Due to a recent conflict regarding the disqualification for wrong romanization of an Artist, I feel like we should reconsider the standartization rule within the Ranking Criteria.

The RC states that used sources must be as official as possible, putting their priority above everything osu-related. Therefore, standartization for the romanization of the Artist's name should only be enforced if the artist himself does NOT provide an official romanization.
We try to credit and respect the musician's work as much as we can, but we fail to respect their "wish" as to what they want to be called.

The recent example I mentioned at the start refers to the artist "Keiichi Suzuki", which is printed on all physical products and is present on all websites.
http://www.keiichisuzuki.com/profile/
https://puu.sh/G13FO/d83647155c.png
https://imgur.com/a/ox4s5

I propose a change to the RC:
Current:
"Artist names are to be romanised in the order they are printed in the unicode fields."

To:
"Artist names are to be romanised in the order they are printed in the unicode fields. If the artist provides an official romanization, it must be used regardless of the standartization."

--------------------------------------------------------------------------------------------------------

#2 Current:
"If an artist has provided an official translation for their name, this may be used in the romanised artist field. Official romanisation may be used for the spelling of an artist's name, but the name order must follow the related rule."

^Get rid of the last part regarding the name order.

Official sources should always be the top priority and should stand above everything RC related, when it comes to metadata. If an artist provides an official romanization, a change regarding the order should not be enforced, as the artist's decision should be followed and respected in order to give them proper credit.

Feel free to disagree with this suggestion in the comments. A healthy discussion is always appreciated, as long as everyone stays civil!

I wholeheartedly support this, official romanization should be top priority, because this is how the original artist intended it to be, and we have to respect that.

I completely agree. If someone tells you call them by such and such name, you should respect that and use it. This applies to official sources as well. Official sources should be prioritized over standardization.

Edit: changed phrasing to prevent confusion

Agreed. Official romanization should be used regardless of the rule, makes more sense to do so.

had to read this several times but that certaintly makes much more sense applying changes to this rule. I'd go for it.

Hello

I disagree

- Many artists have mixed sources which in the past lead to flip flopping names. Which is given? Which is family? If it's inconsistent everywhere between artists or for ONE artist this can lead to lack of clarity and certainty.

Most large commercial products like games, movies, anime, will also follow a set standard rather than doing it person per person per person.

You may argue osu is different because it's not commercial / licensed, but the issue of clarity and consistency is still the same thing being addressed.

- Many artists also have obscure sources for "preferred romanization" such as really old album covers or webarchived talent agency listings that people would use. Some artists would be "obvious" via their website yes, others would not be. The extent of research required to verify accuracy would vary wildly.

- This all basically falls into the same issue as tv size where it often required way too much research to be 100% certain a specific formatting was correct, leading to a ton of inconsistency as "proper" sources are found later on because even with extensive research and finding all information you thought was available, there could still be yet more information still buried.

- Not even to mention the issue where the artist does it one way then changes it later. Resulting in changing the romanisation even though their actual name did not.

- Standardization makes it so it's done just 1 way on osu!, and there is much less room for error. It's not as though we are changing their names, we are following their name as it is in its original language.


and this is all before we even get into political issues surrounding the matter, as that's a mess too. Japan used to use Given Family for romanisation on legal and/or government documents, but relatively recently switched to Family, Given.

There's many reasons an artist may do one way or the other, some is for accessibility to western audiences, some were to fit the old government thing, etc..



Rather than enforcing a specific order for names on osu!, we chose to follow the Unicode field order. This increases consistency while still honoring their original name.

@Noffy

I don't see how the majority of this is problematic, as most of it already is the case for other things regarding metadata, which is not kept consistent.

Artist change their metadata sometimes, that's just how it is. Be it the title of a song, or source changing. How is this any different from romanization?

"Some artists would be "obvious" via their website yes, others would not be. The extent of research required to verify accuracy would vary wildly." - Why isn't it something optional then? If some people are willing to put in the time to this research, it should be ok. It could be handled like tags, some of them being optional.

Also, it doesn't really take much research. Not more than the romanization itself. If you check the tracklist, the album cover, the artist's website, any prints, you will most likely find what you need. This being a rule just doesn't really sound right to me.

"Standardization makes it so it's done just 1 way on osu!, and there is much less room for error. It's not as though we are changing their names, we are following their name as it is in its original language." -It doesn't change the search result anyways, so even if there would be 2 different orders, it would still show up the same, or am I missing something?

i agree, as for keiichi suzuki, someone out there will probably search up his name in japanese, so i think it'll be good to include his japanese name (like in kanji) in the tags

So,

"Primary metadata sources must be used as references for metadata. Do not modify metadata from primary sources except to comply with formatting and standardisation rules below. If no sources are available, use what is most common and recognizable."

