Posted under Tags
AngryZapdos said:
I do not support the "all or none" approach when it comes to consistency qualifiers. It is silly to qualify every single character from a copyright if only a small amount of characters are ambiguous enough for a qualifier.
This is the exact case for Genshin Impact.
Here is a rough count:
Tags linked from list of genshin impact characters: 110 (47 playable, or pets of playables)
Amount of (genshin_impact) qualifiers among those linked tags: 87 (38 playable, or pets of playables)
If my BURs at the start of this thread go through, the number of qualified chartags will drop to 44 (21 playable or pets.)
Vezral said:
I'm surprised these first few comments aren't enough grounds that adding a qualifier from the get-go just makes life easier.
Belatedly, I would opine that cicin ought to be moved to cicin_(artist). While they have a decent amount of Danbooru posts, their most recent illustration is from 2015 on both pixiv and Danbooru. I doubt they will be commonly searched for or tagged ever again. The cicin tag could be either used for Genshin Impact cicins (which would minimize the chance of mistagging, going forward) or left empty with a disambiguation wiki.
Updated
AngryZapdos said:
Did you actually read my reply? I mentioned more than once that I do not think all new character tags from this point on should be qualified. If a series is mostly comprised of characters with fairly unique names but with a few characters just called "Bob" or "Steve" or similar, then only Bob and Steve should get qualifiers.
I did read your reply, and I failed to understand your argument in terms of this situation with the Genshin Impact characters. Based on what you've said, I don't understand why you would disagree with the proposal to drop the qualifiers for all the aforementioned Genshin Impact characters. The majority of Genshin Impact characters do not need qualifiers. Why would you argue, for example, that we can't change "ganyu_(genshin_impact)" to just "ganyu"?
AngryZapdos said:
I mentioned several times that they should only be used if around half or more of the characters from a copyright need qualifiers, and I specifically made sure to repeat this so that people would be very aware I do not support the very black-and-white "all or none" approach.
It looks like you don't support the "all or none approach", but rather the "50%+" approach. That said, in the event that a particular character tips the balance of a particular copyright over 50% of a series having qualifiers, would you support retroactively adding qualifiers to all the characters? Do you support having to calculate the number of characters in any given series to measure whether 50% or more have qualifiers, and then make the decision? To me that sounds like a convoluted and arbitrary standard, and in terms of a black-and-white approach, it doesn't sound much different to me in terms of a switch being flipped.
The only approach that makes sense to me is to handle this on an individual, case-by-case basis. Simply: Is the character tag redundant or is it not?
These are *character* tags, not copyright tags. They should be evaluated on scale of the individual character, not based on the copyright they belong to. This kind of approach can handle issues both present and future.
Akiraka8 said:
Instead of:
Character: ganyu
Copyright: genshin_impactWe have:
Character: ganyu_(genshin_impact)
Copyright: genshin_impactThis is bad. It's not only a 100% redundancy in the tags by way of ugly underscores and parentheses (which need to be escaped in some circumstances, and can be a pain for users to type), and a large amount of bloat, but also a complete copy of the information from the copyright namespace into the character namespace. From a database design perspective, we need to ask why this is necessary. I do think this should to be kept to the logical minimum.
This example is only redundancy so long as no other Danbooru-relevant work has a character, location, weapon, vehicle, etc named Ganyu. And her name doesn't strike me as so unique for to make its recurrence in another work so unlikely we should discount it. And the same is true of the other Liyue characters with Chinese names. As Chinese-made video games and cartoons with anime-inspired aesthetics (in other words, the types likely to have fanart submitted to Danbooru) continue to become more common, the odds of any given Chinese name being used again constantly increases.
mongirlfan said:
Do we have any other chinese anime or game that is also based on chinese culture/aesthetics?
Yes, although they don't have the tag count Genshin does. Here are some anime examples:
Douluo Dalu
Mo Dao Zu Shi
Wan Sheng Jie
Ne Zha
Ling Long: Incarnation
Spiritpact
White Snake
Big Fish & Begonia
Fei Ren Zai
Da Yu Hai Tang
Aishen Qiaokeli Jinxingshi
Updated
Magus said:
This example is only redundancy so long as no other Danbooru-relevant work has a character, location, weapon, vehicle, etc named Ganyu.
Agreed! So by that standard, she doesn't need a qualifier, as is the same for the rest of the characters in the BUR.
her name doesn't strike me as so unique for to make its recurrence in another work so unlikely we should discount it. And the same is true of the other Liyue characters with Chinese names.
