Surely I know that, it's a Japanese DeviantArt after all. But within that 1000+ people there are artists as well, with much more higher percentage on Danbooru, and significantly less people who is only looking for porn.
Posted under General
rantuyetmai said:
Surely I know that, it's a Japanese DeviantArt after all. But within that 1000+ people there are artists as well, with much more higher percentage on Danbooru, and significantly less people who is only looking for porn.
That just seems to bring even more into question what makes the metric of 1000 favorites relevant.
rantuyetmai said:
Surely I know that, it's a Japanese DeviantArt after all. But within that 1000+ people there are artists as well, with much more higher percentage on Danbooru, and significantly less people who is only looking for porn.
So what? If someone is that interested in what the pixiv population thinks, they can go over there and find out. Why do we also need to record this data?
Makes me able to find well drawn Touhou posts with a low score.
Can be used as a tag subscription.
Danbooru is a repository of quality art and this tag is an efficient tool to find quality art. How is that not relevant to this site?
The actual issue is why it should be specific to Touhou instead of being global.
Even then, if one is ready to go through the hassle of finding every danbooru posts tagged 東方Project1000users入り on pixiv you're not going to stop them just because no one else can be bothered enough to do the same with other series.
Updated
+1 for making this tag a pool. I agree that it's useful, but it violates our no-meta-knowledge rule.
Surely making it a pool would satisfy both parties here? Pools aren't nearly as srs bsns as tags are, and the information would still be kept and searchable. What's the problem here?
Quite new with the tag and pool usage here, figure I would ask with this opportunity: what do I depend on to measure as no-meta-knowledge?
I frequent pixiv a lot before knowing Danbooru (and happy because it's English!), most of what I tag are basic characteristics of the character and what pixiv tag or artist comment said. If I tag these, would it be called meta-knowledge? I mean... unless the artist says so, I don't know some characters is a trap 0_0
Are we strict with tag=visual pool=knowledge rule?
Edit: What does negate search mean?
Updated
Saluki said:
I frequent pixiv a lot before knowing Danbooru (and happy because it's English!), most of what I tag are basic characteristics of the character and what pixiv tag or artist comment said. If I tag these, would it be called meta-knowledge? I mean... unless the artist says so, I don't know some characters is a trap 0_0
Tagging what an artist says is fine. Meta-knowlege is more like tagging random remis with vampire or random liliths with succubus while they're not doing anything that indicates either of those things.
Negate just means searching -whatever.
Log said:
Tagging what an artist says is fine. Meta-knowlege is more like tagging random remis with vampire or random liliths with succubus while they're not doing anything that indicates either of those things.Negate just means searching -whatever.
Ah I see, thanks! The negate search could be very useful.
Back to the other thing, so from what I understand, we can base on the written facts over on pixiv to tag? If that's true then a picture tagged with 東方Project1000users入り is a fact that can be used to tag right? Sorry if I'm missing something... not really reading the long discussion.
Saluki said:
Ah I see, thanks! The negate search could be very useful.Back to the other thing, so from what I understand, we can base on the written facts over on pixiv to tag? If that's true then a picture tagged with 東方Project1000users入り is a fact that can be used to tag right? Sorry if I'm missing something... not really reading the long discussion.
The problem is, it's a fact that's not evident from the picture itself. Tags should only reflect things you see in the picture, not things you know from external sources. Otherwise you could just as easily say "It's a fact that Remilia is canonically a vampire, and thus all posts featuring Remilia should have the vampire tag."
In terms of precedent, most of the various Pixiv collaboration projects (e.g. Fantasia, Shadow, War) have copy: tags. There are also some Pixiv memes that have pools.
My understanding is this bears more similarity to the latter than the former thus making it pool material. In the end, it's a fairly meaningless distinction anyway, as pools can be referenced as a metatag.
I did started out this collection with a pool, but turned toward a tag after 300 images because of several reasons:
Coconut said:
The problem is, it's a fact that's not evident from the picture itself. Tags should only reflect things you see in the picture, not things you know from external sources.
I'm not very convinced by your argument. For example, there are lots of personification picture which we would not know without reading the tag on pixiv. How do you feel about we turn that into a pool? It's meta-knowledge.
Updated
rantuyetmai said:
- A whole comic is qualified (e.g. pool #1751). Each page play an equal part, it's silly to put the whole thing in an other pool, but feel wrong to just put the first page in *shrug*
Perhaps you can link to the comic in the pool description, noting that all or the majority of the posts in the other pool apply.
rantuyetmai said:
I did started out this collection with a pool, but turned toward a tag after 300 images because of several reasons:
- Tag auto completion is much faster than putting "pool:xxx", just need to type a "t".
- Negate search ability. Quite a few artists have multiple qualified pictures (e.g. aka_ringo 39, sayori 27). I do -touhou_project_1000_user to take out the ones that had been tagged, and apply tag script to ones that hadn't.
- A whole comic is qualified (e.g. pool #1751). Each page play an equal part, it's silly to put the whole thing in an other pool, but feel wrong to just put the first page in *shrug*
The first two reasons are only for your own convenience, which, no offense, is not something this type of decision should be based on. As for the last, just adding the cover page should be sufficient.
rantuyetmai said:
I'm not very convinced by your argument. For example, there are lots of personification picture which we would not know without reading the tag on pixiv. How do you feel about we turn that into a pool? It's meta-knowledge.
It's certainly not a hard-and-fast rule; few rules on Danbooru are. We try to err on the side of usefulness. In the case of personification, usually that's the biggest focus of the picture, so not having it tagged would be much more problematic. There's also the issue of just how meta the information is. Most people who are familiar with the copyright of the picture would recognize it as such. But even those people would fail to recognize pictures as a 1000users post. You'd only know if you were also a Pixiv user, which is pretty damn meta.
I agree with others above that Pixiv users tend not to be the best judges of quality anyway, so the value of this information is shaky no matter how you put it. That said, I also agree that more metrics by which to judge an image increases the number of interesting images people can find. Having a pool for it sounds like a good compromise, even if it's slightly more inconvenient for you personally.
Valuable information or not, it's personal opinion and I don't want to convince people anymore. The fact that it's there is what made me tag it, i.e. non-subjective.
I just finished going through tagging all the old posts, so it's not inconvenient anymore. Although through this talk I want to understand where to draw the line choosing pool/tag. I've always under the impression that where as it is possible (objective, non-sequential), a tag is preferred.
I don't think we should be discussing whether this should be a tag or a pool, but whether this is worth keeping at all. Why the hell should we care how many favorites a pic had on pixiv or on DA or whatever site it came from?
It might be another way of finding Touhou pics lots of people like (which is completely useless IMHO, but that's just me) but it's not relevant to us. Danbooru's favorites and scores are set by its priv+ users' favorites and votes so they can cater to Danbooru users' tastes and preferences, and pixiv's favorites and scores are set by its users favorites and votes so they can cater to pixiv users' tastes and preferences.
If you want to know what people on pixiv vote and favorite, you go to pixiv. I also agree with Fencedude and Hinacle and the rest, if it's decided that this is worth keeping, a pool is better.
Fred1515 said:
It might be another way of finding Touhou pics lots of people like (which is completely useless IMHO, but that's just me) but it's not relevant to us.
You lost me. Finding good pictures is useless? What are you doing here?
Forgot to say: scores and favorites of Danbooru users are disregarded in assessing picture quality too, in case you don't know. But the fact that it's there is because there might be some use (the popular page, order:score, etc). I don't see why we can't make use of an other metric. You don't use it, let the other use it.
Updated