zzzzBov Oct 9 '15 at 18:18
  • 1
    @zzzzBov In the example I gave in the link, I knew that the question was not a C question, and I was uncertain about whether it should be closed. Consensus over there was "well, don't vote to close if you're not sure", but I still think disabling the hammer if you just edited the relevant tags out of the question would eliminate a foot-gun, at least. – zwol Oct 9 '15 at 18:44
  • @Cerbrus Update: Tags you added are simply ignored. E.g. if you add a tag you have a gold badge for, but there's another tag you have a gold badge for, we'll use the latter. – Thomas Orozco Oct 22 '15 at 9:20
  • That makes sense. And tags that someone else added can be used to close a question, right? – Cerbrus Oct 22 '15 at 9:22
  • @Cerbrus Yes, that's the essence of this change : ) – Thomas Orozco Oct 22 '15 at 9:24
  • @zzzzBov Duplicates are a special kind of closure. Duplicates tend to be good-ish. It (can) mean that two people had the same question but the latter's search terms didn't result in the former's question. Now both search terms should provide a question which should both point to the same set of answers. And with that in mind, it makes sense for duplicates to have the right tags at least (and non-terrible titles). – nhgrif Dec 3 '15 at 1:26
  • "we now look at the question's current tags instead of its original tags, unless you were the one that edited them (either by doing it yourself, or by approving an edit from someone else)" - what about if you were the one that put them in because you asked the question? (It's rare to see someone dupehammer their own question, but I have seen it happen.) – Rand al'Thor Jul 12 '16 at 21:09
  • It just happened to me: I dupe hammered a question tagged with Unix. I retagged from unix to bash and now I cannot single-handedly reopen it, even though I also have Bash gold badge. – fedorqui Nov 30 '16 at 15:29
23

I noticed this too. Beyond the fact that it took me a minute or two to realise that the explanatory description is now on-click rather than on-hover, which is mostly my own fault and I can get over, my one note-worthy issue with the change is that it changes what it means to click on a tag.

That is, whereas everywhere else on the site (AFAIK?) clicking on a tag span (with its particular, identifiable styling) takes you to questions about that tag, in this one scenario it leads to a pop-up about the dupehammer. That's not a consistent and userproof UI.

  • 22
    Agreed; easy enough to fix, though: just add a handle! – Josh Caswell Oct 7 '15 at 23:29
  • @JoshCaswell that was a good one. It totally needs to be implemented – NathanOliver Oct 8 '15 at 0:18
  • Agreed. IMO it should go back to on-hover and leave clicking a tag to go to the tagged questions. – eis Oct 8 '15 at 5:25
  • 1
    Hey, thanks for the feedback! We're going to check if we can restore the on-hover effect. However, I'm not sure what you mean by "clicking this identifiable styling takes you to questions about that tag"? I can see two places where it's used (badge tracker and badges list) where that's not the case. Mind pointing me to what you had in mind? Did you mean that it's simply too similar to the blue-and-squary tag style? Thanks! – Thomas Orozco Oct 8 '15 at 7:39
  • @ThomasOrozco: Hmm fair point you're right. Then I'll fall back on the latter :) – Lightness Races in Orbit Oct 8 '15 at 9:35
  • 2
    @LightnessRacesinOrbit Thanks for confirming! I think it'd make sense to send users to that badge page if we can have the popup on hover and update the copy there (to be clear: the former is an if, the latter is obviously do-able). – Thomas Orozco Oct 8 '15 at 9:36
  • Fan-art: dupehammer with a handle github.com/AlexanderSharykin/StackOverfortress/blob/master/… – ASh Dec 21 '15 at 9:06
17

If you have multiple gold badges in common with the question's tags, it chooses one. Here, I have both the sass and css badge, but it chose the css tag for the closure. Not sure if it goes by the order the tags appear in or the age of the badge when deciding (I got the css badge first).

enter image description here

  • Me, I'd say it should go by the rarest tag badge first, but that may not be practical to precompute. – Nathan Tuggy Oct 7 '15 at 22:09
  • 9
    I think it's more likely a case of doing a SELECT ... LIMIT 1. – cimmanon Oct 7 '15 at 22:57
  • 2
    @cimmanon negative, we do not use MySQL :) – Nick Craver Oct 7 '15 at 23:34
  • 1
    @NickCraver MySQL is not the only DBMS that uses LIMIT. – cimmanon Oct 7 '15 at 23:44
  • 3
    @Nick: You store JSON within a field inside a relational database??? :( – Lightness Races in Orbit Oct 8 '15 at 0:18
  • 12
    { "comment": { "inreplyto": "@Lightness Races in Orbit", "body": "JSON is the future, embrace it ヽ༼ຈل͜ຈ༽ノ" } } – BoltClock's a Unicorn Oct 8 '15 at 5:08
  • 1
    @ThomasOrozco: What criteria is that selection based on, though? First earned badge? Highest rep one? – Cerbrus Oct 8 '15 at 7:47
  • 5
    It should take the most prominent tag; this morning, I saw a question tagged with “python” and “list” being closed with the “list” badge when it should be the “python” badge instead (imo). – poke Oct 8 '15 at 8:48
  • 4
    @LightnessRacesinOrbit Absolutely. Everyone who says you can't do that (almost always with no further reasoning) thinks I'm insane for using JSON. On the other hand, I think they're insane for not using the best tool for a given job. Why limit yourself because of preconceived notions? We handle 5.8 billion HTTP requests issuing 19 billion SQL queries a month (at a ~5% CPU load on SQL), trust that we have a bit of experience at optimization here :) – Nick Craver Oct 8 '15 at 10:30
  • 2
    @Nick: I don't think accusing normalisation fans of scoffing at JSON-packing an RBDMS field of doing so only due to cargo culting and "preconceived notions" is a valid argument ... but invoking your experience of scale data probably is. :P – Lightness Races in Orbit Oct 8 '15 at 10:31
  • 2
    @LightnessRacesinOrbit think of it at this high level: table joins inside SQL have scaling limits. Deserializing that JSON elsewhere does not. We can shift the load to the web tier. If we don't need to query it (key factor), we're safe to use JSON. SQL 2016 adds support for it even, but even that's minimal and we'd only be using that functionality in migration scenarios. It's just about being practical :) – Nick Craver Oct 8 '15 at 10:40
  • 3
    @Lightness Races in Orbit: Or, for those who don't speak hashtag, it's "👍" – BoltClock's a Unicorn Oct 8 '15 at 13:03
  • 3
    @BoltClock: Or, for those who only speak HTTP status codes, it's "200 OK" – Cerbrus Oct 8 '15 at 14:19
  • 1
    @poke ... someone has a gold badge in the list tag? And that's enough to get a dupehammer?! I feel another meta post brewing. – Gus Oct 8 '15 at 18:57
  • 1
    I guess I always assumed the dupehammer was only applicable to tags associated with a language or framework. I guess >1000 rep in >200 questions is a high enough bar that we don't really care if it's in a ridiculous tag like [String] or [arrays] – Gus Oct 8 '15 at 19:04

You must log in to answer this question.

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