Why would we even go as far as to manipulate metadata the comply with our own self-claimed correct standards, when we are throwing what has been intended by the released product overboard? Why are we being forced to change metadata to be technically incorrect, simply because some people aren't too keen on doing research until they have actual confirmation? Why are we getting rid of the initial purpose of metadata RC, to be "as accurate as possible"?

It would only be a fair compromise to give people at the very least the option to use something like album cover scans, or other physical / otherwise official products. As it stands now, we are giving ourselves the "right" to mess with that and change things around to our own liking. Which is, frankly, an absolute arse move to do, and the only thing I'm getting from that is that morale is missing.

Variations among different products and sources may very well happen. So what? If that's what is established by the product, what's wrong with complying with that?

I get the idea of making things easier for mappers by just following the ol' order of the unicode name. Can spare you some tedious research. If someone however does find an official source which conflicts with that given rule, should we really make their effort obsolete? Should we make the released source obsolete? In all honesty, I don't see how we are in the position to just do that, it's rude and completely goes against the very premise of the top rule of metadata. Inferno was another case that was outright ridiculous, because of how the mapper got forced to slap (TV Size) onto the full song, which kills the very implication the vast community gets from that very format, but that's a different story.

Which leads me to the following:

Artist names are to be romanised in the order they are printed in the unicode fields. If an official romanization is provided, it may/should be used instead


If nothing is found, just go with the order as given by the unicode field. If a source is given, that takes their place. All we're asking for is some accuracy in metadata again.


tl;dr I wholeheartedly agree with the proposal, the accuracy it'd give us would make much more sense with what we have currently

+1 to noffy's post (i agree with the points she made)

i disagree with this proposal because even if it's "optional", where is the line drawn for that? in more than one case it has been seen that artists can write their name in varying orders, depending on where you are looking. (ie say anime or agency site vs their own site being different. they would both be official because they own their site and they work under an agency. so which one is right?)

standardization is meant to give at least some level of consistency (it is not perfect, of course). but how many conflicting and confusing cases would this cause by being "optional" is my concern.

Other than the fact that this proposal has already been done before and had been met with disagreements there, I agree with Noffy. The political argument is something I never considered but is one I wholeheartedly agree with especially considering that Japanese names are a huge part of their culture. This rule also simplifies things by a shit ton and doesn't change search results as Capu said.

In pishifat's words: 'metadata isn't supposed to align perfectly with an artist's preferences for the sake of on-platform standarization. i understand why that idea can be unappealing, but that's how metadata is going to be handled for the time being'

While I'm still not too fond of the way it is handled currently, the political aspect and the names playing a big role in japanese culture is something we should also account for, I can live with that reasoning as to why it is the way it is.

+1 for strong disagree with this option.

Places that list romanised words are often not the same source as the unicode source which significantly increases the amount of work that needs to be done for metadata.

Seto Koosuke and Seto Kousuke are both completely valid romanisations under different methods. Each individual metadata site picks one method which is fine in the context of just their site, but we have to aggregate multiple sources together. Having different romanisation methods in the same context causes confusion and just sucks.


It'd be like the osu wiki allowing both American and British english depending on who wrote the article, it's set to British to stay consistent.

i agree with the proposal.

eiri- wrote:

in more than one case it has been seen that artists can write their name in varying orders, depending on where you are looking. (ie say anime or agency site vs their own site being different. they would both be official because they own their site and they work under an agency. so which one is right?)

That's the case where romanization rule apply. If there are multiple source stating artist's name in different order, it should be romanized in the order of the unicode field. But also in more than one case, the artists only write their name in one order, and that's the one that should be followed.

eiri- wrote:

standardization is meant to give at least some level of consistency (it is not perfect, of course). but how many conflicting and confusing cases would this cause by being "optional" is my concern.

the standardization itself is inconsistent. they allowed to write "Aimi x Sakura Ayane x Maeshima Ami x Aiba Aina x Itou Miku" as "Aimi, Sakura Ayane, Maeshima Ami, Aiba Aina, Itou Miku" and both are considered correct. Why not allow us to write the romanization in the artist's preferred order? both are correct too.

if i had to choose i would definitely prefer to respect official romanization over doing standardisation. would be very helpful for artists like Yoko Shimomura, Koji Kondo, Nobuo Uematsu, Keiichi Suzuki and wouldn't even affect other japanese artists as they would just have their romanization standardised when no official romanization is available.

but to be honest i don't see this rule being applied if it's completely going against what the metadata gatekeepers want, rewording the proposal to have both romanization methods acceptable (moving one method to allowances or something) is the way to go in my opinion.

Sinnoh wrote:

Seto Koosuke and Seto Kousuke are both completely valid romanisations under different methods.

