Re: CJK Numbering in gnome-doc-utils

[Shaun McCance <shaunm gnome org>]

On Sat, 2005-12-10 at 10:40 +0800, Funda Wang wrote:
> Shaun> I don't understand at all.  Using the characters in [JA],
> Shaun> without any zeroes, I get 1002 = åä and 1200 = åäç.
> Shaun> No ambiguity.
> 1200 is speaking as äåä in oral language of Chinese, which will
> be confusing if we speak 1002 as äåä also. So, 1002 is written as
> äåãä, and 1020 is written as äåãäå.

So it's largely an artifact of how you speak the
numbers?  The reference page for Japanese seems
to indicate that ã is virtually never used for
Japanese numbers.

It might get a little tricky for me to get the
ã correct, since it seems multiple successive
ãs are collapsed.  But I'll see what I can do.

> Shaun> For instance, 1000 is å, but one
> Shaun> could also write it as äå.
> 1000 is always written as äå, while nobody writes it as a single åã

All right, is that the case for å, ç, and ä
as well?  Are there differences in this between
Japanese and Chinese numbering?

> Shaun> Furthermore, [ZH] also provides characters for 20, 30,
> Shaun> and 40, rather than just prefixing 10 with the correct
> Shaun> digit marker.
> This is wrong. å is only used in Chinese Lunar Calendar and some old
> articles. Nowadays, nobody uses å to represent 20, and there are a lot
> of people who don't know how to pronounce å correctly. So, 20 should be
> written as äå. The same applies to äå means 30.

Good.  This is how I have it implemented now,
and it's the considerably easier way to do it.

> Shaun> Please,
> Shaun> everybody send me details of the numbering systems you
> Shaun> need, and then double-check my results with the script.
> Do you need an complete result list of Chinese numbering system?

A comprehensive list would be difficult, since
I need to check the algorithm up to some pretty
big numbers.  A characteristic sampling would
be sufficient, but we'd really want it to test
key points.

I have checked test-numbers into gnome-doc-utils
CVS, under the i18n directory.  You can use this
to check the results of my implementations.  For
example:

[shaunm verso i18n] ./test-numbers japanese 103843 103850
åääååçååä
åääååçååå
åääååçååä
åääååçååå
åääååçååä
åääååçååå
åääååçååä
åääååçäå

This has been a lot of fun so far, so please
don't hold back on my account.  Hit me with
your hardest numbering problems. :)

--
Shaun