[Zope-dev] Zope 2.6.0 ZMI Problem for CJK(Collector 623) patch.

Heiichiro NAKAMURA nheiich@quantumfusion.com
Tue, 10 Dec 2002 12:52:06 -0800


On Tue, 10 Dec 2002 08:52:09 +0000
Toby Dickenson <tdickenson@geminidataloggers.com> wrote:

> On Tuesday 10 December 2002 12:03 am, Kazuya FUKAMACHI wrote:
> > On Mon, 09 Dec 2002 13:18:26 -0800
> >
> > http://www.etforecasts.com/pr/pr1202.htm
> >
> > At least 21.91% of Internet users are CJK users.
> > (Japan + China + South Korea + Taiwan)
> > Since using UTF-8 for their pages is still not in common
> > in these area, it is preferable to accomodate such workaround.
> 
> I dont follow the logic here. Using unicode in the server doesnt force you to 
> use utf8 in the browser.



I guess probably what he meant was:

   1) 21.91% of Internet users in the world are CJK users.
   2) UTF-8 is not preferred in CJK area.
   3) In these regions, preferred text-encodings is regional-specific, 
      context-specific, device-specific, etc (ex. Big5, Shift_JIS, EUC-KR).
   4) Current Specification of Unicode Support of Zope 2.6 does not
      satisfy their requirements in practice.
   5) Therefore, they have to use 8bit-string object in order to
      use their language text in Zope.
   6) If we could use UTF-8 without trouble in any situation, things
      are much easier since you can assume anything in Unicode like:
      <input name="var1:utf8:ustring">,
      REQUEST.set('management_page_charset','UTF-8'),
      RESPONSE.setHeader('Content-Type','text/html; charset=UTF-8'),
      "Text is handled as Unicode Object in Zope internal",
      But it's just a fantasy that I18N engineers often stick to.
   7) Overall, the only practical way for them is not to use Unicode
      in Zope and deal with text as raw 8bit string.

or something like that.



I have no intension to discuss about this topic for now.
I guess it is something which will be raised in the near future..



Regards,
---
Heiichiro NAKAMURA <nheiich@quantumfusion.com>