Danbooru

Baffling...keeping dupes?

Posted under General

This topic has been locked.

Why are yall keeping dupes? I tried reporting one and was basically told not to.

So if I just take an image and downscale it like 10 times into 10 different versions and post them all you guys would just keep them? How does that make sense?

Oh and the mod who talked to me about it was less than helpful. You know if you're tired and jaded on a certain subject then make a copypasta that actually explains why a policy exists and just paste it into messages...rather than wasting more time being like "it's our policy", which is a shitty non-answer that only prompts more questions. That's what a smart person would do.

Updated by Hillside Moose

Dyrone said:

Why are yall keeping dupes? I tried reporting one and was basically told not to.

So even though the post flagging dialog explicitly states that this isn't a valid reason to flag a post, you went and did it anyway? Dude...

So if I just take an image and downscale it like 10 times into 10 different versions and post them all you guys would just keep them?

They'd probably just get automatically deleted after sitting in the modqueue for a few days. And then you'd just get banned like others who have tried pulling this sort of stunt.

Oh and the mod who talked to me about it was less than helpful. You know if you're tired and jaded on a certain subject then make a copypasta that actually explains why a policy exists and just paste it into messages...rather than wasting more time being like "it's our policy", which is a shitty non-answer that only prompts more questions. That's what a smart person would do.

A smart person would also search the forum for older threads about the very same subject instead of re-opening the discussion yet again. This has been covered time after time after time. TL;DR: there's no harm in keeping duplicates, and if we start deleting them, then we also lose associated information like favorites and comments.

Most comments aren't worth keeping and favorites could be moved to a parent post. As for harm, assuming we're talking about images that are otherwise identical aside from resolution, it creates bloat in search results. I don't think anyone wants to find the same image 10 times if one is, or even a few are, objectively superior. Seems like a simple argument to make, but if the current consensus is to keep them I don't mind.

As for the OP: if you just want to complain about stuff to people who aren't really responsible for how things are then don't be surprised if you get sassed back, especially if you are as obnoxious as in the OP. No one is volunteering their time to listen to people whine.

I think another issue is one similar to what we've had with the post replace feature, in that we could have revisions or other worthwhile differences accidentally deleted. Not as big of an issue as in that case (as they would still be accessible and not eventually gone forever), but tends to be something I think about as for why we keep dupes. We'd have to check every post being flagged as a dupe to see if it actually is a dupe, which is more (IMO) unnecessary work.

EB said:

I think another issue is one similar to what we've had with the post replace feature, in that we could have revisions or other worthwhile differences accidentally deleted. Not as big of an issue as in that case (as they would still be accessible and not eventually gone forever), but tends to be something I think about as for why we keep dupes. We'd have to check every post being flagged as a dupe to see if it actually is a dupe, which is more (IMO) unnecessary work.

So much this.
That's also the reason why only moderators are actually allowed to replace posts because approvers (and these users are trusted) misused that feature for their own good.

Rampardos said:

Most comments aren't worth keeping and favorites could be moved to a parent post. As for harm, assuming we're talking about images that are otherwise identical aside from resolution, it creates bloat in search results. I don't think anyone wants to find the same image 10 times if one is, or even a few are, objectively superior. Seems like a simple argument to make, but if the current consensus is to keep them I don't mind.

[...]

There is the duplicate tag that we can blacklist. Though we would have the same problem : lot of work to add the tag to each picture.

iridescent_slime said:
This has been covered time after time after time.

LOL Those threads you linked are 9 years old, 11 years old, and 11 years old respectively. So basically you're kinda proving that the topic hasn't been covered recently. Also if I commented in any of those threads you'd probably complain that I was necroing. Can't win with a person like you.

iridescent_slime said:
and if we start deleting them, then we also lose associated information like favorites and comments

So? Comments are not valuable at all, neither are upvotes. In fact having 2 versions of the same image simply splits the upvotes over time so you'll NEVER get the actual tally in the end. Can't believe you guys are saving dupes for comments and upvotes.

Rampardos said:
if you just want to complain about stuff to people who aren't really responsible for how things are then don't be surprised if you get sassed back

There wasn't really any sass, still the conversation dragged over many messages because I didn't get an up-front answer. Would be much simpler to proffer that answer from the beginning rather than being glib about it....all I'm saying. Would save the mod time in the end.

EB said:
We'd have to check every post being flagged as a dupe to see if it actually is a dupe, which is more (IMO) unnecessary work.

If you're already adding the tag "duplicate" and giving dupes a parent/child relationship then you're already doing the same amount of work it would take to identify which dupe is the superior image, which is quite easy in 99% of cases. All I'm hearing is "we're lazy" and "this is the way it's always been". Not good answers.

OP is not wrong. Duplicates cause numerous obvious problems and the objections to deleting them are poor and outdated. And he's not wrong either that we do a poor job of documenting our policies.

(I should clarify that I'm talking about third-party duplicates, not Twitter->Pixiv first-party duplicates. Those are also problematic but that's a whole other topic.)

