Posted under General
I assume this is just the default reason and most don't bother to change it given the volume of posts they deal with, but why is it even an option at all? There are a lot of reasonable looking posts that don't get approved (just look at the appeal thread to see some), with only 0-1 moderators providing a reason that isn't this one. It makes the process look heavy-handed and lazy even though it may not be.
And why is the deletion request thread stickied but the appeal thread not?
Nell said:
I assume this is just the default reason and most don't bother to change it given the volume of posts they deal with, but why is it even an option at all?
Because it IS the default reason. Approvers are explicitly told that they should "only approve of posts that they personally like". If they don't like the post, they can either ignore it and let it linger in their queue, or hide it with "no interest" button. Doing the latter adds to the "didn't like the post enough to approve it" counter. I don't see how is it heavy-handed or lazy.
Re: sticky, no idea, but probably because appeal thread resurfaces often enough to not need a sticky. It's linked in topic #8235 anyway, so it's no big deal.
An explanation I believe I saw somewhere on here is that it takes the disapproval of all users who review an image to delete it, but only one user who reviews it to like it to approve, so if not a single one likes it, the image may likely not be very high quality - or with fewer presumptions, (on paper) this system makes it much more balanced in favor of the uploader of the image.
It's not really a perfect metric either way, because, as is simply the case with people, it's easy to let preference cloud judgment, and not everything will meet everyone's preferences, thus leading to images that are just fine from mostly objective standpoints (more scrutinized themes excepted) getting the boot for a variety of largely subjective reasons. Though obviously, going through the list of deleted images, it's plain as day to understand why a large majority were deleted, whether a reason was given or not.
On the flip side, despite those imperfections with this system, it may very well be the best way to handle it, as it's not exactly reasonable to expect every single user with approval privileges to share the same standards as the next, or the same knowledge on what is considered well done or not, and so on. I guess it's just a shame to see those diamonds in the rough that have a lesser chance of ever being approved due to how little of an impact appeals seem to have, especially given that - if I'm not mistaken - the ones handling those are the very same who didn't like an image in the first place.
Of course, this is all from the perspective of someone who believes approvals to be done based on objective value over subjective preference, as seems to be the purpose of the system in the first place. The potential of the latter being true is a whole can of worms I'd rather not touch.
In the end, a simpler answer may be, if the message was something like "Some number of moderators did not find this image to meet a standard of quality," it'd be equally vague as to why an image was deleted, for the same reasons you mentioned. Just my two cents (or twenty).
Updated
Approvers are trusted to apply an objective standard of quality in what they approve, regardless of their personal taste for the content - much in the same way contributors are trusted to apply an objective standard of quality in what pictures they bypass the queue with.
That said, the approval process is mostly subjective in nature: the main metric to approve something is whether you like it and it is what you're explicitely told to do as an approver.
So I just approve things that look good to me - and my approvals probably make it quite clear what kind of art I'm after. Franchise bias is mostly not a thing for me though, for example lots of Kemono or Fate pictures in my approvals even though I have not watched or played or w/e either of those things.
Cephalopa said:
It's not really a perfect metric either way, because, as is simply the case with people, it's easy to let preference cloud judgment, and not everything will meet everyone's preferences, thus leading to images that are just fine from mostly objective standpoints (more scrutinized themes excepted) getting the boot for a variety of largely subjective reasons.
One of the reasons behind having a large pool of approvers is to mitigate the role of individual bias. Most approvers here probably have certain art styles or subject matter they absolutely loathe and would happily see deleted regardless of how well executed or visually interesting it may be, but this is compensated for by other approvers with entirely different sets of likes and dislikes. Practically any half-decent upload has a fair shot at approval; art appealing to niche interests like furry and guro gets approved here on a regular basis, for instance.
I guess it's just a shame to see those diamonds in the rough that have a lesser chance of ever being approved due to how little of an impact appeals seem to have, especially given that - if I'm not mistaken - the ones handling those are the very same who didn't like an image in the first place.
From what I've seen, those "diamonds" that somehow go unnoticed in the queue are very few and far between. The overwhelming bulk of posts mentioned in the deletion appeals thread is the same old generic fanart/pinups over and over again - not bad art, per se, but nothing the likes of which we haven't seen thousands of times already, and definitely nothing I'd want to frame and hang on my wall. Just because an upload lacks technical faults doesn't mean it adds value to this site.
