Re: Using UTF-8 for non-ASCII Characters in URLs

Larry Masinter (masinter@parc.xerox.com)
Fri, 2 May 1997 01:58:06 PDT


Message-ID: <3369AC9E.281F@parc.xerox.com>
Date: Fri, 2 May 1997 01:58:06 PDT
From: Larry Masinter <masinter@parc.xerox.com>
To: "Martin J. Duerst" <mduerst@ifi.unizh.ch>
CC: URI mailing list <uri@bunyip.com>
Subject: Re: Using UTF-8 for non-ASCII Characters in URLs

This is a great start at dealing with the issues that would
otherwise cause great confusion.

Other issues:
The bidi issues for RLT languages in conjunction with
normal punctuation used in and around identifiers. (Will
the identifiers present themselves 'correctly' without
these characters in all cases?)

Using UCS in identifiers that are normally "case insensitive"
in ASCII, and the issues, e.g., similar upper-case forms,
the role of accents and equivalence.

I think "white space" or spacing characters in general
need to be addressed.

You need to decide whether you're doing canonicalization/normalization
or just equivalence. Equivalence is probably easier to define,
and less politically sensitive, even though not as useful.