W3C home > Mailing lists > Public > www-tag@w3.org > April 2004

Re: Text for charmod last call comments

From: Bjoern Hoehrmann <derhoermi@gmx.net>
Date: Sat, 03 Apr 2004 00:25:20 +0200
To: Chris Lilley <chris@w3.org>
Cc: www-tag@w3.org
Message-ID: <406dbf02.1731301630@smtp.bjoern.hoehrmann.de>

* Chris Lilley wrote:
>BH> Which is not surprising, since the TAG considers IRIs not part of
>BH> what the TAG considers the web.
>Please don't guess what the TAG thinks or what individual TAG members
>think, especially if you guess wrong.

No guessing here, <http://www.w3.org/TR/2003/WD-webarch-20031209/>:

  Constraint: Identify with URIs

  The identification mechanism for the Web is the URI.

Where "constraint" is defined as 

  An architectural constraint is a restriction in behavior or
  interaction within the system. Constraints may be imposed for
  technical, policy, or other reasons.

My interpretation of this is that using IRIs is considered bad (since
I violate a Web Architecture Constraint) and that I have left the Web
if I do so (since the identification mechanism for the Web is the URI,
not IRIs or yet something else). Most parts of this document support
this interpretation, already the Introduction reads

  The World Wide Web (WWW, or simply Web) is an information space in
  which the items of interest, referred to as resources, are identified
  by global identifiers called Uniform Resource Identifiers (URIs).

This is a definition of "Web" that clearly excludes IRIs, isn't it?
I really don't mind this by the way, URIs are a way too complex and
confusing topic even for experts, adding yet an additional layer of
complexity (IRIs) is a bad idea. IRIs should be called URIs and the
IRI specification should obsolete the URI specification, that'd be
way more useful for the community (which in cases like
http://tidy.sf.net/bug/924809 thinks this has happend already).

The Web Architecture Draft refers to URIs dozens of dozens of times,
will it be updated later this or next year to talk dozens of dozens
of times about IRIs and mention URIs as an historical artifact? I
doubt it. And if, then I would question the utility of talking about
URIs in the first place. All you get from this URI/IRI distinction is
confusion resulting in improper software, content, and specifications.

(I acknowledge that the Webarch document also says "The integration of
internationalized identifiers into the Web architecture is an important
and open issue." but I would prefer if the TAG first determines what
the web is and what its identification mechanism before it publishes
recommendations on this matter).

>This particular TAG member considers IRIs very much part of the Web
>and argued for most of the telcon that they should stay in this
>document, as the minutes show.

Yes, and I appreciate that. Heck, all the section is question is trying
to ensure is that I can refer to my hometown via http://www.dagebüll.de
I could probably live with a more general wording to the effect that
specifications MUST ensure that, but this *is* necessary and fundamental
and thus properly placed in Charmod Fundamentals, a request to strike it
from that document is unreasonable, in my opinion.
Received on Friday, 2 April 2004 17:25:59 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:32:41 UTC