The large number of approvers is specifically because the rules are stacked heavily in favor of approving something (one random approver that likes something can overrule every other approver hating it) and they are explicitly told to be completely subjective in what they approve. As has been argued many times before on these forums, approvers are a ring species that have developed around specific tastes in a system that overtly encourages Eternal September, explicitly incentivizing territorial behavior. If you look at the comics on this site, you'll see the same four approvers approving absolutely everything, and there are some approvers whose names I only see when attached to some of the porn I had explicitly marked as "poorly drawn".
I, personally, go through every uploaded post and will try to apply quality standards, although there certainly are things I don't like and will judge more harshly, but with that said, there are even filters you can apply to the mod queue to make it so you don't see any post that doesn't fit certain criteria. Between "only approve what you like" and that, you can see some real differences in what approvers will approve. Considering some wild variance in how often different approvers actually log on, you can get some notable differences in what quality is approved just by when approvers with lower standards are active.
And yeah, because of that, I've apparently gained one of the highest rates of approvals among approvers, and I generally have been good about going through new additions every day, so it's not that things go unnoticed, it's that I find reasons not to approve things.
Keep in mind that even if something looks good, there may be other reasons not to approve something. It may be tagged as a sample, it may be a really detailed official art that has really terrible jpeg artifacts from someone doing horrible compression ripping it from the source, the uploader may have screwed up uploading or tagging somehow. There's also the whole anatomy thing, where there's reason to hold off on approving something with questionable anatomy.
And Iridescent Slime is right - there are more things in consideration than flat artistic skill, alone. Sweetpea likes to talk about it in terms of a point scale, where if an image winds up with a positive score, it gets approved. So you start with a score from artistic skill, then add or subtract points for different things like bad anatomy, good composition, and whether or not the image actually manages to be entertaining in some way, and the overall score determines whether you approve it or not. This post, for example, was likely deleted in spite of being technically passible because it's just the same character about which we have plenty of pictures just standing there in front of a blank white space. This post, however, has the character actually doing something in a setting with actual context that makes it more amusing and worthwhile as something to look at while already browsing images of that character. This post, meanwhile, is a fine art parody, which makes it amusing enough that I'd approve it in spite of having some awkward art style. (Not that I had a choice in that one, it was uploaded by an unlimited uploads user.) Likewise, I will approve comics of much lower artistic merit if they're funny or tell an interesting dramatic story because those elements will add a significant positive score to counterbalance the negatives of generally hastily-done artwork.
As said before, however, there are real differences in how people weight these things, so you'll wind up with some approvers approving basically all the comics on the site while other approvers seem to hate the fact we have comics at all.
I do get where you're both coming from, but if liking an image really is the sole valid reason for approval as you all say, aren't those with a more lax standard for approval working a bit more in line with how the flag/appeal system works at present?
For instance, if the images that get through are to be of consistently high value, doesn't that detract from how freely usable the flagging system is? The restrictions on the appeal system do make sense with both instances, but with those, why have appeals make so little impact if the cases where a "valuable" image gets missed are supposedly so few and far between? Especially considering the point you brought up that not all approvers are on at all times of the day.
It just feels so backwards to me, but I guess as just another user it's not really my place to say anything. I don't intend to be standoffish by any means, but I've just been getting a lot of mixed messages in my time on the forums and it's pretty frustrating as someone who loves the site. I'm probably overthinking it. lol
The thing is, flags generally don't go through, and generate a TON more drama than simply not having something approved in the first place.
There's also the presence of the "poor quality" and "breaks rules" buttons as well as the "no interest" button when approving, which tells other moderators, essentially, "Don't approve this." Combined with the capacity to flag (IF they notice it was approved, anyway), this carries some weight of threat with it. An approver can lose approval privileges if they have too many of their approvals flagged, and it's more than slightly frustrating and spurs conflicts to have such overt clashes, so in general, it acts as a soft push against other approvers that is sorely needed in generating some form of consensus, at least among the more active approvers who care what the other approvers do. (Notably, most of the active approvers are newer, and have lower ranks, like, say, everyone in this thread but Type-Kun, while the "green name" janitors and "yellow name" moderators have been around much longer and are as a group far less active and give less of a damn about what everyone else is doing...)
Besides that, the flagging system isn't really that freely usable. You have one per day for most users, and 10 per day for higher-ranked ones, and some having unlimited flagging based upon going specifically after things like samples. Most flags that aren't along some really narrow broadly accepted line of argument (samples or bad anatomy) tend to be overturned for bucking the consensus.
The problem would be that this is a system that doesn't really have rules, it has an uneasy consensus forged between a dozen active users who mostly work things out because they have to on a daily basis, peppered occasionally by some rare faces who have totally different goals and viewpoints from the others who don't care about the others' consensus. The net result is that the rules are a subtle thing prone to changing any time the balance of approvers changes for whatever reason, and there isn't any obvious communication to the users when this backroom politics happens.
