Donmai

The Problem with Prolific Uploaders

Posted under General

This topic has been locked.

This has been a persistent issue that upsets many users on the site that I wanted to spend some time to try and address.

Some users are prolific uploaders. Whether by virtue of their schedule or time zone they can monopolize most new content for some tags. This isn't a problem in of itself, but there can be a lot of resentment that they get all the credit while someone who is only one or two minutes late does not.

I think there are explanations for that resentment. Maybe other users find uploading to be a futile exercise and it decreases their engagement with the site. What I worry, however, is that some users are treating uploads as a game where they try to maximize their score by uploading as much as possible. That is definitely not the purpose of this site.

With some trepidation I am starting this topic to: 1) explore the motivations different people have for uploading; 2) explore possible solutions to reduce the angst. I will ask people not to point fingers or name names.

My own motivation for uploading is to share good art I find and to contribute to a database where I can easily search for it in one place. This means I don't care who gets credit for the upload. But I'm not going to force my opinion on others.

Some solutions I've heard:

  • A hard cap on all users. Unfortunately the existence of comics makes this very hard to enforce fairly.
  • Hiding the uploader information. This would certainly demotivate a lot of users, and it seems unprecedented among other sites.
  • Highlighting other information over the uploader, such as the most prolific tagger or noter or wiki writer. This has some technical constraints but it is doable.
  • Do nothing. The status quo isn't great but maybe any attempted fix would just end up being a net negative.

If you have opinions that you don't want to state publically, feel free to message me privately.

Updated

This is a topic I've been highly opinionated about in my own server, so I'll respond later. However, in regards to what I think is perhaps the best solution, it's (2). More on that later, since I'm busy with other things, but I can do a little (or, well rather large) writeup later.

I think most people have noticed I've been getting more detached from the site, and while I like to cover it up with "IRL" things or other obligations, it is the fact that I dislike the mentality that rockstar uploaders bring to the scene, contrary to what we as archivers and uploaders should be doing.

There's a reason why uploading is limited to an increasingly smaller subset of users, and the gallery is being less up to date. Attracting users is the most difficult part, but most of these prolific "rockstar" uploaders are the ones that aren't interested in that. In fact, that would be contrary to their goal, to only have their name on the best content.

Highlighting good contributions in the form of tag gardening, translations (including commentary), artist updates and even comments could improve it. I'd also like if there were like an expanded pools concept where a user can manage their own curated lists which could highlight niche concepts and styles and could add on select other users to help, so quite different from pools or favlists currently. It would also help if the threshold for quality was increased, since some of those high numbers are greatly inflated compared to the contribution.

Definitely against hiding uploader info for reasons you mentioned and because it seems to cause unneeded technical hurtles (searching by user:chinatsu would have to be nixed?). Hard caps not too fond of. Becoming a builder/unlimited uploader is a really hard-earned achievement and would be a big downer for many me included. Though maybe a limit based on community reception of posts could be useful.

I thinks sites like Exhentai are already highlghting other stuff other than simply the uploader name, like adding tags are who does what translations.
I meab, we have the user reports on Isshikei that are already doing this, and that was also my goal when initially suggesting this, but these reports are fairly hidden under a link under "More" now.

A hard cap is still the worst solution of all of this. Like, you have found a big batch on Pixiv (that happens sometimes) and then you can't even upload all of them, despite being so-called "unrestricted". That sounds like it will create tons of frustration, actually. If that's done then the cap should be pretty high, like around 200 uploads and shouldn't cap at 50, since it would also slow down Danbooru and would lose content with that. Admittingly, that's the worst case, but it should get avoided.

I personally am for number 3, like I said above.
Highlighting other stats shouldn't be that difficult. The data is already gathered, there should only be a way to make it better accessible and visible.
But on the other side, number 3 doesn'T seem to have anything to do with "prolofic" uploaders. If the problem is really that only a few users are uploading the "best" stuff then it doesn't get solved. The apparant focus of uploading will simply be shifted to more communical work since other stuff gets attention, but the same users will keep on uploading.
While I highly support this, it has nothing really to do with getting other users to upload or pulling some users back. It is nice, but is an independant act, actually.

