Posted under General
Actually I'd like 'oh' for おお/'ou' for おう. It's still consistent, and it avoids the problem where English speakers can't pronounce 'too' without special training (and since Rin lives in London this is surely an important problem). Of course, this is just for character names, nobody even wants to think about how place names are romanized.
...and I just looked and ufotable officially uses "oh" for おう in etoh_mei, but it's not like anyone remembers that series?
I don't think the mispronunciations of casual/non-Japanese-inclined users/fans is relevant to this discussion for either "side", seeing as how most people can't even pronounce "manga".
I dislike 'oh' in most cases, but as long as we don't use it for おう I'm not going to get myself worked up over it. I'm sure we'll discuss this (again) and reverse it (again) six months from now anyway.
(And Manabi Straight is awesome)
albert said:
goes against the current "use whatever the copyright owners use" rule.
That's big news! I had no idea there was such a rule.
The aspirations of the Oono Kanako legions will be crushed.
Wait, they *want* their heroine to be called "Oh no!"?
Shinjidude said:
I don't think I've expressed an opinion on this before. But I'm very much in favor of the "oh" romanization of "おお". It's consistent (with a common scheme), in popular use, and avoids the issue where "oo" is typically pronounced "u" in english rather than "o:".
That point is itself inconsistent. Every single vowel is pronounced wrong in English! We'd have to mangle Japanese names thoroughly if we were to attempt to follow how an English speaker would instinctively pronounce them, and we still couldn't get him to say them right. What's the point of singling "oo" out?
Not to mention azumanga_daioo looks absolutely retarded.
王 would still be romanized as "ou", not as "oo". Besides, you don't seem to worry about how silly a name like "Shinjidude" would look to the general public. Even if we were to use "daioo", you'd get used to it pretty quickly.
albert said:
Maybe they are. The fact that oo is used in several tags already is another legitimate counterargument. But what's been done has been done and I'm not really in the mood to revert it based on a whim that goes against the current "use whatever the copyright owners use" rule.
You must be new here. We have discussed that over and over again, and every time we have agreed to use tags that are *correct*, as opposed to "official". We have already decided to use toosaka_rin, which is why it was in place, as opposed to the earlier tohsaka_rin.
There's no point in half-assing this. Give me a list of every *oh* tag we need to fix. I will need to have other JP speakers verify every one.
Okay.
You also need to give me an all encompassing list of romanization rules. Saying "Hepburn" isn't enough; Hepburn has a dozen variants. If the majority of Danbooru signs off on it, then I will revert tohsaka_rin back to toosaka_rin. The rivers will run red with blood of Toono Akiha fanboys committing suicide. The aspirations of the Oono Kanako legions will be crushed. But you will have your way.
You mean like the ones in forum #12961? Will do. But I'd like you to note that the community has already pretty much agreed to that, as reading said thread reveals. And I'd be careful with saying "majority", the majority of danbooru users are drooling morons who can't see past the first vag○○○ in a pic, I wouldn't really ask them for informed opinions.
And lastly, WTF? It was memento_mori again? Why does he do that all the time? Was the super_deformed alias not enough damage yet?
LaC said:
Every single vowel is pronounced wrong in English! We'd have to mangle Japanese names thoroughly if we were to attempt to follow how an English speaker would instinctively pronounce them, and we still couldn't get him to say them right. What's the point of singling "oo" out?
I assume you mean every Japanese vowel is pronounced wrong in English. In actuality though the problem is that there English has multiple orthographic conventions for each vowel, as well as often multiple pronunciations for each written convention. In at least some context, every one of the vowels the conventional romanization scheme is correct in English orthography (provided you are willing to treat [ɯ] as ). That is with the exception of "oo". Since English doesn't convey length distinction, there is no natural reason to have such an orthographic convention of "oo" -> [o:].
王 would still be romanized as "ou", not as "oo".
Damn, you're right. I just picked on "oh" example that would sound silly as "oo" without thinking of the underlying kanji. It is technically bad to be using "oh" for おう even under the scheme I was referencing.
Besides, you don't seem to worry about how silly a name like "Shinjidude" would look to the general public.
I think I've explained before that my alias was meant to be a throw-away name that I never gave much thought to. If I'd known I'd be using it as much as I am, I'd probably have picked a different one. But meh, I guess it's established at this point.
In the end, I guess I can't defend my position completely logically except to say that the おう->"ou", おお->"oh" version of Hepburn is internally consistent and in much greater use than おお -> "oo". As such it preserves the most commonly used romanizations for affected tags.
Bleh, Azumanga Daioh, why must you be such an example of bad romanization?
In seriousness, if 葉月 makes a list, I'll volunteer to check through it :P Honestly I find romanization distasteful in general, and am kind of inclined to start typing romaji as I would in IME (what is known as "waapuro-shiki"). Or should I say, "start typing ro-mazi in wa-purosiki"? Yes, I probably should.
But anyway, as many of our tags are also in English, there's a bit more complexity here than just "what the most sensible way to romanize Japanese is". It's more a question of "how does one most sensibly represent Japanese on an English website". However, finally, I agree with LaC - English orthography is so fucked up that I would never recommend throwing a romanization scheme into disarray just to accomodate English "phonics".
Japanese is a pretty much monophthonic language, so to speak, so we can usually convince ourselves that we're romanizing "phonetically" by using the corresponding digraphs (with a few trigraphs like "shi" or "tsu") for the various kana, and call it "sensible". But why not just make the romanization correspond to, say, the kana that Japanese speakers actually use? I don't really mind writing "ji" instead of "zi" or "ju" instead of "dyu" (as haduki, er, hazuki insists), but ideally (imo) we would even go that far.