I agree that there is no real reason to keep third-party duplicates since we prefer first-party sourced images. First-party duplicates, as evazion mentioned, is a whole 'nother can of worms though. Also compared to ~9 years ago we now have the option to move favs to parent posts when deleting images so that point is moot, upvotes and comments don't get moved however.

Dyrone said:

LOL Those threads you linked are 9 years old, 11 years old, and 11 years old respectively. So basically you're kinda proving that the topic hasn't been covered recently. Also if I commented in any of those threads you'd probably complain that I was necroing. Can't win with a person like you.

So? Comments are not valuable at all, neither are upvotes. In fact having 2 versions of the same image simply splits the upvotes over time so you'll NEVER get the actual tally in the end. Can't believe you guys are saving dupes for comments and upvotes.

There wasn't really any sass, still the conversation dragged over many messages because I didn't get an up-front answer. Would be much simpler to proffer that answer from the beginning rather than being glib about it....all I'm saying. Would save the mod time in the end.

If you're already adding the tag "duplicate" and giving dupes a parent/child relationship then you're already doing the same amount of work it would take to identify which dupe is the superior image, which is quite easy in 99% of cases. All I'm hearing is "we're lazy" and "this is the way it's always been". Not good answers.

Instead of insulting everyone, please tell us about how you think we should take care of this "problem".

seika0 said:
Instead of insulting everyone, please tell us about how you think we should take care of this "problem".

Umm...start deleting dupes? ¯\_(ツ)_/¯ Also love how you put problem in quotes...it is a problem as your compatriots have already admitted.

evazion said:
I should clarify that I'm talking about third-party duplicates, not Twitter->Pixiv first-party duplicates. Those are also problematic but that's a whole other topic.

In that case you would just delete the Twitter image in favor of the Pixiv...as in most cases the Twitter image is going to be vastly inferior.

Dyrone said:

In that case you would just delete the Twitter image in favor of the Pixiv...as in most cases the Twitter image is going to be vastly inferior.

There's a whole bunch of issues that makes it much more complicated than just "delete all inferior Twitter uploads".

Unbreakable said:
There's a whole bunch of issues that makes it much more complicated than just "delete all inferior Twitter uploads".

Care to explain? I actually know a lot about this subject and believe it or not...it's not all that complicated. Of course the images would have to be evaluated one by one, but beyond that...it's not hard to tell which image is better.

Updated

Here are all of the reasons I can think of against deleting objectively inferior duplicate posts.

—Dupes often have different commentary to the superior versions. For example, I often see Twitter compilations posted to Pixiv where the commentary is just "Here are my posts from Twitter over the last week/month!". Meanwhile, the commentary on each individual Twitter post is much more specific to the individual pieces of art.

—Favorites (and to a lesser extent, score) would need to be merged to the superior version. If I actually used favorites, I would be pretty annoyed if suddenly I have all these deleted posts in that list that I needed to go through and change over.

—Believe it or not, comments do actually have some value most of the time. Examples include (but aren't limited to) giving context to memes that the artist didn't give in the commentary and actual discussions on current events. Not every comment is a cringe-worthy roleplay about the pleasure of being cummed inside.

—It would require a massive effort to go back and delete the large number of dupes that already exist.

—Innocent users' upload limits could be destroyed. Additionally, it would artificially inflate some long-time uploaders' deletion counts, making them look like worse users than they really are.

—It might become less obvious which users are actually doing the work. Suddenly deleting duplicates could inspire a culture of 1-uppers who only copy tags from Twitter posts to their own superior Pixiv uploads, which in turn makes them look like a model user despite having put next to no effort into contributing to the site.

—For some artists, paid rewards would trump publicly released versions in terms of which is superior. I don't know about other users but to me making a paid reward the only active post seems like a shitty thing to do - that, however, might be another discussion entirely.
A less morally-rooted problem is the potential for a culture of piracy to rise up, where posting superior paid rewards whenever possible to nab the superior upload becomes the norm. This could have the same problem I talked about in my previous point, but may also result in a lot more banned artists due to Danbooru suddenly showcasing their paid rewards much more blatantly.

Now, despite all of these problems I'm actually in support of deleting inferior versions (except when it comes to paid rewards) - however, rather than deleting them perhaps a new post status should be created for inferior duplicates.
It would be similar to status:deleted or status:banned in that it would hide posts from search results, but different in that it would not affect a user's upload limit or increase their deleted posts count. Perhaps it could also avoid showing a green border on the superior post in searches, but still show any inferior versions when actually on the post page itself (this would still allow users to check for any commentary/comments that inferior posts may have).

AngryZapdos said:

...

These are all good points and I can't find anything to really disagree with in there. It's true that duplicates (as in first-party, twitter/pixiv duplicates for example) needlessly bloat out search results, but at the same time axing them entirely brings about a whole can of problems.
A new status specifically for them (such as status:duplicate) that doesn't appear in search results by default (users could opt-in to make them appear as they currently do or somesuch) but does not count in deletion ratios and appears in other places would be good.