That lack of rules seems baked into the system on purpose, essentially just to give the people who were at the top of the heap first the ability to be as arbitrary as they wanted, but it basically falls upon those who are actively manning the ship, as it were, to try to figure out how to make something resembling a consensus in the wake of their passing.
EDIT: When I said "everyone but Type-kun" to "everyone IN THIS THREAD but Type-kun".
Updated
I generally agree with NWSiaCB, but maybe the issue will become more obvious with these slightly political examples.
If Danbooru were a democracy with moderators/approvers elected by all users, we’d probably have even more bickering because more users think their opinions could make a difference. We’d also have more of what the currently non-vocal part of the user base wants, which I guess is porn and more art that’s not so good (aka. mediocre).
If Danbooru were an autocracy with Albert setting fixed rules, we’d probably have little to argue over and the content would mostly match what he thinks is right, possibly high-quality art. Autocracies are generally perceived as stronger and more stable, which is why many people, historically and currently, prefer strong leaders over democracy. Unlike most real world autocracies, Danbooru users would have the choice to just leave if they don’t like the fixed rules of this site, so the remaining user base would like it here, thus leading to the aforementioned less arguing. I guess neither Albert nor the other admins want to micro-manage a chaotic bunch of idiots, so Danbooru is not an autocracy.
Instead, I’m getting the feeling that Danbooru is a twisted god game, with Albert and the other admins providing the infrastructure and assigning some individuals from the population to certain roles based on how compatible their goals seem to be... and then watching all their subjects mingle and organize themselves within the lax rules of their world, occasionally correcting undesirable behavior by divine intervention. Even though I’m just one of the subjects, I watch this chaos unfold every day, wondering what it will turn into next and hoping that I will like it. :) It’s of course possible for anyone who doesn’t want to be part of this god game to just leave.
The more I think about it, the more Danbooru looks like a god game or a resource management game.
kittey said:
The more I think about it, the more Danbooru looks like a god game or a resource management game.
- Subscriptions are slowing down the site; we must research saved searches.
- Our storage is filling up; we must research off-site storage.
- Our subjects are unhappy about advertising; we must research other sources of income.
- Our subjects are complaining about Failboorus; we must repair the servers.
Jokes aside, I'm pretty sure that the rules were made lax on purpose, and I wholeheartedly support that. We are judging art, which is incredibly fluid and unpredictable, and it's hard to formulate the rules while accounting for all the edge cases. It goes much like a counter-argument for Rule utilitarianism - the rules will have so many exceptions that it's better to have no rules at all. So we have a few hard guidelines to define the purpose of gallery (no real-life porn, for example), a few more soft guidelines (furry and scat are acceptable on conditions the art is VERY good and catches approver's fancy), and the rest is up to individual judgment. Having stronger rules with explicit note that any of them might be broken "at approver's discretion" is even worse, because it will generate horrendous amounts of drama from people who are keen on following rules to the letter.
Of course, one could argue that "individual judgment" can be done by the means of "vox populi", mass voting, and IIRC Albert considered switching to that at one point. However in my opinion it wouldn't bring the desired effect of "keeping good art", but would flood Danbooru with mediocre porn instead - see "order:favcount" pages besides the first few.
Type-kun said:
Jokes aside, I'm pretty sure that the rules were made lax on purpose, and I wholeheartedly support that. We are judging art, which is incredibly fluid and unpredictable, and it's hard to formulate the rules while accounting for all the edge cases. It goes much like a counter-argument for Rule utilitarianism - the rules will have so many exceptions that it's better to have no rules at all. So we have a few hard guidelines to define the purpose of gallery (no real-life porn, for example), a few more soft guidelines (furry and scat are acceptable on conditions the art is VERY good and catches approver's fancy), and the rest is up to individual judgment. Having stronger rules with explicit note that any of them might be broken "at approver's discretion" is even worse, because it will generate horrendous amounts of drama from people who are keen on following rules to the letter.
The problem is, I find that not having any rules doesn't teach those who want to live strictly by the rules learn to be more lenient, it just teaches them that there are no rules besides quality, so only a few objectively definable things like lineart and anatomy matter, and everything else is fair game. (Well, or "no photographs". When someone uploaded one of those recently, every single mod that got to it before me hit "Breaks Rules", so apparently, photographs is consensus hard-line off-limits.) It's fine to say something about how, the further away from anime things get, the stricter the standards need to be to overcome the "negative weight" of increasing irrelevance to Danbooru's main purpose (such that something like the Mona Lisa is past a hard-line cutoff), but as the argument in the Doom image thread shows, all that really does is invite people to say "off-topic is allowed", that there are no genres of artwork that are any less at home on Danbooru, and "Stricter Scrutiny" is a meaningless fluff word nobody even recognizes when you try to use it as a flag reason.