In summary, orthography being consistent across various representations of a language is more important to me than orthography reflecting phonetics (or English phonics) accurately. A failure in the second can easily be overcome by reading a short wikipedia-style article, whereas the first continues to be annoying in perpetuity.
tl;dr I'm with LaC and 葉月 here. Ideally romanization should come close to reflecting kana. On an English site, allowances can be made, but tohsaka_rin goes too far, IMO.
Well, you're not *really* with me, as I hate nihon-shiki and friends with unrelenting hatred :). IMHO "as close to kana as possible" is a rather worthless feature of a romanisation scheme, it should be fairly close (so as to give some reasonable clue about the spelling), but not so much so as to sacrifice more important characteristics, such as being non-ugly and consistent about pronunciation clues[*] (that's pretty much the only point of using a romanisation scheme; if you want to be really close the original JP text, you should just use the original text and stop being a pussy). Still, I agree with the general principle of being consistent and close to the original when possible.
[*] Mind you, that's consistent within itself, not with whatever pre-existing ortography the reader might be familiar with. You shouldn't insult it gratuitously (which is why "shi" and "wa/wo" are a good idea), but don't try too hard to stick with it where things don't match; English isn't the only language anyway and speakers of many languages will be using your scheme.
I don't really care as long as we don't end up using retardedness like Okochi.
memento_mori said:...and I just looked and ufotable officially uses "oh" for おう in etoh_mei, but it's not like anyone remembers that series?
Her character CD used Etou though (post #107224).
T5J8F8 said:
I still insist on pronouncing "Danbooru" as "daanbuuruu" (would've typed "dahn", but that sounds like an impeded "dawn").
So you pronounce oocyte as uusait? And Oort cloud as uurt claud?!
But thanks for giving me the best argument ever. To the "oh" camp: you will give up your faggotry until the site itself is renamed to danbohru. Ketsuron.
葉月 said:
IMHO "as close to kana as possible" is a rather worthless feature of a romanisation scheme, it should be fairly close (so as to give some reasonable clue about the spelling), but not so much so as to sacrifice more important characteristics, such as being non-ugly and consistent about pronunciation clues[*] (that's pretty much the only point of using a romanisation scheme; if you want to be really close the original JP text, you should just use the original text and stop being a pussy). Still, I agree with the general principle of being consistent and close to the original when possible.
I disagree - the main reason for romanization is not to give pronunciation clues, but to give a standard representation for a language in the Roman alphabet, because speakers of many languages can instantly read latin characters, unlike, say, kanji/kana. For example, I could easily remember how to spell a Polish word without having any idea how to pronounce it, whereas remembering what different kanji look like is something I find more challenging.
It does help that orthography in European languages is vaguely consistent (a is usually close to /a/, o is usually close to /o/, for example), but look at how for example hanyu pinyin uses the letters x, c, q, etc. and diacritics in rather unintuitive ways (to a speaker of a European language) but to great effect. Of course, it then goes on to do some weird shit with the finals, but hey.
"Consistency within itself" is easily achieved by making sure that spelling follows pronunciation. Other than that, I'm not sure what "consistency within itself" really means, other than that there should be an easy algorithm for spelling a word and for deciphering one.
Consistency with kana not only helps students of Japanese in the transition to using kana and kanji, but it also doesn't conflict with "following the pronunciation". There may be two ways to spell the same pronunciation (づ and ず, for example), but there aren't two ways to pronounce the same grapheme, as there are in some other languages I could name (c in English, for example). This means that if we follow kana, we are not losing any information. Furthermore, there's a nice mapping from kana to latin digraphs: there is a canonical "kana table" which has rows and columns, and it is quite natural, orthographically, to assign digraphs to its entries by mapping the column to the first letter (the "consonant") and the row to the second. One column in the kana table, 「たちつてと」, becomes "ta ti tu te to" in ローマ字. Compare with Hepburn "ta chi tsu te to". Sure, the latter looks like how we'd spell the actual sounds in English, but I find it more "gratuitously insulting" and less "non-ugly" than the former (nihonsiki), to be honest. By the way, this table isn't in any way arbitrary either, e.g. the stem of a verb ending with つ will end with ち (育つ -> 育ち). The fact that pronunciation of the initial consonant in the column I gave differs from row to row is likely a historical morphosis.
Consistency with "general European standards for how certain latin letters should sound" I find arbitrary and a non-issue. As I said previously, any failure in the intuitiveness of the spelling in a romanization system can easily be overcome by the prospective user reading a short wikipedia-style article. ローマ字 shouldn't be a crutch for (and solely the domain of) people who can't be bothered to learn a few simple rules.
Not to throw more trash on this fire, but I'll note in passing that when Japanese people go to the bother of Romanising something, it tends to come out sort of..."willy-nilly" might be a good way to describe it; I wouldn't call it rare to see a write switch style mid-word.
As far as this in particular goes, yeah, the seemingly-random switch is a little annoying at some level. But if you disabuse yourself of the notion that an alias creates a modification and instead consider that an alias establishes an equivalence normalised for cruft-management's sake, it's like nothing ever changed.
I dislike Hepburn's style, myself; I prefer Kunrei, but unless you feel like writing a patch for bidirectional aliases and an option for Romanisation-style on the user's end, we're never going to be able to please everyone.
Or hey, there's always JSL-style!
DschingisKhan said:
I dislike Hepburn's style, myself; I prefer Kunrei, but unless you feel like writing a patch for bidirectional aliases and an option for Romanisation-style on the user's end, we're never going to be able to please everyone.
I was considering doing something like this, actually. Then I came back to my senses.
