Regraz wrote:
Please at least give a look on the threads I listed above, I don't think dates are written in English there. :C They are all concuded before the UBKRC was founded. I don't think it is proper to abandon all the information these threads provide, just because there are Chinese characters, since some of them are written in English.
I know that these threads are older than the UBKRC. But those people still can join the discussion. Nobody is ignoring their opinions. As long as they: 1. Raise a concern 2. Give a reasoning. Unfortunately, majority of the Chinese community can't handle basics of English, so their communication with rest of the community is close to none. I'm not blaming these people, but if they want to be heard publicly, they shouldn't discuss these things behind closed doors without any communication with the team that is responsible for metadata. The threads without translation don't give us the concrete reasons for why this romanisation is bad or why this change overall is bad. We'd rather get the feedback here without the need to translate 2+ years old threads. There's enough time for everyone to give their opinion, if we had to translate and check every single metadata discussion, we couldn't possibly propose anything, ever.
It's extreme delay for us if we actually have to go through all the discussions that a detached group of people made. Please, work more with the community and say your opinion here. Now it's the best time!Regraz wrote:
Put aside your statement of "language tells its boundaries itself", which is not really convincible first. Let's assume the boundaries between language are clear. However, you need a boundary for the implementation of a rule in osu to deal with cases when these boundaries are trespassed. For example, a Chinese word might belong to different dialect, they look the same but differ in meaning and pronunciation. Since you stated in this proposal that, other languages within the Chinese language system are not considered. Then there will be potential issues that if you require mappers to utilize the method of romanization provided by this proposal, they will say that "The title belongs to Cantonese or other languages within the Chinese language system". How will you know if it is Mandarin or another language? If you continue to forcing other languages using the Mandarin romanisation system, that is already deviating the idea of your proposal of letting international people know about Chinese titles conveniently for they will be in different meanings and pronunciation after romanisation under your scheme. You have to tell in the rule clearly, instead of saying "it is a case-by-case" problem since it is related to quite a few Chinese words.
Excuse me, but you cannot just tell me to put aside my argument because it's not convincible and give yours. That's not a rational discussion. Same way, you could argue that we can never know if a language is Japanese, if it contains characters that are also Chinese and doesn't include Hiragana/Katakana. There always need to be research as to what language it really is and it is up to modding community and BNs to check that. As it goes for any other language, even though the words may be exactly the same, you have to figure out what language it really is. This is what we have to do when checking metadata even today. Respectively, we always had to do that. And yes, no matter how vague the difference between a language may be, it's up to research to figure out what language it is.
Regraz wrote:
Sadly it is your business. Different from Japanese, there is not a single standard/system of romanisation could cover all, or at least most of the cases. And the cases you cannot cover is myriad. If you pick up one, for example, the system mentioned in the thread, you have to make relevant rules regarding what you cannot cover, like the names from Taiwan.
What I'm saying is none of our business are political issues. These don't belong to this game and shouldn't be considered in any way (ignoring some extremely violent cases). If by choosing a Romanisation system, you are basically "picking side" in terms of Chinese politics, they you are "picking side" even by not choosing it, it works vice-versa. Romanisation systems don't have to be official in the country from which the language comes for it to be correct. It's intended primarily for users of Latin alphabet, so if it's a standard in (in fact majority) of countries that use Latin alphabet, the evidence seems to be that that's good enough, rather than a system that is only used within that country and is not respected by nearly any other country. Hanyu Pinyin seems to be built on a heavy amount of linguistic research and the reasons it is official in majority of countries using Latin alphabet seem to be pretty reasonable. Yes, if you want to Romanise Chinese Cantonese or Taiwanese Mandarin, or any dialect, you will pick a system that most suits the needs and that makes most sense with Latin alphabet. Now maybe 15% of Chinese/Taiwanese songs on osu! will be Cantonese (never said there shouldn't be system for Cantonese) or Hokkien, the rest will be Mandarin, creating a system for each dialect is not needed because there will be maybe 5 maps overall, insignificant languages will still be judged case-by-case.
As for the names from Taiwan, that will already be covered by the romanisation system you use for Taiwanese, so we don't have to include rules for that. We only have to include rules for what you have to do different than the system, because of some compatibility reasons.
Regraz wrote:
No matter what system you choose, and what reason you provide, you have to deal with practical issues. Here're some cases where spacing by word is hard to implemented as posted in p/5192373 , they are answered by Lyric in p/5192457 but they are not convincing as different people have different explanisation toward the spacing scheme:
Pata-Mon and re-write in English for ease reading wrote:
Word: 痛彻心扉
Problem: If 痛彻 could be treated as different words?
Word: 大侠艾吃汉堡包
Problem: This title could be romanized into both "Daxia'ai Chi Hanbaobao" and "Daxia Aichi Hanbaobao", they are both reasonable but different in meaning and pronunciation. Please do not say "It is not related to people who do not speak Chinese". Because you have to give an answer which one to use for the title of the map. If with the current character spacing scheme, no problems will arise: "Da Xia Ai Chi Han Bao Bao"
Word: 学不会
Problem: It should be romanized into Xuebuhui or Xue Buhui? Same reason as above.
and so on
"no problems will arise" is again such a strong claim without any thought or evidence.
So, what is suggested here is that we solve an actual issue of the language (sequence of characters having a 2 different meanings or 2 different Romanisations) by making it incorrect entirely. If the reasoning of an issue is that they have different meaning by Romanising them in 2 different ways, then in result, Romanising them syllable by syllable will yield the exact same Romanisation for both contexts/meanings. That doesn't only mean the problem is not fixed (the text still may not align with what the meaning is supposed to be), it also means introducing second problem (now, the meaning is completely undetectable). If you have 2 possible Romanisations and have to choose by a meaning, you will have to do your research on what the meaning is supposed to be (
we do that for Japanese too!). If you accidentally choose the wrong title, people will think the meaning is different than what it's supposed to be, but with a system that completely disregards this,
you won't know the meaning at any point. Romanising separate syllables is not solving the problem, it's skipping it, that's the problem with the original system.Regraz wrote:
It is of course everyone has the right to discuss as osu! is quite community-oriented. But romanization is realated to two languages, and requires detailed research and work regarding both languages. If people barely understand Chinese and are going to build a romanisation system possibly just after searching something in google and wiki, that is not authentic at all either.
I'm not at all saying that these people can't give their opinion, but they shouldn't be above anyone else. We are not building a Romanisation system, we are just using it. You can't go wrong about using a Romanisation system because it has a clear set of rules that you have to follow. Assuming that we're basing this just on wiki is not making it less authentic, and searching it on google is not really an argument, because that doesn't say how intense or good was the research.
What we are promoting is a system that indeed does differentiate between different Romanisations of the same text, based on meaning (because we do that for the other languages too, for a long time already) (ofc we can't cover everything because we can't use Unicode characters, but current system doesn't solve that either, in fact no system will) and we also promote a system that is correctly Romanising it with respect to the Latin-based languages. What the current system does is that it ignores the different meanings and also breaks the rules of Latin-based languages. What we suggest in the proposal solves both of these issues.
