Re: language_names

[Yoshizumi Endo <y-endo@ceres.dti.ne.jp>]

"James A. Treacy" <treacy@debian.org> writes:

> > > > Also, "character entity references" can handle multibyte-characters
> > > > with ISO10646 code set.
> > > > 
> > > >   Japanese (Chinese characters) -> &#x65e5;&#x672c;&#x8a9e;
> > > >   Chinese  (Chinese characters) -> &#x4e2d;&#x6587;
> > -snip-
> > > One thing that confuses me (which may simply be a symptom of a
> > > greater misunderstanding): will this work when the translations are
> > > encoded using their native charset, e.g. iso-8859-1? My impression was
> > > that the pages would need to be encoded entirely in something like
> > > ISO10646 for this to work.
> > 
> > Yes, this (which is called "numeric character reference") works
> > independent of the charset of pages too.  It is not necessary to
> > change the charset currently used in any pages.
> > 
> > I think that to use "character entity references" and "numeric
> > character reference" is a simple method to actualize multi-lingual web
> > page at the moment.
> > 
> 
> I wish you were here when we were originally discussing this. :)

I send this mail only to debian-www.

> If we do this, we should add some information on sources for fonts
> and how to get a browser to recognize them. Any pointers? Even
> better would be if you could write up something for us.

Unfortunately, I don't know any pointers.

As far as I know, OE5 can handle properly "numeric character
reference" and get necessary fonts automatically via the Internet.

Mozzila M12 can handle it too.  But users should get manually some
necessary fonts.  Debian users will need appropriate xfonts-*
(ex. xfonts-cjk) packages.

-- 
Yoshizumi Endo <y-endo@ceres.dti.ne.jp>