I personally don't see how another copyright would be as audacious as to use the name "甘雨" (ganyu) for one of their characters. That would be an obvious rip-off. It's not impossible, but I don't think it's likely. That goes for the rest of the cast in my opinion, though of course I can't be sure about all of them. But all that said, both of our opinions are subjective. I don't think it's good to use subjective criteria to determine how to tag if it can be avoided.
As Chinese-made video games and cartoons with anime-inspired aesthetics (in other words, the types likely to have fanart submitted to Danbooru) continue to become more common, the odds of any given Chinese name being used again constantly increases.
This is true, but I don't think it's cause for too much concern. Danbooru already has thousands of Japanese copyrights, so this reality applies to Japanese works as well. In spite of that, there are only a few notable situations in which we need to worry about qualifiers. (Just for fun I looked up the most common Japanese surnames (Satou, Suzuki, Takahashi, and Tanaka) to see how many characters with those surnames ended up needing qualifiers, and the number was quite low.)
Even if the scenario that you mention is more likely than I think, the process of adding more Chinese names to the site will still happen over the scope of years, and it's not going to be too hard to adjust to.
I understand that people aren't as used to dealing with Chinese names on Danbooru as much as Japanese names, but let's not throw the baby out with the bathwater or use "It's Chinese" as a reason to break from usual convention.
It's better to stick with a two-step process:
Is the canonical name duplicated?
Yes: Add a qualifier
No:
Does the canonical name belong to a list of categories that have reasonable grounds to need a qualifier? (as mentioned before, historical and mythological references, etc.)
Yes: add a qualifier
No:
Use the plain canonical name
This process should cover any character tags present and future.
Akiraka8 said:
It's better to stick with a two-step process:
Is the canonical name duplicated?
Yes: Add a qualifier
No:Does the canonical name belong to a list of categories that have reasonable grounds to need a qualifier? (as mentioned before, historical and mythological references, etc.)
Yes: add a qualifier
No:Use the plain canonical name
This process should cover any character tags present and future.
100% this, really, for exactly...
blindVigil said:
Sorry for what? We're not preventing a disaster, it'll be one BUR that'll get approved almost immediately and won't require any discourse or further thought.
... this reason. I really don't see the harm in keeping "possibly-maybe-eventually-perhaps-ambiguous-in-the-near-or-far-future-but-not-at-the-moment" character tags as unqualified and BURing them if the time ever comes that they become ambiguous. Especially so for characters that are completely original and not from history or this or that.
It's not like some great evil will befall all of Danbooru from having to BUR one chartag or two once in forever and we must safeguard ourselves by qualifying every chartag. To me qualifiers have always ever been about disambiguating when needed, and outside of egregious examples like Fate but really >fate tagging imo it should be kept this way, rather than stapling a qualifier onto chartags for "consistency" when 98% of those tags won't ever need disambiguation anyway (and if they do a quick and simple BUR wham bam approved fixes it without problem.)
Akiraka8 said:
This is true, but I don't think it's cause for too much concern. Danbooru already has thousands of Japanese copyrights, so this reality applies to Japanese works as well. In spite of that, there are only a few notable situations in which we need to worry about qualifiers. (Just for fun I looked up the most common Japanese surnames (Satou, Suzuki, Takahashi, and Tanaka) to see how many characters with those surnames ended up needing qualifiers, and the number was quite low.)
Qualifiers are very rarely needed when a character has both a given name and a surname. But if they only have one name, it's very easy for the name to end up getting repeated in other works. Especially if it's a relatively generic name.
Magus said:
Qualifiers are very rarely needed when a character has both a given name and a surname. But if they only have one name, it's very easy for the name to end up getting repeated in other works. Especially if it's a relatively generic name.
But we're not talking about generic names, and there are loads of single name characters on site who don't have and have never needed a qualifier. When another character named Zhongli appears on Danbooru, then we can talk about "needing" a qualifier.
Single word names are not inherently generic or ambiguous just because they're a single name.
Magus said:
Qualifiers are very rarely needed when a character has both a given name and a surname.
I did overlook this, and so it wasn't a great example. But I think the point still stands that whether or not a name has one or two words isn't actually the point behind qualifiers. It's whether or not the name is duplicated. Also, there are provisions for generic names and a few other cases to get qualifiers in advance as well.
The only thing I can really say more to this is that I think the thumbs-up side looks to all be arguing for the same idea and also has the same idea of how to accomplish it (correct me if I'm wrong). However, while the thumbs-down side talks about consistency, each person seems to have a different idea of what to do and why it's a problem. There's been the "Use preemptive qualifiers to avoid them in the future" argument. There's also the "Each series has to be consistent among itself" argument. There was the "A series should have qualifiers if over half the characters need them" argument. There was also an "It makes it easier to see which copyright a character comes from" argument.