The only thing that would actively attack "prolific" uploaders is hiding the uploader's name, but I see some problems with that: If the uploader's name is hidden, then the original uploader loses the responsibility to tag their uploads.
Now, let me explain what I mean: Having no name means that you can't "snipe" each other anymore. Therefore, both users will start tagging one post and I can ensure you that there are different ways to apply tags to posts since not all definitions are hard, like breasts size or hair color. I fear that this could lead to tagging wars and it already has lead to tagging wars in the past, even with the uploader's name present.
Especially during Pixiv rush (midnight in Japan) when a lot of images are shared in a short time, I could imagine quite well this scenario and users not going bananas over uploading but instead over tagging.
Another issue I see with making the uploader's name nameless is really the cut of responsibility. Tagging your uploads is mandatory, so it would make that tagging is actually reserved to people that do actually care that an image is well tagged and that are not that many, actually. It's especially hard to keep track of it, since there are so many posts changes and uploads each day.
There are some search queries that will show you low tagged posts, but then these posts getting tagged with tagme, and that not too short. Meaning that a lot of posts will be tagged like that and you can't really get behind the workload if you try to responsibly tag every image under this tag. Tagme is already treated by a handful of people already...

Regarding well-tagged uploads, I fear that exactly this will happen. I don'T say that it's 100% sure, but some users are already only tagging the bare minimum despite knowing more tags and I fear with less influence to push users to tag "their" stuff from a handful usersthe community, more posts will be tagged low.
For posts that are posted during rush, too many people will tag an image when it should only be one and creating tagging war and other stuff that aren't multiple persons are going for are probably ending up with mediocres or less tagging.

The other problem would be posts like post #2974764. Third-party edits. Since these uploaders can't be monitored that well anymore, this stuff might get uploaded en masse then. Now you could say that we can look at the tagging history of the posts that reveals the tagger, but then you could just have the uploaders name. It's far less complicated. On top of that, only Mods can see that. To contact every time a Moderator to tell this user to stop uploading this crap is rather irksome.

You could do it like that: Hide the uploaders name, but only up to Builder level users. There are plenty of Builders on the site and they could still take care of this pretty easily. Of course then you have to ask why we should hide it at all, because I think don't think it will be a motivation for Members to start becoming Builder simply to see who uploaded what.

Yes, hiding the uploader's name will solve that dilemma that only one user is uploading from Touhou, but there are some unwelcomed side-effects as well and it's easier to abuse uploading per se then.

Updated

Without beating around the bush, I'm fairly sure everyone already knows that what's causing strife between uploaders is the practice of "sniping" and nothing else. New users get frustrated easily when they see a few users monopolizing a copyright or artist simply because they already know the tricks of the trade (distill, using /uploads/new instead of the bookmarklet to be faster, the dreaded mintagging or even 0-tags-on-uploads-tagging), and it's currently practically impossible for anyone inexperienced to upload new posts from popular artists because there's always 5-6 users at any time using monitoring tools on them. As an example, see the comments in post #2953054 for a prolific uploader that recently stopped uploading precisely for this reason.

If you want to take a look at the problem, check the uploads page at midnight JST/10 AM EST/4 PM CET (when Pixiv rush hour happens). It's a battlefield.

The problem is that you either play the game by its current rules and only add copyright and character tag on upload, or you waste your time trying to tag the picture only to see it taken by (and credited to) someone who used two tags on upload. Of course that'd cause people to get angry at what they perceive is bad practices.
I really don't care about credit either, which is why I scrape artists for missed posts rather than camp the feed for popular uploads, but it's undeniable that getting redirected to the duplicate page after having tagged a post fully because the other guy used two tags on upload is pretty tiresome, especially when it happens frequently. And not everyone thinks in the same way - there's users who want to upload new content, but as I stated above it's impossible for them.