Dyrone said:

Care to explain? I actually know a lot about this subject and believe it or not...it's not all that complicated. Of course the images would have to be evaluated one by one, but beyond that...it's not hard to tell which image is better.

See AngryZapdos' reply, it pretty much sums it up. I have around 1500 Twitter uploads that also has a parent post, suddenly gaining 1500 deleted posts is not something I'd like.

Unbreakable said:

See AngryZapdos' reply, it pretty much sums it up. I have around 1500 Twitter uploads that also has a parent post, suddenly gaining 1500 deleted posts is not something I'd like.

Does it matter what you like :3?

But I also agree with him.
Deleted posts should be an indicator for bad uploads. Duplicates are not bad uploads unless you upload the duplicate afterwards. You also get banned if you do that excessively. There was a ban just like that a few weeks ago.
However, if duplicates are counted as deleted uploads then this sheds a wrong light onto the user. Deleted posts should only be posts that have bad anatomy or have gotten flagged for a good reason.

AngryZapdos said:
Favorites (and to a lesser extent, score) would need to be merged to the superior version. If I actually used favorites, I would be pretty annoyed if suddenly I have all these deleted posts in that list that I needed to go through and change over.

I was told there was a tool for merging favorites. I don't believe score is very important...the score has already been corrupted by the mere fact that there are 2 or more version of the image. The actual score the image would have received if it hat been uploaded as one version is unknowable...simply merging the two scores might actually artificially inflate the score higher than it would have been had it been just a single image. However if there was such a tool developed I wouldn't be opposed to such a thing...I just don't know how willing/easily you guys are able to implement new features.

AngryZapdos said:
Believe it or not, comments do actually have some value most of the time. Examples include (but aren't limited to) giving context to memes that the artist didn't give in the commentary and actual discussions on current events. Not every comment is a cringe-worthy roleplay about the pleasure of being cummed inside.

That's true...but I would say that's probably like .001% of comments are valuable. Very rare. If there is a valuable piece of information then presumably whoever is evaluating the dupe could take it over to the better post.

AngryZapdos said:
It would require a massive effort to go back and delete the large number of dupes that already exist.

Yeah but it doesn't have to be done overnight...simply chipping away at it over time would be fine. The important part is changing the policy now so that the problem doesn't just keep growing larger and larger.

AngryZapdos said:
Innocent users' upload limits could be destroyed. Additionally, it would artificially inflate some long-time uploaders' deletion counts, making them look like worse users than they really are.

That would require some new tech so that one could delete an image without penalizing users.

AngryZapdos said:
It might become less obvious which users are actually doing the work. Suddenly deleting duplicates could inspire a culture of 1-uppers who only copy tags from Twitter posts to their own superior Pixiv uploads, which in turn makes them look like a model user despite having put next to no effort into contributing to the site.

Honestly...I think the whole Twitter gold rush that's going on should be curtailed in some way. Too many people are posting Twitter images knowing full well that the artist will upload to Pixiv or Newgrounds very soon...they don't care they just want to be the FIRST so they're willing to upload a vastly inferior version to do that. It's also very harmful as sometimes people simply see the Twitter version and figure "meh, guess that one already got uploaded...I'll just move on then." and the HQ version NEVER gets uploaded. It's very damaging to rush to upload the Twitter versions. Whether you're deleting the dupes or just flagging and parenting the dupes...it creates needless work that could easily be solved if people just wait like...a day in a lot of cases.

I think if you post a Twitter pic and a better version is available and uploaded within 3 days, you should be penalized for that. Stop rushing to Twitter when you know the artists are just going to upload a better version to Pixiv like 2 seconds later.

Dyrone said:

I think if you post a Twitter pic and a better version is available and uploaded within 3 days, you should be penalized for that. Stop rushing to Twitter when you know the artists are just going to upload a better version to Pixiv like 2 seconds later.

You don't always know that.

Dyrone said:

The important part is changing the policy now so that the problem doesn't just keep growing larger and larger.

You keep maintaining that there is a problem, but you have yet to explain exactly why the existence of duplicate images is a problem.

Because it splits up the scores? It can't be that; you say scores aren't important.

Because it adds a bunch of identical copies to your search results? Good thing you can add duplicate to your blacklist.

Because it's filling up the hard drives on Danbooru's servers? Deleting them won't make a difference in that regard, since status:deleted isn't the same thing as destroying the data.

I think if you post a Twitter pic and a better version is available and uploaded within 3 days, you should be penalized for that. Stop rushing to Twitter when you know the artists are just going to upload a better version to Pixiv like 2 seconds later.

Stop and think for just a second about the consequences of what you're suggesting. What happens when the artist accidentally uploads an absurdres or uncensored image to Twitter and then replaces it with a downscaled or censored revision, or uploads something to Twitter but shortly thereafter deletes their entire online presence without warning? The image in question never gets posted to Pixiv as it originally appeared. The window for archiving it closes and it's lost to us forever.

Whatever your opinion of duplicate uploads or the quality of Twitter's image compression, having uploaders who race to grab art from Twitter without delay is a good thing.

1 2 3