W3C home > Mailing lists > Public > www-tag@w3.org > April 2003

Re: Grinding to a halt on Issue 27.

From: Robin Berjon <robin.berjon@expway.fr>
Date: Tue, 29 Apr 2003 11:42:51 +0200
Message-ID: <3EAE491B.8040303@expway.fr>
To: Julian Reschke <julian.reschke@gmx.de>
Cc: WWW-Tag <www-tag@w3.org>

Julian Reschke wrote:
> By all means, yes, the latter. As far as I understand, IRI support never
> appeared in the requirements for XML 1.1, and thus delaying the specs
> because of the reference to the unfinished IRI spec seems to be the
> absolutely wrong thing to do.

XML had strong requirements for I18N, to me that would *very* strongly mandate 
IRIs. In fact it would be ridiculous to support Unicode and not IRIs. The spec 
may be in flux but such is life and if we have word from the people working on 
it that it's stable enough then imho it's usable.

As far as I can tell issues of equivalence are the same in URIs and IRIs, in 
fact the most frequently quoted example -- %7E -- is a character in the US-ASCII 
range. The way namespaces are compared today is using string equality, 
irrespective of escapes. Hard as I try I fail to see how this will change. I 
totally fail to see the benefit in changing what we have now to something that 
will simply cost more. In fact, the idea of someone wanting to change it makes 
me shudder with the thought of the thousands of lines of software I have out 
there that just use string comparison and would suddenly be broken (and that 
without counting the fact that Java is broken[0] on this issue, "language of the 
internet" har, har).

As Larry Wall put it, it's easier to fix a broken spec than to fix 100.000 
broken implementations. I guess the reverse is right too.


Robin Berjon <robin.berjon@expway.fr>
Research Engineer, Expway        http://expway.fr/
7FC0 6F5F D864 EFB8 08CE  8E74 58E6 D5DB 4889 2488
Received on Tuesday, 29 April 2003 05:42:51 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:55:58 UTC