I'm not sure there's an actual solution to this problem.

  • Preventing uploads with less than (say) 10 gentags will severely hinder uploads for posts that actually can't reach that amount (some scenery pics, minimalistic artwork and the - possibly irrelevant in this context - ending/credit pages of comics). Preventing submission with less than 5 gentags will do nothing as artist+copyright+character+1girl+solo already is five tags.
  • Hiding the uploader's name will be even worse because being credited is already a big part of the reason why people upload. Take that away and a good percentage of active uploaders will just stop doing it. It also opens the door for abuse, as people will just upload bad posts with no worry as they won't be credited or held responsible by anyone but users of enough level to see it.
  • A hard cap will do absolutely nothing for this and will simply stop people like me who mostly upload old posts. I frequently up more than 100 posts a day and I know of at least another two users who do, and I've gone past 200 posts a day several times before, and yet none of us "monopolizes" a single copyright. Furthermore, if you look at the contributor report you'll see there's only three users that upload more than 35 posts a day.
  • Highlighting the tagger as well is possibly the best solution of the bunch. It has been proposed in the past to give the credit of uploading to the tagger if they're responsible for, say, more than 80% of a post's tags (easily doable on a sniped post), but that would probably cause more foul play as people try to add wrong tags just to make count - though that could easily be stopped with feedbacks and bans. As a side note for this, crediting the translator of a post too would also be a pretty welcome addition.

Also, automatically merging tags on duplicate uploads would probably help a lot in reducing the issue - some users told me they just give up merging if they get the duplicate page because it's just too bothersome a process to copy the tags from the static page.

Updated

I honestly don't think we need to worry about running out of uploaders. The ones uploading "for the fame" quitting doesn't really worry me at all in terms of the health of the site. However I figure that removing the uploader's name will only result in a more inflamed version of what's already occasionally happening with post score; "a bunch of my recent posts have been downvoted recently, that guy I had an argument with three days ago must be attacking me" and it turns out to be nothing of the sort.

Absolutely agreeing with OOZ.
Now, the percentage of how many posts get uploaded have switched, but if we are talking about two things:
1. Total amount of uploads
2. Total amount of uploaders
I want to make a comparison between the most recent reports and reports from over one year ago, more specifically the first report ever created:
https://isshiki.donmai.us/user-reports/contributor_uploads/2016-10-28_v1.html (October 2016)
https://isshiki.donmai.us/user-reports/contributor_uploads/2017-12-31_v5.html (December 2017)
It seems to me that the fear we are running low on uploaders in this section isn't really justified.
We have aroubd the same amount of uploaders if not more. Of course the names are switching, but it's growing.
The total amount of uploads is generally increasing.

Now for user that don't have unrestricted uploads:
https://isshiki.donmai.us/user-reports/member_uploads/2016-10-28_v1.html (October 2016)
https://isshiki.donmai.us/user-reports/member_uploads/2017-12-31_v5.html (December 2017)

Now what do we have here: Also a solid amount of uploaders, and the numbers are staying the same over the time (except @Rignak whose uploads decreased drastically).
Now, these are only numbers of course and can only reflect so much. Numbers don't explain the feeling that users apparantly "own" a copyright and upload only the best stuff. But it seems that most users are untouched by it, since the amount of uploaders is increasing.
Users keep quitting and coming, that's nothing unusual.

Sacriven said:

^in other words, Danbooru is stagnating then. That statistics are interesting, but that doesn't mean Danbooru is okay in the current way it is.

We can play these statistic games further:

http://danbooru.donmai.us/counts/posts?tags=date%3A2015-01-01..2015-12-31 (Total amount of uploads 2015)
http://danbooru.donmai.us/counts/posts?tags=date%3A2016-01-01..2016-12-31 (Total amount of uploads 2016)
http://danbooru.donmai.us/counts/posts?tags=date%3A2017-01-01..2017-12-31 (Total amount of uploads 2017)

Well, these are only numbers of course. And those should get interpreted precisely if we want to know if Danbooru really is stagnating.