The problem is, that's too inconsistent, and so far it doesn't have a clear process or policy for adding qualifiers, so it would be a mess in the future.
But the thumbs-up solution is very simple: Just give a tag a qualifier if it *needs* one, and follow simple steps to determine whether or not that's the case.
I am sure this BUR will pass anyway since it's aesthetically pleasing, but let me just say it's crazy how people prefer to have BUR bureaucracy over a solution where BUR won't be needed for characters, ever.
Akiraka8 said:
However, while the thumbs-down side talks about consistency, each person seems to have a different idea of what to do and why it's a problem. There's been the "Use preemptive qualifiers to avoid them in the future" argument. There's also the "Each series has to be consistent among itself" argument. There was the "A series should have qualifiers if over half the characters need them" argument. There was also an "It makes it easier to see which copyright a character comes from" argument.
The problem is, that's too inconsistent, and so far it doesn't have a clear process or policy for adding qualifiers, so it would be a mess in the future.
But the thumbs-up solution is very simple: Just give a tag a qualifier if it *needs* one, and follow simple steps to determine whether or not that's the case.
I felt the need to point out the ridiculousness of this comment, so I'll be making a brazen overgeneralization: the downvoters are simply saying that we should qualify every character tag. That's. It.
Those "arguments" that you put in "quotes" are the pros of doing so, not how we should deal with future cases.
Basically, imagine a Danbooru worldline where all character tags have qualifiers and its implications:
Vezral said:
- "There won't be a need for character BUR" (new!)
Except multiple Genshin characters already proved this wrong. Ganyu needed a BUR to alias Ganyu -> Ganyu (Genshin Impact) because, as it turns out, giving a new, very popular character with a completely unique name an unnecessary qualifier results in hundreds of missed searches. We created problems for our users and more work for ourselves. In the end, we guaranteed the necessity of a BUR by giving her a qualifier, when leaving her without a qualifier might not have ever needed a BUR at all.
Everyone keeps talking about tagging efficiency, which I'm still of the mind this approach barely improves, but no one seems to care about searching efficiency, either. Unnecessary qualifiers directly hinder search efforts, as well as forcing us to make a BUR just to fix the problem we created.
Oh, and now, if a second Ganyu should ever actually appear on this site, we'll then need to do another BUR to remove that alias, so that Ganyu can be made into a disambiguation page. Two BURs total, for a character that never should have even needed one.
Updated
blindVigil said:
Everyone keeps talking about tagging efficiency, which I'm still of the mind this approach barely improves, but no one seems to care about searching efficiency, either. Unnecessary qualifiers directly hinder search efforts, as well as forcing us to make a BUR just to fix the problem we created.
I only use Danbooru on desktop and this might be a mobile-specific issue, but I can't imagine searching for anything on this site without using the tag autocomplete.
I can't search using "reimu". I can't search using "hakurei". I can't search using "hakurei reimu" either. I have to enter "hakurei_reimu" into the search bar for it to work.
Like wtf? How the hell is a new user supposed to know that they need to use exact term with underscores? And if you need exact term to search for anything, trying to search on Danbooru without autocomplete is basically suicidal.
Vezral said:
I only use Danbooru on desktop and this might be a mobile-specific issue, but I can't imagine searching for anything on this site without using the tag autocomplete.
I can't search using "reimu". I can't search using "hakurei". I can't search using "hakurei reimu" either. I have to enter "hakurei_reimu" into the search bar for it to work.
Like wtf? How the hell is a new user supposed to know that they need to use exact term with underscores? And if you need exact term to search for anything, trying to search on Danbooru without autocomplete is basically suicidal.
The vast majority of our external traffic comes from google, which uses the most common terms without qualifiers.
If you're the average internet user when you search for a character on google you type "ganyu art", not "ganyu (genshin impact) art".
Furthermore, if you land on danbooru by searching for generic anime art, and you've never used the site, you probably won't even use underscores at all. That's why whenever a new popular series is released we get an explosion of missed searches for just the first name of a character: it's new users getting here through google.
Updated
nonamethanks said:
The vast majority of our external traffic comes from google, which uses the most common terms without qualifiers.
Could you elaborate further on how this works? I tried to search for "ganyu danbooru" on google and the top three hits are:
If the concern is that danbooru won't come up when people google for "ganyu", then then there shouldn't even be a discussion. This BUR should just pass otherwise danbooru userbase would dry up.