It's not about different romanizations, it's about the order in which the romanization is written.

adding to my previous post supporting both methods being acceptable:

more acceptable formats = less metadata dq's

is name order consistency really important? i think it's fair enough to expect consistency when dealing with songs that have been ranked before with correct metadata, but for new songs, as long as the names are at least spelled correctly they will always return the same relevant maps in beatmap search. doesn't take much brain cells to realize that "Keiichi Suzuki" and "Suzuki Keiichi" are the same person for example.

pimp wrote:

is name order consistency really important? i think it's fair enough to expect consistency when dealing with songs that have been ranked before with correct metadata, but for new songs, as long as the names are at least spelled correctly they will always return the same relevant maps in beatmap search. doesn't take much brain cells to realize that "Keiichi Suzuki" and "Suzuki Keiichi" are the same person for example.

i don't know. 安野希世乃 is technically Yasuno Kiyono, but her name is actually Kiyono Yasuno.

for english artists and etc, i don't think someone would put Swift Taylor for the artist box if they were mapping a Taylor Swift song. so i think it can also depend on the culture, as japanese culture goes by family, given.

i definitely agree and disagree with this proposal. as noffy said, there's mix ups. for example, in the artist box, would you put Takagi, or Takagi-san? because both are technically correct (Takagi-San being apart of the anime name). BUT then 高木 says Takagi, and only Takagi. BUT then at the same time, the anime is called Karakai Jouzu no Takagi-San.

and there's tons of other examples. heck, even with my username, who is based off of the character Li Syaoran, but he literally is translated into so many other names, like Shaoran, Xiaolang, etc etc

yes, i was referring to japanese artists ofc. Taylor Swift and other non-asian artists are not affected by this proposal as they don't use romanised artist field.

"-San" is a japanese honorific, not a given name or family name. this doesn't seem to be related to the proposal as i never seen any artist having a japanese honorific in their artisic name, and even if some actually have, i don't think any official metadata source would romanize a japanese honorific before the actual name.

the way your username would be romanised also doesn't seem to be related to this proposal since you are refering to spelling rather than name order.

here's a few sets that use Takagi-San instead of Takagi for the artist box:

beatmapsets/751065
https://osu.ppy.sh/beatmapsets/761718
beatmapsets/749297
https://osu.ppy.sh/beatmapsets/938468
beatmapsets/755289
https://osu.ppy.sh/beatmapsets/1004335
beatmapsets/784563
https://osu.ppy.sh/beatmapsets/718691
beatmapsets/894653

none of the ranked sets use Takagi in the artist box.

they will not credit this character as "Takagi" if that's not how they are written in official metadata sources.
in the few official links i found on these maps they are credited as "高木さん" and the correct romanization is what they are using already.

again, this proposal is about using official romanization vs using standardized romanization. your example doesn't have anything to do with this proposal from what i can tell.

"respecting the artist's intention" in all cases is a pointless obsession imo. simple standardisations like this make sense for easy understanding by osu! players with no downside to search-ability or identification. see the forums that specialise in databasing music which have (based on the ones I know) unanimously concluded that metadata's a giant mess when you always attempt to use exactly what the artist gives you. pishifat says the intention well at community/forums/posts/7087013

I'd be willing to bet most of the affected artists didn't even consider this on purpose, and the "respect" you think it'd be showing is only in your interpretation of how the artist presented their name. Japanese names are sometimes romanised as First Last just to make sense for western distribution

clayton wrote:

"respecting the artist's intention" in all cases is a pointless obsession imo. simple standardisations like this make sense for easy understanding by osu! players with no downside to search-ability or identification. see the forums that specialise in databasing music which have (based on the ones I know) unanimously concluded that metadata's a giant mess when you always attempt to use exactly what the artist gives you. pishifat says the intention well at community/forums/posts/7087013

I'd be willing to bet most of the affected artists didn't even consider this on purpose, and the "respect" you think it'd be showing is only in your interpretation of how the artist presented their name. Japanese names are sometimes romanised as First Last just to make sense for western distribution

i mean i am for and against this.

it just, really depends ig. i don't expect people to put コシュニエ when it's actually Cö shu Nie

there's no truly correct answer to this issue since it's up to interpretation (do you find the artist's preference more important or is standardization more important?)

for osu, standardization is arguably more important because of the scale of the songs listing. adapting the romanized artist field specifically for each artist leads to inconsistencies in the listing (eg sorted by artist, "Suzuki Keiichi" and "Keiichi Suzuki" aren't grouped together). preferring the artist's choice sucks because it leads to inconsistencies when sources vary and allowing both sucks because it intentionally allows inconsistencies. standardizing based on unicode artist aims to avoid this

some other responses here linked it, but just to reiterate: community/forums/topics/883238?start=7087013 is still applicable

The sorting argument makes me sad cuz it's true. I've never thought about that before