perl-unicode

Re: UTF-8 in web pages

2001-08-06 18:40:33
At 10:02 01/08/06 -0700, Tim Walters wrote:
Markus Kuhn wrote:

> Andrew McNaughton wrote on 2001-08-05 10:38 UTC:
> > > You can send Unicode directly to the Web browser. Just make sure you
> > > announce in the HTTP header that the body is encoded in UTF-8.
> >
> > Sounds nice, but in practice this excludes a lot of users still.
>
> Not really. All M$-Win32 browsers support UTF-8 flawlessly.

I'm glad UTF-8 display is working well for you. Unfortunately, we have run
into a few UTF-8 bugs in Netscape Navigator which were a problem for our
application. I hope it's not too much off-topic to list the two big ones, in
case other readers run into them also:

- There are some non-negligible setup issues (nicely detailed in one of Alan
Wood's Unicode Resource Pages,
http://www.hclrss.demon.co.uk/unicode/netscape.html).

As that page says, what you write below applies to NS 4.x. For NS 6, the
situation is much better. In particular:

"Navigator 6 can use characters from several Unicode ranges to display a
single Web page, and appears to be able to interrogate the operating system
to identify fonts that include characters from any required Unicode range."

So upgrading to NS 6 is another option.

Regards,   Martin.



Unfortunately, you
have to set up a *single* "unicode" font for dealing with all characters,
whereas most machines just have fonts for specific scripts lying around
(i.e., a font for Japanese, a font for Korean, etc.) If you select a
Japanese font for Unicode display it will work fine for Japanese and other
characters in the font, but won't be able to deal with Korean or other
characters which aren't included. Netscape misbehaved and crashed a lot when
I tried to get it to use the huge Microsoft Unicode font with all of the
characters.

- Text input boxes in forms only accept characters in the "ANSI" system code
page. You can't enter or edit Japanese data, even if you can display it.

IE5 does not have these problems and, of course, it's much more stable. If
you (and the browsers viewing your pages) can stick to IE5, UTF-8 shouldn't
be a problem.

Cheers,
Tim Walters

<Prev in Thread] Current Thread [Next in Thread>