[Nuxeo-localizer] Re: Translation of QuickStart in Japanese

[Kazuya FUKAMACHI]

Hi  David,

Thank you for answering my question.
 BTW, I changed my email address to one for mailing list use. 

> First, thanks a lot for the japanese translation of the Quick Start,

It's my pleasure.

> I'll add it to the web site. I'll open the Localizer section too, so
> you will be able to translate it directly. I'll tell you and the
> mailing list when it's done.

Thanks. I'm going to start translating other pages and also
Localizer's po file.

On Sun, 08 Sep 2002 15:23:04 +0200
Juan David Ibanez Palomar <address@hidden> wrote:
> The encoding of the PO file has no influence.

> You can load the page and then change the encoding. However, Localizer
> expects the data to be in UTF-8, bad things could happen otherwise.

> The charset specified in the Properties tab only is used to import/export
> the message catalog from/to PO files.

> But the data stored in the ZODB will always be Unicode strings. Unless you
> change it directly, with your own interfaces, instead of using the Localizer
> management screens.

Now,  I understand.

I sould better to consider  to use  UTF-8 encoding extensively.
It's a bit combersome subject, because encoding of pages may affects to
Zope Splitter and ZCatalog. If I choose UTF-8 for some page, I should
change whole site to UTF-8 to enable whole site search.
Maybe it's inevitable to use UTF-8 for i18n site.

Firstly, I should make a patch for JSplitter, Japanese Splitter, so as
to handle UTF-8, maybe. Korean and Chinese, ...  users should do
the same things for their Splitter, maybe.

FYI
Many of Japanese sites use EUC-JP and Shift_JIS for their page encoding.
And also  use ISO-2022-JP, mainly for sending e-mails.
(for example, we should patch MailHost  to send mails  ISO-2022-JP
 encoded  and also subject mime-encoded)

Thus, Japanese should handle a few encodings always.
It's troublesome, but we should tough through until most of  tools and
programs should handle UTF-8 and programmers familliar with it, also.
It takes time, maybe. Anyway, UTF-8 is far better than ASCII encoding.
Many programs only treat ASCII characters as literal characters.

When I  localize Zope product, first thing to do is grep -i iso-8859 and
also search such code,  "if ord(c) <= 31 or ord(c) >= 127:" (grin)
We, double byte character set users,  need 8bit through coding.

Regards,
Kazuya Fukamachi