What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

I see that the "close as duplicate" feature continues to undergo changes. Unfortunately, I must report negatively on this newest one.

This time, it's that a banner is added to a question as soon as the very first dup-vote is cast:

a banner added to questions as soon as the first dup-vote is cast

This just seems:

  1. Ripe for Abuse: you can vote for absolutely anything and get it added to the question in a manner that looks official;

  2. You get an official-looking sanction for a dup-vote, regardless of the fact that you haven't yet attracted enough agreement to meet the requirements to actually close the thing.

A single duplicate vote is hardly worthy of such a prominent display — if a question is truly a duplicate, worry not! Your excellent high-reputation regulars, even though you don't seem to trust us nowadays, are more than capable of identifying a true duplicate and closing it in short order. There's really no reason to undermine our community-powered discipline by advertising the initial vote in this fashion.

I would ask why so many changes are being applied without community input? And, if you need a non-OCD practical question, what problem does this change solve?


Consider this situation:

  1. Question X becomes hot due to Reddit or the Multicollider.
  2. Bob sees question X but wants attention for his (unrelated) question Y.
  3. Bob casts dupe close-vote on X with destination Y.
  4. Much of the traffic pouring into X is now redirected to (unrelated) question Y.
  5. Tons of nasty comments from confused users over why the banner is there for an unrelated question Y.

Yes, this type of abuse would warrant severe moderator consequences. But unlike vandalism edits, this banner cannot be removed (by normal users).

So you're allowing a single user to do so much damage that cannot be reversed without a moderator.


And here's the other (very common) example pulled from the comments:

  1. Bob sees a question. He (honestly) thinks it's a dupe.
  2. He casts a dupe-vote.
  3. 5 other uses point out that it is not a dupe.
  4. Bob agrees. But he can't do anything, other than to delete the auto-inserted comment.

The banner is still there. Left to mislead anyone who visits the question.
For how long? Until the vote ages away (5 days)?

share|improve this question
    
For the record, I invited @Mysticial to make this edit because his title is better than mine. :D –  Lightness Races in Orbit Feb 6 '13 at 5:52
5  
As a demonstration, I've "abused" this "feature" to draw traffic from this question: meta.stackexchange.com/questions/166428/… –  Mysticial Feb 6 '13 at 6:05
3  
For 1, I say "wait and see if it actually becomes an issue". For 2, I partially agree, but I also like the big freaking notice that may just dissuade superfluous answers on a true duplicate. Perhaps the threshhold for the banner can be bumped up to 3 close votes for (the same?) duplicate. –  Josh Caswell Feb 6 '13 at 6:16
5  
The other (more likely) scenario that I didn't mention is that the dupe-vote could have been an accident. In the past, if you mess up, you can just delete that auto-inserted comment. Now, if you mess up, that banner sticks for (5 days)? –  Mysticial Feb 6 '13 at 6:17
4  
@Mysticial: Yes, in fact, I just now successfully voted to dupe-close a question with a bad dupe. This is entirely my fault, and I am chagrined. There are, however, other good duplicate candidates. Until a few hours ago I could correct the dupe banner, and now I have only a few bad choices: edit the other links in anyways, leave a comment, or try to get the question re-opened and then re-closed. –  Josh Caswell Feb 6 '13 at 6:30
12  
The banner disappears as soon as there's a single reopen vote, too?! i.stack.imgur.com/EyVbq.png That's...equitable, I guess, but even more stupid. –  Josh Caswell Feb 6 '13 at 7:07
1  
How about I merge the two of you so you can share the rep amongst yourselves ;) (I can't; they removed that option for mods very recently.) –  BoltClock's a Unicorn Feb 6 '13 at 7:16
9  
I can't wait until I hit 100k rep so I can finally use the production version of Stack Overflow; having to use the test servers is a bit annoying. What I find strange is that this privilege isn't listed on the privileges page. –  R. Martinho Fernandes Feb 6 '13 at 7:16
1  
Related: Improve UI flow for "close as duplicate" I edited it a bit so it could serve as a possible improvement to the current process, partly addressing the issues raised here –  Pëkka Feb 6 '13 at 8:08
2  
The ultimate improvement to the dupe situation, IMO, is still Shog's suggestion about having dupe suggestions as a special kind of answers that can be upvoted and downvoted. I don't understand why they are not following up on that. Looks like the team are afraid to introduce big changes (I probably would be, too), but some tags are drowning in dupe crap. –  Pëkka Feb 6 '13 at 8:09
9  
IMO, once a dupe vote is cast, everyone with the privilege to cv should be forced to review the dupe before they can submit an answer. Users with not enough reputation should not be able to submit answers at all. This will educate people that we want them to act on dupe votes. –  Gordon Feb 6 '13 at 9:07
3  
@Gordon: forget dupes, I want that for all close reasons :) –  Manishearth Feb 6 '13 at 11:06
6  
@JoshCaswell "Stupid" is putting it a bit mildly. It's such a glaring design flaw. Question gets closed as dupe, a reopen vote comes along, and now the question will stay closed for all eternity -- because all records of which question it supposedly duplicates have just been wiped out. Hell, even the review queue doesn't tell you. –  lunboks Feb 6 '13 at 14:39
2  
Okay, so also, under-3k users apparently don't see the banner: meta.stackexchange.com/a/166545/159251. So my support for the banner is gone, because the only reason I liked it was to dissuade answers from low-rep users who couldn't close vote anyways. /cc @Mysticial –  Josh Caswell Feb 7 '13 at 7:30
1  
See also meta.stackexchange.com/questions/167503/… and meta.stackexchange.com/questions/915/…. This is ridiculous, AFAIC. –  Dave Newton Feb 21 '13 at 19:50