Here are some day-by-day data : http://i.imgur.com/0JvKzty.png
We can see that the number of pictures uploaded is slowly increasing.

In the same time, the rate of upload on pixiv seems to decrease :
id frame | time frame | number of days
64M -> 65M | 09/17/2017 -> 11/22/2017 | 66d
54M -> 55M | 12/12/2015 -> 01/31/2016 | 50d
44M -> 45M | 06/09/2014 -> 07/29/2014 | 50d
34M -> 35M | 03/04/2013 -> 04/14/2013 | 41d

I'm more in favor of the option of highlighting the tagger. And even more if the merging of tags is developed.
An hard cap is only removing the unrestricted uploads permission. It will affect a very limited number of people while making no distinction between the snipers and the others.
Hiding the uploader's name is quite hazardious, we don't know how the users will react. Perhaps, it will cause less uploads and a decreasing quality in the tagging.

I would like to suggest “reserving” uploads again, which I already did in forum #131461. That would go well with the suggestion to show the top tagger(s).

The main difference to my previous suggestion is that there would be an additional field like “Top tagger(s)” on a post that showed exactly that. If Alice added almost all the tags, she would be the top tagger. If Alice added 30 tags, Bob 20 and Carol 2, Alice and Bob would be top taggers, etc. (The numbers are just examples.)

Now that might start tagging wars, which would be undesirable as well. That’s where the reservation system comes in.

Basically it would work like this:

  • If the uploaded uses the reservation feature, an image is uploaded immediately to Danbooru but won’t be accessible to other users for a short amount of time.
  • The uploader can take the time to properly tag the image, probably easily securing the top tagger spot.
  • When the uploader submits the tags, the image will be visible for everyone.
  • If the uploader doesn’t properly tag the image during the reservation time, the post is up for anyone to tag and the uploader loses that reservation slot for a while. My previous suggestion was for the upload to be discarded entirely so that someone else can upload it instead.
  • If another user wants to upload the same image during someone’s reservation time, an error will be shown so that the second user doesn’t waste any time on tagging in vain and can just move on to another upload.

More details and ideas how to combat abuse of that system in forum #131461.

While it won’t solve all the problems, I think it would improve the situation a lot because it would fix the two probably biggest problems:

  • Wasting time tagging an upload in vain because someone sniped it in the meantime.
  • A few users hogging most of the uploads with min-tagging (during Pixiv rush, etc) if every user has only a limited amount of concurrently usable reservation slots and is motivated on spending some time on tagging before uploading much more.

Bonus advantage: Currently, many new users feel that uploading is the only way to contribute or make yourself known, often leading to sub-par uploads. If the top taggers are shown and there’s not much tagging war going on, the importance of proper tagging would be highlighted and more users could be motivated into helping with tagging.

Problems not solved:

  • It’s still possible to not use the reservation feature and not care about the top tagger spot by sniping uploads by uploading with pretty much no tags at all.

My suggestion for this reservation feature isn’t final, of course; further ideas are welcome.

kittey said:

  • It’s still possible to not use the reservation feature and not care about the top tagger spot by sniping uploads by uploading with pretty much no tags at all.

That's the deal though, I doubt anyone interesting in "uploading fame" or whatever reason they have to upload would care about having another name next to theirs as long as they get the uploader credit, and furthermore it's not guaranteed that anyone else would be able to tag the mintagged post, as the uploader is the first person that sees the post as they're redirected to it while everyone else involved gets the duplicate page. This means a fast typer with an average connection can immediately fully tag the picture even after 0-tagging it, simply by adding tags three-four at a time and quickly submitting them.
I'll in fact add that it's pretty difficult to get most of the tags in a sniped post unless you're very fast and ready for them, and you're also tagging in the way I described above, otherwise you'll end up getting your tags sniped too.

