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 08:40:25 -0800
Message-ID: <7874BFCCD289A645B5CE3935769F0B52750786@tigger.pureedge.com>
To: "Chris Lilley" <chris@w3.org>
Cc: "Bjoern Hoehrmann" <derhoermi@gmx.net>, <public-xml-id@w3.org>

Hi Chris,

>One can, but one need not. There are examples of both types of change policy.

>The change policy for the XML namespace has already ben demonstrated,
>for example when xml:base was added to it.

Was this a conscious decision about the change policy, or just
everyone not understanding something about namespaces?
I find it quite hard to believe that anyone really gave
this point much thought.

>The meaning of 'reserved' is fairly clear. 

Please look carefully at what is reserved.  The names of prefixes are 
reserved.  This is orthogonal to reserving the meaning of the namespace URI.

>Another specification (such
>as XML canonicalization) should not be making assumptions about future
>registrations into the XMl namespace, since that namespace is reserved.

The namespace is not reserved in the sense you mean it.  The namespace is
defined by XML 1.0, for which C14N is the canonicalization algorithm.

Post hoc, the namespace definition is changing.  Namespaces in XML 
defines a namespace to be URI+(collection of names).

Cheers,
John

JB> Or, more appropriately, use Bert's suggestion of the
JB> prefix xmlid, which would then be bound to a *different*
JB> URI than the one associated with XML 1.0.

JB> Thanks for the link!
JB> John Boyer



JB> -----Original Message-----
JB> From: Bjoern Hoehrmann [mailto:derhoermi@gmx.net]
JB> Sent: Monday, February 07, 2005 3:47 PM
JB> To: John Boyer
JB> Cc: public-xml-id@w3.org
JB> Subject: Re: Change namespace of xml:id


JB> * John Boyer wrote:
>>Just in case it doesn't get noticed in the XML Signatures conversation,
>>I'm recommending that you change xml:id to another namespace
>>because adding it to the xml namespace does an end-run around
>>the purpose of a namespace that happens, in this case, to break
>>something.  Notably, canonicalization.

JB> In http://www.w3.org/mid/16178.32403.534822.783054@lanalana.inria.fr
JB> Bert Bos proposed "xmlid" which is indeed reasonable as all attribute
JB> names starting with "xml" are reserved. The Working Group unfortunately
JB> refused to formally address comments on the requirements document.
JB> There were other important requests aswell, e.g., Gustaf Liljegren in
JB> http://www.w3.org/mid/3.0.6.32.20030811231834.00b23128@pop.bredband.net
JB> arguing for numeric id values which are popular xml:id values indeed,
JB> http://www.nntp.perl.org/rss/perl.i18n.rdf uses them for example.




-- 
 Chris Lilley                    mailto:chris@w3.org
 Chair, W3C SVG Working Group
 Member, W3C Technical Architecture Group
Received on Tuesday, 8 February 2005 16:41:07 UTC

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