7 Answers 7

up vote 18 down vote accepted

See my answer - we're now only showing the pre-closure banner "This question may already have an answer here" to the question owner. No one else, not even mods, will see it until after the question is closed.

share|improve this answer

I think the banner has the most value for the asker themselves and anonymous/low rep users. The verbiage is great for that use (oh hey, I want this answered, click) but I can see how it needlessly enforces groupthink for those who can close.

If it turns out to be a problem a possible idea is to only display the banner on unclosed posts to low rep/anonymous users and the asker of the question. This way new users see "oh hey, my precious answers!" and users with close votes just see another question with a close vote, no banner. That way the workflow is the same for closing users (ideally the review/close dialog should give them the info they need to make the decision, not the banner) but you keep the benefit for the new users, which I assume this change is intended for.


Also this? (Image by Josh Caswell) That bug was fixed

share|improve this answer
    
Agreed with your first sentence. In case you haven't seen this, it looks like users without close votes don't see the banner. I don't understand that at all, unless it's also a bug. –  Josh Caswell Feb 7 '13 at 18:12
    
@JoshCaswell that's no good... –  Ben Brocka Feb 7 '13 at 19:10

what problem does this change solve?

Well, for one, it helps make it more obvious that there is a possible duplicate. Until now, dupe suggestions were often buried in long comment threads, and easy to overlook. I like this change.

Your excellent high-reputation regulars, even though you don't seem to trust us nowadays, are more than capable of identifying a true duplicate and closing it in short order.

Maybe, but many other users are not. The impulse to answer is stronger than the impulse to dupe-vote, and any education about a possible dupe is a good thing in my book.

I say give this feature a chance; increasing the threshold so the banner is displayed only after the second close vote would be an option, though.

Consider this situation:

Bob casts dupe close-vote on X with destination Y.

That would be massive abuse and should result in serious moderator consequences; I agree it should be possible for the community to roll back the banner, though. The most obvious way for this would be the ability to cast pre-emptive close votes.

share|improve this answer
5  
But this doesn't at all address the issues with the change. –  Mysticial Feb 6 '13 at 7:40
    
2nd close vote sounds fair to me. But there needs to be a way to remove/edit the banner once it appears. –  Mysticial Feb 6 '13 at 7:41
    