If you're going to generate a system whereby people need to rule by consensus, then at least make a system that actually requires people to come together and form a consensus to do things. As it stands, this is mostly rule by whoever doesn't get noticed or generate enough drama to actually raise enough alarm that someone steps in and stops them. Flagging, again, is the only thing that really stands between unlimited uploads users and uploading literally anything that is of a file type Danbooru accepts, but the only way for someone to flag is to go out of their way to watch the uploads of the particular users that are making those uploads, and that just invites people being decried as showing bias against those users.
Updated
Flandre5carlet said:
Approvers are trusted to apply an objective standard of quality in what they approve, regardless of their personal taste for the content
According to this quote that's not true:
Type-kun said:
Approvers are explicitly told that they should "only approve of posts that they personally like".
This isn't the first time I've heard that and I think I even remember seeing a screenshot showing it. Approvers are told to approve based on subjective personal standards rather than objective quality/ToS based standards. Honestly, I see that as being a big problem. It seems bizarre and counterintuitive to keep the approval process like that since it almost turns the site into those few users' personal picture folder when the site is suppose to be for everyone.
I think it would be far better if the approval queue would tell them to approve in a way that Flandre5carlet described rather than personal preference.
My approval habits are more along the likes of what Flandre5carlet describes, but I understand that not every janitor does things this way, especially the old ones. Most of the new janitors are the most active ones, especially with Provence's departure.
It kind of does seem like we're on this eventual cycle of janitors though... for a long time NOOU was the prime "safety catch" post approver, then it was Provence, and now it's a multitude of users with Qpax standing at the top as of these past few weeks. Assumably, these users aren't approving everything they simply 'like' but they at least gave every post a deserving chance of being seen. As NWS mentioned here:
(Notably, most of the active approvers are newer, and have lower ranks, like, say, everyone but Type-Kun, while the "green name" janitors and "yellow name" moderators have been around much longer and are as a group far less active and give less of a damn about what everyone else is doing...)
This is infallible fact as shown in the most recent report here. One thing to also note is that the disapproval options are of no use by them, so we can't actually determine how much they're actually looking at the queue.
Back when I started uploading, the queue struggled with active moderation. Perhaps these janitors just didn't bother using the actual options, but it was definitely a perceived effect -- the deletion threads would continuously get bumped with uploaders asking for more reviews (given the only 3-4 who saw them regularly). And so did the standard get perceived to be higher, because Provence was the sole carrier of those approvals during his tenure here. Now that he's left, it's in the hands of a few regular users of the site that have been chosen for their high level of scrutiny and measure of each post's quality.
I wonder how long this system might last, but even if we were to argue in some way how the userbase could get to "select" images by way of a democratic process, this is inevitably bias towards questionable and explicit images since those tend to be what users most look for. Yet what arguably keeps the site active and basically not a hentai site is other grassroots uploaders that don't upload all the porn and the like. This includes that niche ground for translated fancomics which NWS largely dabbles in.
I think the best way to sort of ease this problem is to continue making efforts to make it easier to upload content for normal users. There are a couple standout member-level users that might get promoted in the future, but it's up to the moderators to notice and select them. Doesn't take too long, although this was partly encouraged by the "user reports" Provence would keep sending back and forth between primarily promoting Moderator+ users like Wypatroszony (who promoted me). Then perhaps this can continue to spiral forth and from that pool we can also select more active Janitors in the approval process.
Hoobajoob said:
-snip-
Well, that's more or less how I personally approve things anyways, and that's how I understood albert's comment in the last Western thread:
albert said:
Approvers are selected based on their ability to select good art regardless of where it comes from.
To me, I interpreted that as meaning that Approvers are trusted not to approve bad art even if they might like it, and so that means to apply a certain objective standard to what they approve.
Flandre5carlet said:
Well, that's more or less how I personally approve things anyways, and that's how I understood albert's comment in the last Western thread:
To me, I interpreted that as meaning that Approvers are trusted not to approve bad art even if they might like it, and so that means to apply a certain objective standard to what they approve.
>Approvers are trusted not to approve bad art even if they might like it
Doesn't that imply that "Approvers are trusted to approve good art even if they might not like it"?
