W3C home > Mailing lists > Public > public-i18n-cjk@w3.org > October to December 2013

Re: [css-counter-styles] japanese-informal counter style

From: Kent Karlsson <kent.karlsson14@telia.com>
Date: Sat, 02 Nov 2013 20:14:16 +0100
To: "Tab Atkins Jr." <jackalmage@gmail.com>, Koji Ishii <kojiishi@gluesoft.co.jp>
CC: MURAKAMI Shinyu <murakami@antenna.co.jp>, www-style list <www-style@w3.org>, "CJK discussion (public-i18n-cjk@w3.org)" <public-i18n-cjk@w3.org>, George Rhoten <grhoten@apple.com>
Message-ID: <CE9B0D98.283CF%kent.karlsson14@telia.com>

Den 2013-11-02 00:37, skrev "Tab Atkins Jr." <jackalmage@gmail.com>:

> On Sat, Aug 17, 2013 at 5:25 AM, Koji Ishii <kojiishi@gluesoft.co.jp> wrote:
>>>> Maybe there are just different styles taught across the country?
>>> "一千" is sometimes used, especially "一千万円" (ten million yen) is often used
>>> rather than
>>> "千万円", but this case is exceptional and not required for japanese-informal
>>> style.
>>> 
>>> I tested Excel's [Format Cells - Number - 漢数字/Japanese] and got
>>> "一万千百十一" for 11111. (See the attached screen shot) I believe Excel's
>>> numbering is
>>> correct at least for Japanese styles.
>> 
>> "一千" is ambiguous for me, and can be counted as "different styles" as Tab
>> says. My preference varies by the numbers, and quick thinking is that, I
>> guess bad cases for not having it is worse than having it, such as:
>>   一万一千百十一 vs 一万千百十一: whichever is fine
>>   一千 vs 千: the latter is slightly better
>>   一千万 vs 千万: the former is much better
>>   一億一千百十一 vs 一億千百十一: the former is slightly better
>> 
>> Interestingly, Word and Excel disagrees on "千". I recently heard that ICU has
>> a function to format i18n numbers, does anybody know what ICU produces?

Try http://st.unicode.org/cldr-apps/numbers.jsp?locale=ja.
Open the Edit view, and you can change the values tested as well
as the rules. Please submit any proposed changes as CLDR tickets.

> Based on this, I'm leaving the rule as it stands, so that it produces
> a leading 一.  It seems that it's better to do so if we have to always
> choose one rule, even though in some circumstances it's fine or even
> preferred to drop it.

I would tend to agree that it is best in general (unless explicitly
dispreferred) to include the "one" multiplier explicitly, for all
"spellouts" in all locales; it is more formal and clearer. For "ja",
CLDR currently does that only for the "financial" variety.

    /Kent Karlsson

PS
CLDR also covers "hanidec" numbering, which is using CJK characters for
0-9 (one select variety) in a decimal-positional style. This does not go
via RBNF, though.

> ~TJ
> 
Received on Saturday, 2 November 2013 19:14:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:10:24 UTC