@Mysticial ok, but who should be able to remove it? 3k+ users? 5k+? The OP? The latter would defeat the point IMO. –  Pëkka Feb 6 '13 at 7:45
1  
Well, as it is right now, I don't see a good solution other than to rollback to previous behavior. If there was a way to "downvote" the banner or cast pre-emptive re-open votes, then it could work. But this seems to be over-complicating the issue. –  Mysticial Feb 6 '13 at 7:47
2  
+1 for doing it on the second vote. That at least is a sign that there is some consensus the question is out of place. And I have seen close vote abuse from at least two different users on a beta site before. Even with a suspension, their votes didn't get revoked/age for a long time. –  Troyen Feb 6 '13 at 8:21
2  
close (1) is as visible as necessary IMO. –  Matthew Read Feb 6 '13 at 16:58
1  
Y'know, it's not just that the impulse to answer is stronger than the impulse to close as a dupe. In some cases I've posted an answer to something I strongly suspected was a dupe simply because I could write a complete answer off the top of my head in less time than it would have taken to find the original question. Lazy? Sure, but if the choice was between dupe-hunting and ignoring it entirely, I'd probably do the latter, which doesn't help anyone. –  McCannot Feb 7 '13 at 17:42

Before the "close as duplicate" behavior we all know and love was introduced, there was a semi-standard practice among those of us closing duplicates: edit in a link at the top when you vote to close.

This was nice, primarily because that way you could be sure a link existed (the system didn't do any enforcement, so a lot duplicates at one time didn't have any links to the "original", or if they did they were in comments).

It also had the advantage of putting the link front and center: if you were the author of the question, or were thinking about answering it, you really couldn't miss it.

This latter advantage is finally captured by the new behavior.

What the new behavior doesn't capture was the ability of the OP to dispute closure by simply rolling back the edit. This sort of edit is discouraged for closed questions (and there have been several suggestions to make it impossible), but when the edit was made prior to closure it was a pretty powerful way for the asker to make it clear that the question was not a duplicate.

With the behavior that existed before today, the OP could reply publicly to the automatically-generated comment, which was somewhat less emphatic but serviceable. However, there's no way to publicly disagree with a system banner.

The other potential issue I see with this is that the ability to edit in additional duplicate links (or edit out bad ones) is lost. It may be necessary to add explicit support for some sort of modification to these lists, if simply accumulating the various links from close votes does not suffice.


rm -rf had another idea for enhancing this... Allow the question asker to confirm the duplicate as helpful, and instantly close the question.

...This brought to mind Pekka's suggestion for enhancing the dup-closing UI - almost all of the pieces are now in place:

  • Potential duplicates are shown to voters immediately
  • The new close dialog has a fairly decent in-place preview functionality, with the ability to also go back and select a different question.

...we just need a button on the former that takes you straight to the latter!

The only thing this doesn't solve is bogus links. There needs to be a way to get rid of them, even if the question is closed, even if other, better links are also included.

share|improve this answer
5  
"Allow the question asker to confirm the duplicate as helpful, and instantly close the question." I can name at least 5 times when I would find that useful –  Bartek Banachewicz Feb 7 '13 at 1:41
1  
The bogus links could decay like flags. If enough people think 'that's bogus', then it goes poof. But that's a lot of buttons and knobs in something that we're trying to make simpler already. –  Tim Post Feb 7 '13 at 1:48
    
@TimPost: Instead of adding a bogus link in the main linkbar, add it to the dupe menu. Then you can click a "this is not a dupe" button somewhere there. A single "this is not a dupe" vote will remove the temporary banner until/unless it gets closed. –  Manishearth Feb 7 '13 at 7:38

I think the key problem in this case is the inability of community members to remove the notice by themselves.

If a community member is given a way to reject what's said in the banner (without even necessarily rejecting the vote itself), the problem will be solved in its entirely.

A user won't be able to simply vote on a reddited question, and have all traffic being redirected to his question of interest, he would, but for a very short time, and we all know you can't CV twice on the same question.

So yeah, the key solution in my eyes is to allow community members to reject the suggested duplicate.

share|improve this answer

I guess the discrepancy comes from the level of activity on the site. Not all sites get questions closed in a jiffy. Some of them take time. I personally like this feature. Why? Because I participate on smaller sites (Physics and Chemistry). This would be perfect there, and the greater visibility would help root out non-dupes. On these sites, I trust the 3k/500 rep community not to abuse this (besides, it would easily be caught).

