''Romaji, Hepburn''

[wahaha]

Well, the help-text for anime-titles mentiond that they should be written in Hepburn-romanization. This is, however, obviously not really working in conjunction with the idea to have the romaji titles in lower ASCII, when it comes to long vowels.

For a start, see wikipedia: Hepburn (features)

To get more clear about the modifications used for romaji, let me suggest to:

• literally transcribe long vowels (ああ=aa, おお=oo, おう=ou, ...)
• write the particle 「を」 as "wo"
• always write 「ん」 as "n" ("sempai" -> "senpai")
• For Katakana:
  • ... if the katakana represent a non-japanese word (that can be properly spelled in lower ASCII), transcribe it in the original spelling.

[Skywalka]

What is lower Ascii?

(sorry, I just thought that I am allowed one silly question too once in a while ^_^)

[wahaha]

Uh... I *knew* I forgot to get a link for it. But then again, "lower ascii" worked quite well

The Ascii table is split into two sections, the lower ascii table and the upper ascii table. The lower ascii table defines all numbers beteen 0 and 127 inclusive. This is the officially standardised section of the Ascii table and represents all the most common characters.
[...]
The upper Ascii table has not been officially standardised, and tends to vary from computer to computer and from font to font.

Basically, lower ASCII is what you can rely on really being the same everywhere (and thus of course: being supported by most programs/filesystems).

[Skywalka]

Ah, so I already understood you but had to make sure

You are talking about "ō" for instance, which is not part of lower Ascii, right? ^_^

[Skywalka]

[*]For Katakana:

• ... if the katakana represent a non-japanese word (that can be properly spelled in lower ASCII), transcribe it in the original spelling.

[/list]

Why? I love to read stuff like the japanese representation of "Düsseldorf" for instance ^_^ What was it again? Dosudorufu? ^____^

[wahaha]

You are talking about "ō" for instance, which is not part of lower Ascii, right?

exactly

Düsseldorf -> デュッセ ルドルフ* -> Dyusserudorufu*

*) probably... not sure, though

[gholovo]

If you are going to write the particle を as wo, then are you going to put the particle は　as ha? That might seem a bit odd to quite a few people. Ditto actually writing senpai.

Of course, I might just be unhappy because my creq on Sailor Moon isn't valid if we do this...

There are other forms of romanization, but all the standard ones require high ASCII for long vowels. I suppose we could make a new AniDB standard (as suggested), although that does sort of lose the benefits of using a standard.

And at last:

Why? I love to read stuff like the japanese representation of "Düsseldorf" for instance ^_^ What was it again? Dosudorufu? ^____^

I recall watching a fansub of either Stellvia or Planetes in which Lagrange points were subbed as Lagurangu points. Amusing, indeed.

[Skywalka]

I understand your point but I guess the reason for lower Ascii was to get working ed2k links no matter what presets you have for the filenames.

I know this might pose a problem but I think the reason AniDB was put in place was not to teach proper romanization of japanese writings

So I guess there need to be a couple of small changes to the usual rules to fit our purpose.

[wahaha]

If you are going to write the particle を as wo, then are you going to put the particle は　as ha? That might seem a bit odd to quite a few people.

I'd go with Hepburn here: wa if it's a particle, ha if it's not.

Ditto actually writing senpai.

Either way, there should be a rule for it. I'm in favor of using "n", because to me, it feels awkward to have a single "m" in romaji, while the pronounciation-shift could still be derived from seeing 'n' + next letter.

Of course, I might just be unhappy because my creq on Sailor Moon isn't valid if we do this...

Ah, now you gave away the cause of this thread (the creq itself - not to get an excuse to deny it)

In any case, there has to be a rule for ō vs. oh vs. oo/ou. So that we can avoid the chaos of having all different kinds of romanizations.

I suppose we could make a new AniDB standard (as suggested), although that does sort of lose the benefits of using a standard.

Well, it's not like I totally made this up. Though not exactly a reliable source for proper spelling or grammar, the suggestions are heavily influenced from fansubs. In my experience, these modifications of Hepburn are being used (in fansubs), just without having been written down.

[gholovo]

Either way, there should be a rule for it. I'm in favor of using "n", because to me, it feels awkward to have a single "m" in romaji, while the pronounciation-shift could still be derived from seeing 'n' + next letter.

It can be derived, but only if you already know that n + p = mp. Further, it's quite common practice to write 'sempai.'

In any case, there has to be a rule for ō vs. oh vs. oo/ou. So that we can avoid the chaos of having all different kinds of romanizations.

If you're going to modify it, I'd say to follow the kana oo = oo, ou = ou, uu =uu, and that should be an obvious way to write extended vowels.

In case it's unclear from what I've written, I do agree with you in that we should modify this and stick with lower ASCII. I just like to argue.

And, finally:

Ah, now you gave away the cause of this thread (the creq itself - not to get an excuse to deny it)

<outfit="superman">
Then I'll rush to the scene and revoke it before you get a chance to deny it!
</outfit>

[analogued]

• literally transcribe long vowels (ああ=aa, おお=oo, おう=ou, ...)
• write the particle 「を」 as "wo"
• always write 「ん」 as "n" ("sempai" -> "senpai")
• For Katakana:
  • ... if the katakana represent a non-japanese word (that can be properly spelled in lower ASCII), transcribe it in the original spelling.

I agree with wahaha on this... I was thinking along the same lines.

[analogued]

I think this (and a few other threads, like the one about release dates) should be stickied somewhere ... or, better yet, they should be summed up and added to the Anidb Documentation Forum

[wahaha]

Come to think of it, the "n" case isn't complete. What I wanted to aim at is the n<->m problem. However, there's also romanization like "しんいち" -> "Shin'ichi". Strictly following "write ん always as 'n'" would result in "Shinichi" though, which may mislead to think it was "しにち" in Japanese.

IMO, the " n' " as used in Hepburn is clearer in such cases, so I'd stick to it. Opinions?

[analogued]

Come to think of it, the "n" case isn't complete. What I wanted to aim at is the n<->m problem. However, there's also romanization like "しんいち" -> "Shin'ichi". Strictly following "write ん always as 'n'" would result in "Shinichi" though, which may mislead to think it was "しにち" in Japanese.

IMO, the " n' " as used in Hepburn is clearer in such cases, so I'd stick to it. Opinions?

You're right ... I was paying more attention to the first two bullets so I missed that. Indeed, the hepburn romanization in such cases is clearer. Thoigh I don't think there are a lot of cases, that's no reason not to do it.

Also, another thing... how should we write long vowels such as "をー"
The wikipedia page proposes a few methods... we obviously can't use a few of those since we have to use ASCII characters so we're left with:

1. o: not indicated at all
2. oo: doubled
3. oh: using an h
4. ou: using u

Which one of these should we use? For katakana this is obviously not the case since we should write the word intended in the original spelling. If that includes oo, ou or oh is clearly not a japanese language matter.

Also, just to make sure, you're saying we should use the Modified Hepburn (regarding the n <-> m problem) as described on the Wikipedia page.. right? I'm in favour of this... I just wanted to make sure.

[wahaha]

Also, another thing... how should we write long vowels such as "をー"

I don't know who added the "This romanization style is popular with fans of anime (otaku)"-passage to the wikipedia article, but I'd literally transcribe long vowels (おう=ou). As for the "ー", I'd just double the vowel, so "をー" -> "woo".

Also, just to make sure, you're saying we should use the Modified Hepburn (regarding the n <-> m problem) as described on the Wikipedia page.. right?

Exactly