W3C home > Mailing lists > Public > public-xml-id@w3.org > February 2005

RE: Change namespace of xml:id

From: John Boyer <JBoyer@PureEdge.com>
Date: Tue, 8 Feb 2005 11:36:05 -0800
Message-ID: <7874BFCCD289A645B5CE3935769F0B52750789@tigger.pureedge.com>
To: <veillard@redhat.com>
Cc: "Chris Lilley" <chris@w3.org>, "Bjoern Hoehrmann" <derhoermi@gmx.net>, <public-xml-id@w3.org>

Your *opinion* is that it is broken, and your conclusion that a 'mistake'
was made is based on that opinion.

But your opinion is based entirely on not interpreting the word 'identify' 
in the same way that I do.  How do you justify identifying two non-equal
sets of names with the same identity?

You can be annoyed with a technical opinion that differs from yours,
and the W3C is welcome to 'fix' C14N if it likes so that only the
two attributes described in XML 1.0 are handled in the way described.

But the post hoc addition of names to the namespace identified by
a URI (once a recommendation is let) remains an abuse of the notion 
of namespace due to the normative definition of namespace appearing 
Namespaces in XML.  You can do it now, but don't expect this to be the 
only problem that crops up over time.  The W3C has policies for assignment 
of namespace URIs to versions of languages that come out, and it has
these policies for a reason.  Welcome to the reason.

John Boyer



-----Original Message-----
From: Daniel Veillard [mailto:veillard@redhat.com]
Sent: Tuesday, February 08, 2005 10:44 AM
To: John Boyer
Cc: Chris Lilley; Bjoern Hoehrmann; public-xml-id@w3.org
Subject: Re: Change namespace of xml:id


On Tue, Feb 08, 2005 at 08:44:17AM -0800, John Boyer wrote:
> Hi Daniel,
> 
> >  Nothing in the Namespace in XML spec implies that this binding
> >is static as you're claiming.
> 
> Yes it does.  The first definition in the recommendation defines
> a namespace to be a collection of names identified by a URI.
> If you change the collection of names, you change the namespace.
> Since the meaning of 'identify' is fairly clear, a different URI must
> be used to 'identify' the new collection.

  A collection is not a list. It doesn't have to be closed. I reserve
the right to use the namespace {http://veillard.com/} to any local name
in the future if I deem it useful, even if right now the collection
doesn't include any.
  The C14N group made a mistake. It's incredible that this is leading
to such a mess. I'm not sure I value interoperability more than getting
broken stuff fixed.

Daniel

-- 
Daniel Veillard      | Red Hat Desktop team http://redhat.com/
veillard@redhat.com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/
Received on Tuesday, 8 February 2005 19:36:48 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:53:49 UTC