If this is a matter of scale, then, for sites like SO and other high-volume sites, let this kick in at 2 or 3 votes (similar to how suggested edit approvals work for different sites). Simple.

The main advantage of this feature is that it prevents answers from being scattered. Many a time I've seen questions closed as dupes which have answers, simple because the answerer didn't pay much attention to the comments. Such posts can't always be merged (merging is when the questions are pretty much identical), and you end up with the answers being scattered. This feature mitigates the problem.

  • Ripe for Abuse: you can vote for absolutely anything and get it added to the question in a manner that looks official;

  • You get an official-looking sanction for a dup-vote, regardless of the fact that you haven't yet attracted enough agreement to meet the requirements to actually close the thing.

I personally don't see why it looks more official -- this is only because you're used to the old display (and what's wrong with it being official-looking anyway). It just says "This question may already have an answer". A perfectly legit thing to add -- if one user feels that the question is a dupe, it may indeed have some answers there, with a high probability.


I would ask why so many changes are being applied without community input? And, if you need a non-OCD practical question, what problem does this change solve?

Regarding the changes, again, let me mention that this is SE's way of testing stuff. They may remove a feature if the community doesn't like it. They do this all the time, you've just noticed it for the first time. Most of these changes go unnoticed. So I really wouldn't call them out for not asking the community. Aside from that, they have been consulting a few community members, just not in public (I don't know much about this).


An afterthought: If the question gets a "leave open" vote in the queue, then no header magic should be applied till the question gets closed. Otherwise, this header will stay for all eternity on a q with a lone CV. Even better, have this/or this in place. Give users a "this is not a dupe" button in the close menu (next to the "exact duplicate" option) that acts as a preemptive reopen vote (or a negative close vote), and removes the banner. Basically, the banner should get removed whenever the system realizes that there is disagreement over it being a dupe.

And removing the banner when there's a reopen vote is just silly. That just creates broken windows, and the folks trying to reopen have to jump through extra hoops (unless they come from the queue)

share|improve this answer
1  
At the least it needs to be two votes (if not three), even on beta sites. I've seen a user go through a bunch of old questions and revenge-vote them as "too localized". Nobody else agreed, but they also didn't age very quickly being on a smaller site. Two close votes is a more sure sign that something is up with the question. –  Troyen Feb 6 '13 at 8:18
    
@Troyen: On beta sites, aside from the community being generally nicer, there is more visibility of such things. I mod a beta site, and I check the review queues as well as any new stuff (edits/etc) on the front page (other mods and users do the same). Such things get caught very easily. –  Manishearth Feb 6 '13 at 8:28
    
Regarding revoking/aging, that ought to be fixed with my afterthought^ –  Manishearth Feb 6 '13 at 8:28
1  
They do this all the time, you've just noticed it for the first time It's been more lately. –  Lightness Races in Orbit Feb 6 '13 at 11:00
    
@LightnessRacesinOrbit: No, it's not. They test small tweaks every now and then on SO/SU and/or MSO users. These aren't really small, but they've done that. They just don't get noticed much, since they are in corners of the site. Regarding the current stream of new closing related features, has it occurred to you that it's a part of a much larger feature that's being planned? –  Manishearth Feb 6 '13 at 11:08
    
Anyway, no point arguing on how often they live test features on us. The point is that they do that, and there's nothing wrong with doing it without community input. –  Manishearth Feb 6 '13 at 11:14
3  
If you're going to incrementally test significant functionality changes, live, on your community of software developers, you might as well ask for their input while you're at it. And, yes, I know they have always tested small tweaks every now and then -- but I'm not talking about small tweaks. –  Lightness Races in Orbit Feb 6 '13 at 16:42
    