To take it even further: even if the tag merging was implemented, people getting sniped would still not get credit most of the time because if you take 60 seconds to fully tag a post, by then a sniper will have already mostly tagged the post as I described above.

I don't want to point fingers and I'm not particularly bothered by sniping so I'll avoid making names, but all the usual mintaggers have an average tag number (sometimes well) above 25 as shown here, which also means we can't really identify and stop people who frequently snipe just by numbers alone.

Frankly I don't think there's a definite solution to this "problem" that won't greatly upset either side, and I'm not sure it's a real problem at all - I know mine is a fairly unpopular opinion, but if people want to upload for e-peen they need to step up their game and compete against anyone else that does. There's plenty of ways to score fast, popular uploads, and the methods that all the currently active fast uploaders use are publicly known and freely available - in fact anyone can read a full article on them by searching for "upload guide" in the forums.
Popular artists and copyrights naturally get the highest scores - I find it absurd that someone interested in getting those high scored posts would get angry when others are competing for the same reason, or that they'd get frustrated when they aren't even trying to be as fast as the competition.

The only thing I can see happening with a reserving feature is people hoarding the most popular artists even more and leaving the "scraps" to everyone else, which will in turn fuel pointless anger as the ones left behind still want the "best" uploads.

As a final note, I feel people are needlessly aggressive and too easily upset about the whole thing - rather than seeing uploading as a mean to complete this site's collection and use it as a high quality art database, they are quick to antagonize and hold grudges against each other. I doubt the people who got angry at the user who is apparently "hoarding" a copyright (I'm sure most here already know who it is) even tried to contact said user at all to ask for help or ways to improve.

As a somewhat uploader myself, I'm happy enough if others upload more of my favorite artists, copyrights and characters. In fact, if I knew that they were being covered sufficiently by other users, I probably wouldn't upload at all. I've got more stuff to do anyways, and curating art can be hard work.

As far as ideas go, I've got a couple that could be used in part or as a combination with other ideas.

1. What about hiding the uploader name just for a short duration... like a day or so? Most posts are looked at in their initial period anyways, and so it may curtail any competitive behavior, but it would still give credit to those that went through the effort to upload.

2. What about locking the post for editing on the uploader for up to an hour after upload? This would more strongly highlight those users that mintag, and it would allow others to get tagging credit if that becomes a thing. Admittedly, this one would require a bit more work as I've heard complaints about the usability of the uploads page. I find that page just fine myself and usually fully tag before upload, but I'm also not in a rush to upload pictures before others can snipe me.

nonamethanks said:

That's the deal though, I doubt anyone interesting in "uploading fame" or whatever reason they have to upload would care about having another name next to theirs as long as they get the uploader credit, […]

I’m sure that system can be tightened up more, which is why I’m hoping for suggestions for improvements. Attempts to game the system, like adding nonsense-tags to fulfil a minimum tag requirement, could be a bannable offense.

[…] and furthermore it's not guaranteed that anyone else would be able to tag the mintagged post, as the uploader is the first person that sees the post as they're redirected to it while everyone else involved gets the duplicate page. This means a fast typer with an average connection can immediately fully tag the picture even after 0-tagging it, simply by adding tags three-four at a time and quickly submitting them.
I'll in fact add that it's pretty difficult to get most of the tags in a sniped post unless you're very fast and ready for them, and you're also tagging in the way I described above, otherwise you'll end up getting your tags sniped too.

I’m not sure I understand how you understood my suggestion...

The uploader, i.e., first submitter of the image, will have a short amount of time, let’s say five minutes, to tag the post without anyone interfering. Any further upload attempts during that time will simply lead to an error message with no option to add tags until the post is released. In other words, there is no tag merging during that time because it’s not necessary. Afterwards, be it because the uploader finished tagging or let the reservation time expire, tagging (and tag merging) works as it does now past upload. Attempting to snipe-tag during the reservation period is kinda stupid because the uploader might already be adding all the tags. Unless you expect the uploader to let the reservation expire, in which case you have the full reservation time to prepare your tags and submit them later, in which case tag merging will work. If you meant attempting to tag-snipe a post that the uploader let expire and compete with other tag-snipers, well, it’s up to you if you really want to fight that battle or just come back later and relaxedly tag whatever is still untagged.

