I see no reason why this should be forbidden from being a pool. It's quantifiable information that at least two people think would be at least somewhat useful. Just looking at the tag, I've already found a number of great posts that I never would have seen before due to their low scores.
No one is saying that we should recognize Pixiv user's eye for quality as being accurate. Just that it's information that should be somehow recorded. The more metrics an image is judged by, the easier it is to find good ones, even if the metrics themselves aren't very accurate.
The tagged images are, for the most part, very good.
Also, touhou score:>x doesn't quite cover it. There's a number of very good posts that have low score/favcount. Not to mention all poor quality posts with scores in the 10+ range.
We've basically proven here on Danboo that the average user has incredibly shitty taste, and I doubt that Pixiv is all that much better.
I search for tags, and I also visit the popular page - which is useful in its own way. It's foolish to distinguish "our users fav" and "that site's users fav". After all it's just a means to tell you something about the quality.
No, they are not better than us. What they have is the mass, it's incredibly hard to have 1k people to like the same "shit". Give me an example of what you'd define as mediocre art within touhou_project_1000_users. If you can, I'll gladly nuke it myself and shut up.
theadonicus said: Since there are other ___1000users tags on Pixiv but Touhou is the only one featured here, it feels like there's some Touhou bias going on. Says the guy whose uploads are 70% Touhou....
Over a fifth of the posts on Danbooru are Touhou. Of course there's a Touhou bias.
rantuyetmai said: Give me an example of what you'd define as mediocre art within touhou_project_1000_users. If you can, I'll gladly nuke it myself and shut up.
I was going to point out that there are two deleted posts with the tag...but then I realized they were deleted at the artist's request.
rantuyetmai said: I've always under the impression that where as it is possible (objective, non-sequential), a tag is preferred.
In retrospect, this tag is not even slightly objective, and I have no idea why I thought it was in my first post. It's based on a collective of users' personal opinions about what constitutes good art; that's practically the definition of subjective, even if said collective of users happens to be Pixiv and not Danbooru. A tag is really out of the question here.
That said, the arguments for not even allowing this as a pool sound a lot like the arguments you see on Wikipedia about notability. We're not gonna run out of 0s and 1s, guys, and I just don't see this pool 'opening the floodgates' and prompting dozens of pools like it for other sites (which would be a problem, if it happened, and is the only real counterargument I can think of here). Given that, this concept has no less inherent validity as a subjective pool than "Disgustingly Adorable".
It is a lot of work with very little gain. Someone will have to keep a constant eye on the pixiv tag and since you can't sort tags by the date they were added to a post only by date they were posted you're going to have to watch the entire set constantly to keep it updated.
Tag subscriptions simply will not work, subscriptions inject the post into your subscriptions based on when the post was posted on danbooru not when the tag/pool was added. Seeing as how very few if any posts will be posted after they receive 1000 favorites on pixiv this will not help anyone find newly 1000ed posts.
eidolon said: Also, touhou score:>x doesn't quite cover it. There's a number of very good posts that have low score/favcount.
The median score for touhou is around 2 or 3 and ~90% of touhou_project_1000_users has a score of 3 or higher. So most touhou_project_1000_users posts have a higher than average score relative to regular touhou posts and will in fact be covered by a touhou score:>x search. I don't think this tag actually helps that much with finding underrated posts.
Pool it for reasons giving by glasnot, Coconut, and Fencedude.
I'd also like to point out that this concept would be an "unstable" tag because it has nothing to do with the actual content of the image but what others think of the image. What if a user removes an image from their bookmarks and the image only has 999 favourites? Should the tag be removed or does that fact that at one point in time 1000 Pixiv users liked that particular image enough support for an image to keep the tag?
There's also the problem of images going bad id. When one of these posts is deleted from Pixiv, as some already have been, then there's no way for us to know if it really was a 1000 users post.
glasnost said: In retrospect, this tag is not even slightly objective, and I have no idea why I thought it was in my first post.
This will steer away from the matter at hand, but I'll entertain you with an answer.
Anyone who is familiar with the Quantity and Quality basic philosophy will understand: when the quantity reaches a certain threshold, the quality change and vice versa. 1000 is not a small number in the context of liking something. When we gather that much opinion, it changes the individual taste to public opinion, i.e. subjective to objective.
To test this, just try to pick up any post in the tag that you honestly believe is mediocre or worse.
Log said: Tag subscriptions simply will not work, subscriptions inject the post into your subscriptions based on when the post was posted on danbooru not when the tag/pool was added. Seeing as how very few if any posts will be posted after they receive 1000 favorites on pixiv this will not help anyone find newly 1000ed posts.
This... is a very good point. I can see the advantage of making it a pool now.
rantuyetmai said: Anyone who is familiar with the Quantity and Quality basic philosophy will understand: when the quantity reaches a certain threshold, the quality change and vice versa. 1000 is not a small number in the context of liking something. When we gather that much opinion, it changes the individual taste to public opinion, i.e. subjective to objective.
40% of Americans do not believe in evolution. Does that mean evolution doesn't work in America? Public opinion is not objective.
It's also funny that you mention basic philosophy, because the millenia-old sorites paradox suggests the opposite of what you're trying to claim.
rantuyetmai said: To test this, just try to pick up any post in the tag that you honestly believe is mediocre or worse.
rantuyetmai said: This... is a very good point. I can see the advantage of making it a pool now.
I didn't mention this is in PMs because any underused tag has the same issue, though to a lesser degree somewhat.
Log said: Tag subscriptions simply will not work, subscriptions inject the post into your subscriptions based on when the post was posted on danbooru not when the tag/pool was added. Seeing as how very few if any posts will be posted after they receive 1000 favorites on pixiv this will not help anyone find newly 1000ed posts.
It needs to be paired up with tags that don't get updated too frequently (such as certain artists) or not overly popular, or put alone in a separate group (all of this obviously being troublesome, I reckon). Most recents posts with the tag only are a few days old, so the subscription will still evolve.
Bad example on the Creation belief. Are you trying to compare artistic taste with scientific discoveries? If the whole world suddenly favor scat pictures, I don't necessary have to accept the fact that scat is appealing, but base on the context I can see the reason why it is liked. Sciences do not work that way I'm afraid.
I see nothing in sorites paradox that suggests the opposite of quantity theory. What it does is presenting the problem of "how much is enough" within the latter. Obviously human have to set out numbers for the things they do, because without that there are nothing to base on, crying "but that's not exact!" does not help.
The post you pointed out was "100" instead of "1000", I made a mistake.
I don't really see what's up for debate. It's not our job to decide what people should consider valuable metadata. Presence or lack of subjective valuation is not a metric for validity in a semantic system.
The existence of a group of users that use this is the primary metric that matters when considering its worth as a member of our ontology. As we have that, the only question that should remain for consideration is how we represent it.
