RE: Significant W3C Confusion over Namespace Meaning and Policy

Tim,

Is there in effect a substitution rule being applied to xml:id in xml:id
unaware XML 1.0 processors?  Seems like the substitution rule is roughly
to treat xml:* attributes as regular attributes, modulo the extra
complications.

Cheers,
Dave


> -----Original Message-----
> From: www-tag-request@w3.org [mailto:www-tag-request@w3.org] On Behalf
Of
> Tim Berners-Lee
> Sent: Saturday, February 12, 2005 7:52 AM
> To: John Boyer
> Cc: www-tag@w3.org; Bjoern Hoehrmann
> Subject: Re: Significant W3C Confusion over Namespace Meaning and
Policy
> 
> 
> 
> On Feb 9, 2005, at 16:34, John Boyer wrote:
> 
> >
> > Dear TAG,
> >
> > Some of you may be aware that an issue with the
> > xml:id specification began erupting the day
> > before it became a CR.
> >
> > The issue has flowered nicely into a more general
> > discussion of what namespaces mean and what is
> > the W3C policy regarding their assignment in
> > recommendation track documents.
> >
> > I've been asked to provide this information to you,
> > and as PureEdge AC rep I'd like to please request that
> > the TAG make a formal statement to all working groups
> > regarding these issues as soon as possible.
> >
> > The kernel of the issue is my interpretation of
> > the definition of namespace as it appears in the
> > Namespaces recommendation.  The definition is that
> > a namespace is a collection of names *identified*
> > by a URI. So, for example, the namespace
> > ({lang, space}, http://www.w3.org/XML/1998/namespace)
> >
> > is not equal to
> >
> > ({lang, space, base, id}, http://www.w3.org/XML/1998/namespace)
> >
> > The W3C director directed the W3C to this interpretation in
> > http://www.w3.org/1999/10/nsuri, which states that a recommendation
> > cannot add more than clarifications and bug fixes without changing
> > the namespace URI.
> 
> Good point.
> Well,it says that changes should not be done which break existing
> data or code which reads it.
> 
> Yes, the addition of xml:id tag changes XML ins such a way that
> new documents with xml:id attributes are not XML 1.0 documents,
> if you interpret the XML1.0 spec to rule out the use of attributes
> in the XML namespace.
> 
> I think the argument here is that actual implementations
> (except canon'n) all treat xml:id as a normal attribute,
> and so they don't break.  But of course it is more complicated than
> that.
> 
> In general, the emphasis up to CR is on the WG simply
> setting accurate expectations for the sorts of changes to be made.
> Maybe that should be continued through to Rec.
> 
> It sound as though the TAG should consider this.
> 
> Tim BL
> 
> 
> 
> > Thanks,
> > John Boyer, Ph.D.
> > Senior Product Architect and Research Scientist
> > PureEdge Solutions Inc.
> >
> >
> >
> >
> >
> >
> 
> 

Received on Monday, 14 February 2005 17:29:54 UTC