Re: http charset labelling

Masataka Ohta (mohta@necom830.cc.titech.ac.jp)
Mon, 5 Feb 96 17:38:12 JST


From: Masataka Ohta <mohta@necom830.cc.titech.ac.jp>
Message-Id: <199602050838.RAA11677@necom830.cc.titech.ac.jp>
Subject: Re: http charset labelling
To: gtn@ebt.com (Gavin Nicol)
Date: Mon, 5 Feb 96 17:38:12 JST
Cc: masinter@parc.xerox.com, keld@dkuug.dk, uri@bunyip.com
In-Reply-To: <199602021428.JAA02019@ebt-inc.ebt.com>; from "Gavin Nicol" at Feb 2, 96 9:28 am

> >So, the available choices of representation on our namecards are:
> >
> >	1) pure ASCII
> >	2) a few % notations embedded in ASCII
> >	3) a lot of % notations
> >	4) MIME QP with charset labelling
> >
> >Obviously, 1) is the best both for recognizing and for typing in.
> 
> The only problem is that people *expect* to be able to use whatever
> encoding they like in URL's.

Then, all we can do for them is to write an informational, or BCP,
RFC on why localized URL can't be used for internationalization.

> If you look at some of the magazines
> in Japan with articles on creating WWW sites. Many of them have URL's
> that contain Kanji.

Did they? Can you name more specific references?

> If you look even closer, you can see than browsers
> display this as junk.

Such behaviour is unavoidable and is the reason why localization is
no good.

						Masataka Ohta