OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

 


 

   RE: [xml-dev] Naive Question about NS 1.1

[ Lists Home | Date Index | Thread Index ]
  • To: "Leigh Dodds" <ldodds@ingenta.com>
  • Subject: RE: [xml-dev] Naive Question about NS 1.1
  • From: "Manos Batsis" <m.batsis@bsnet.gr>
  • Date: Wed, 10 Apr 2002 12:04:08 +0300
  • Cc: <xml-dev@lists.xml.org>
  • Thread-index: AcHf1VLbh0D1FWnjQLqiksFoxkdY+gAl4Zpw
  • Thread-topic: [xml-dev] Naive Question about NS 1.1



> From: Leigh Dodds [mailto:ldodds@ingenta.com] 
think I'm misunderstanding something, can anyone point 
> where my train of thought is being derailed:
> 
> - The Infoset defines that the namespace information items 
> for an element are all those that are in scope. i.e. defined 
> on this element or it's ancestors.

I was thinking about that specific part too. IMHO, for the Infoset to
serve it's cause without bringing burden to other specifications or
applications such as fragment exchange, the namespace Information Items
must not be the ones "in scope" but the ones that are either present by
a prefix or as a default namespace. However:

> - This leads to revising the namespace spec to allow undeclaring of 
> namespaces, alleviating the issues in the other specifications.

Again IMHO, the ability to undeclare any namespace should have been
there right from the start and it's omission was a big poopoo (sorry
folks), as well as the poor support or bad design of support for
namespaces by other standards. 


> Why let changes ripple out like this? Why can't the infoset 
> definitions be 
> changed? 

Absolutely, that's the part that needs fixing concerning the above.

> Why does an element need namespace information items?

Because they belong in the core :-) Seriously now, namespaces where a
good spec. Simple and effective, although I would want more from it. For
example, I always believed that schemata should be bound to documents
using namespaces or ideally, elements should be bound to complexTypes
using namespaces.

Still, the problem is the set of problems. How about leaving the
existing major versions as is and gather the problems as a whole to
build XML 2.0?

Establish the missing layer (currently almost addressed by RDDL).
Bind schemas to documents using that layer.
Fix namespaces and have them pointing to that layer.
Label all the above as XML 2.0 Core.
Leave XML 1.0 and Namespaces 1.0 as is.


DTDs can keep working perfectly with XML 1.0 and NS 1.0, but I strongly
believe that the XML community should admit the current problems,
discuss them and build the next, clean generation of XML specs. This
time, we just have to build them solid based on a fixed Infoset.

Just my .25

Manos




 

News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS