RE: [xml-dev] DTDs, W3C Schemas, RELAX NG, Schematron?

["Don Box" <dbox@microsoft.com>]

> -----Original Message-----
> From: James Clark [mailto:jjc@jclark.com]
> Sent: Sunday, June 23, 2002 10:15 PM
> To: Don Box; John Cowan
> Cc: xml-dev@lists.xml.org
> Subject: Re: [xml-dev] DTDs, W3C Schemas, RELAX NG, Schematron?
> 
> 
> > > 5) RNG allows a (simple) datatype to be restricted not only by
facet,
> > > but by explicit extension and exception.  One can write
'xsd:integer -
> > > "0"' to mean a nonzero integer (although unfortunately "00" will
still
> > > validate), or 'xsd:integer | "Inf" | "+Inf" | "-Inf"' when a value
may
> > > be integral or infinite.
> >
> > Wow! I was knocked out by this one. What good are datatypes if one
can
> > [not] work in terms of the value space of the type rather than the
> lexical
> > form?
> 
> In RNG, you can work in terms of the value space.  If you want a
non-zero
> integer, you can say
> 
> xsd:integer - xsd:integer "0"
> 
> or
> 
> <data type="integer">
>   <except>
>     <value type="integer">0</value>
>   </except>
> </data>

Thanks for the clarification. This seems reasonable. 

>  > I've found that the most important difference between RNG and XML
> > Schemas is wrt the underlying type model.
> >
> > Specifically, XML Schema uses named types to implicitly convey
intention
> > as well as structure (a la Java, C++, C#). Anonymous types are 2nd
class
> > citizens in XML Schema (they're type-equivalent with nothing).
> >
> > Relax NG has a much looser model (a la Perl) in which two things are
> > compatible if they share structure independent of a common, named
type
> > definition.
> >
> > Each approach appeals to its own community of users in very deep
ways.
> 
> I agree that the emphasis on named typing in W3C XML Schema is a
profound
> difference from RNG.  One interesting question is whether it would be
> possible to build a schema language that supports named typing yet
still
> has
> (at least most of) the simplicity and power of RNG.  

I had this thought myself recently, but frankly, I think lack of named
typing is RNG's primary attraction for most people. 

> However, I still have
> my doubts that named typing is appropriate for XML.  I would speculate
> that
> named typing is part of what makes use of DCOM and CORBA lead to the
kind
> of
> relatively tight coupling that is exactly what I thought we were all
> trying
> to avoid by moving to XML.

I admit that it's a very weird dynamic. 

Correct me if I'm wrong, but even with RNG, equivalence still relies on
common element names unless wildcards are used everywhere, no?  

DB