To take it even further: even if the tag merging was implemented, people getting sniped would still not get credit most of the time because if you take 60 seconds to fully tag a post, by then a sniper will have already mostly tagged the post as I described above.

If you get sniped, don’t tag the post, move on to another one, reserve it, and take your time tagging it. My idea is to reduce any kind of sniping to reservations, which take almost no time or effort that might end up wasted.

I don't want to point fingers and I'm not particularly bothered by sniping so I'll avoid making names, but all the usual mintaggers have an average tag number (sometimes well) above 25 as shown here, which also means we can't really identify and stop people who frequently snipe just by numbers alone.

That statistic shows the total added tags, but the site also remembers the tags present at upload time. We just don’t have a statistic for the latter kind.

The only thing I can see happening with a reserving feature is people hoarding the most popular artists even more and leaving the "scraps" to everyone else, which will in turn fuel pointless anger as the ones left behind still want the "best" uploads.

In forum #131461 I mentioned how reservation slots would be limited, to avoid exactly that. As almost no time or effort is required in acquiring a reservation, I think more users would have the chance to get their share of popular artists.

As a final note, I feel people are needlessly aggressive and too easily upset about the whole thing - rather than seeing uploading as a mean to complete this site's collection and use it as a high quality art database, they are quick to antagonize and hold grudges against each other. I doubt the people who got angry at the user who is apparently "hoarding" a copyright (I'm sure most here already know who it is) even tried to contact said user at all to ask for help or ways to improve.

Welcome to the Internet, where everyone is a nobody, just like everyone else. And as you can’t beat someone over the head to become a somebody, which seems to be the biggest goal for some people, all you can do is play the game the way it is.

kittey said:

If you get sniped, don’t tag the post, move on to another one, reserve it, and take your time tagging it. My idea is to reduce any kind of sniping to reservations, which take almost no time or effort that might end up wasted.

Most of the mintagging & sniping happens at rush hour, which is japanese midnight. That's when pixiv resets its daily ranking counts, so everyone interested in getting their art out there will upload to pixiv at pretty much the same second. This means that all users wishing to partecipate in uploading from the pixiv rush will try to grab as many posts as they can from their feed as fast as they can, often just by looking at the thumbnail from the feed, and only after they covered all of their feed they'll check which ones they didn't get. It almost never happens that you grab one post, hit submit, wait until you know if it went through, then grab another post.

I am uncertain what weight, if any, the opinion of a mere Plat like myself possesses in debates like this, but I agree with Noname's assessment of highlighting the tagger, note stamper, and wiki editor alongside the uploader. It seems like the best compromise we have on the table thus far, to spread the credit around more evenly. I also agree that reserving posts isn't the best idea, as it encourages rush hour poaching, which Noname presented just above me. And enforcing upload caps or removing credit entirely would be more likely to damage the site and community than solve our current problem. As I understand it, the goal is to minimize credit-hoarding, not encourage it. Adding more things to be credited for (translation notes, tagging, etc) could help with that.

With that said, I simply wanted to offer a "civilian's" perspective on the debate. My apologies if this was meant as a Builder-only discussion.

WatcherCCG said:

With that said, I simply wanted to offer a "civilian's" perspective on the debate. My apologies if this was meant as a Builder-only discussion.

This is an open forum discussion. Everyone should take part in it if they want to share an opinion. User rank has barely something to do with it, unless you're an Admin (or more preisely Albert) who has the final say...

My opinion: If we rank tag count on an image, we'll see an increase in padding tag counts by tagging every little thing like the color of minor accessories.

I stand by option 2 in favor of removing uploader names entirely.

Edit: tldr; ranking things creates frivolous competition.

1 2 3 4 5 8