@LightnessRacesinOrbit: Again, some of this seems to have been discussed with a small part of the community beforehand. Look, if they discussed this with the public beforehand, there would be dissent and they'd never take off. This means that a lot of advantages/disadvantages would never be found out (there would just be speculation). Simple philosophy: you never truly know how something will turn out unless you try it. –  Manishearth Feb 6 '13 at 18:49
3  
I guess it's a good thing that a) we're not their customers and b) they have no real competition, because I'm finding myself agreeing, at least in principle, with @LightnessRacesinOrbit here: doing your hallway testing in the live version of your product sounds like a good way to send people running for the exit. –  Josh Caswell Feb 6 '13 at 19:36
    
@JoshCaswell: Gmail does it. Facebook does it. Google+ does it. Wikipedia does it. If you look at it, it's muuuch less annoying if SE does it simply because there is a much better feedback system (namely MSO). Google/FB announce a new feature and that's pretty much it. Folks rant, and sometimes their rants get heard. SE launches a new feature, and then there's a weekish of listening to good feedback on it. And this isn't "hallway testing". They aren't testing buggy code. They're testing a concept, which seems OK to me. –  Manishearth Feb 6 '13 at 19:40
    
A note: while I've been calling this a "test", it isn't a test in the strictest sense. If there isn't much negative feedback, they will keep the feature in status quo. If there is, they shall listen and tweak (I hope). There is negative feedback right now, so I guess they may tweak it. –  Manishearth Feb 6 '13 at 19:42
2  
"They aren't testing buggy code." Hallway testing doesn't imply bugs; it means that you're having a sample user investigate a feature without having announced it or otherwise prepped her beforehand. And note that for all the companies you've named, my two bullets also apply (well, Gmail sort of has alternatives). Plus, seriously, "Everyone else is doing, why can't we"? –  Josh Caswell Feb 6 '13 at 20:25
    
@JoshCaswell: I still don't see the problem with such testing. They put forth a feature, and wait for user input. They get user input. Isn't the process working? They haven't made any huge changes. They've made a medium change that won't have a permanent effect if removed. IMO, that's perfectly acceptable to try out. Also, I'd prefer to stop arguing about this, it's rather pointless. I'd just made the remark about testing because to me, Lightness looked like she was near ranting (for the second time). I just wanted to address that a bit :/ –  Manishearth Feb 6 '13 at 20:54
    
Sure, argument over. (Aside: Lightness is a he; formerly went here by his real name, Tomalak Geret'kal (see blog link in his profile). The profile pic is from a sci-fi TV show, I think.) –  Josh Caswell Feb 6 '13 at 21:29
3  
@Manishearth I think the overhaul of the review system is a good counterexample. They proposed a new system on MSO, got overwhelming positive feedback, and were able to iron out issues before putting it live on SO. Seems like they're doing an overhaul of duplicates/closing questions, but incrementally and directly on the live site. –  Troyen Feb 7 '13 at 2:03

I believe this banner is very helpful and it is great it is now shown before the question is actually closed - this way users can make use of the answers in the original post even before the question is closed.

Still I think there is a place for error here. As OP states this feature leaves a lot of place for abuse which is definitely not good. Even if there is no abuse, a user may cause this banner to appear by mistake. It happened to me recently that I erroneously marked a question as duplicate. Now I really want to remove my vote but what is more important I want to remove the banner as it is very misleading.

The solution for me is to enable some users(as per the community policy probably users with high enough reputation) to remove the banner. This way the banners will still be helpful, but in case of abuse or error, moderators or experienced users will be able to fix that.

share|improve this answer
    
this way users can make use of the answers in the original post even before the question is closed. They already can. The "Related" sidebar exists. I realise that most people have trained themselves to ignore it, but this new feature only enforces that self-training. –  Lightness Races in Orbit Feb 7 '13 at 9:39
    
@LightnessRacesinOrbit related side bar exists even if a question is not flagged as duplicate. Also quality of the banner is supposed to be better as it is result of a human interaction, actually a user flagged a question as duplicate(i.e. related). –  izomorphius Feb 7 '13 at 9:41
    
Regardless, usually that user was able to simply pick the first "Related" result. The "Related" sidebar is already a very good indicator. –  Lightness Races in Orbit Feb 7 '13 at 9:43

You must log in to answer this question.

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