Re: [xml-dev] New approach

["Rick Jelliffe" <ricko@allette.com.au>]

It seems that we have a layered architecture emerging:
   
    combined language or profile (document type), eg. XHTML+SVG+MathML
    standalone language, e.g.namespace such as XHTML, SVG, MathML
    information pool e.g. HTML modules, Ruby, xlink, CALS table
    information unit, e.g. a simple ruby or an extended ruby, a simple xlink or an 
                     extended xlink,  the xml-stylesheet PI
    information item, e.g. xhtml:rb, xlink:href, 

It seems that namespaces are being used for very coarse-grained
element sets, not for fine-grained modules (which is the impression
I get from the namespace spec, but that may be a phantom). 

The combined language or profile seems to be more what a MIME type should
describe, or a RDDL document or an XAR archive.  

Ruby has had quite an ordeal figuring out whether it should have its own
namespace. In the end, they decided against it, because then it
is better for them to be part of HTML rather than separate. They are
clearly a module in the sense that the XML Namespaces has, but given 
that they primarily need to fit conveniently into HTML, they chose not
to have their own namespace.  

I would say Schematron pattern's often correspond to information pools,
and Schematron rules often correspond to information units.

Grammar-based schema languages provide content models
to model information units. XML Schema's <include> supports pooling.
<schema> supports standalone languages.Combined languages and 
profiles are weakly supported using schemaLocation and redefine. 
 

Cheers
Rick Jelliffe
 
P.S.  Anyone who is going to XML Europe 2002, and is interested in software
supporting ad hoc modular construction of documents organized as information 
pools (which may be a namespace), please come and see me in the incubator
area at the Topologi desk.


----- Original Message ----- 
From: "Didier PH Martin" <martind@netfolder.com>
 
I just looked at the new draft: http://www.w3.org/TR/XHTMLplusMathMLplusSVG/ and found the approach interesting. Notably, the concept that an SVG document can be considered the host of XHTML and MathML fragments, idem for MathML and idem for XHTML (As hosts). I played with these profiles and found quite interesting things. Did anybody played with these profiles?