Re: My action item on RDDL/RDF

Tim Bray wrote:
>...
> 
> What else would you need?  -Tim

At first glance, I like your proposal. I think that there is a 
requirement for this sort of thing on one end of the 
explicit-and-inline/implicit-and-non-intrusive spectrum.

At the other end, I think we also need some declarative way to say: "If 
you need to view this XML as RDF triples, here is how." This is 
necessary even if only as a transitional manner. But I would be happy to 
have that as a separate specification.

Also, I think that the prefix-abbreviation idea of namespaces really 
does make sense for data also. Maybe we would just use QNames. After all 
that battle was probably lost years ago. XSLT, WSDL and XML Schema all 
use Qnames-as-data.

But if I can't convince you, then there could be another short-form feature:

<root xmlabbr:rddl="http://www.rddl.org/"
       xmlabbr:schemas="http://www.where.I.keep.my.schemas.com/"
       xmlabbr:iana=
"http://www.isi.edu/in-notes/iana/assignments/media-types/application/"
       xmlabbr:myns="http://my.namespace.com/">

...
<R r="${myns}#">
<PV p="${rddl}#purpose" v="${rddl}/purposes/#validation"/>
<PV p="${rddl}#nature" v="${iana}/xml-dtd"/>
...
</R>
</root>

I don't care about the syntax but I think that short-form is necessary.

The propertyBase proposal is only helpful if most properties come from 
the same namespace but the more popular RDF gets, the more mixed the 
documents will get. Even your simple RDDL examples use a variety of 
different namespaces.

And anyway, I think that a short-form proposal subsumes the features of 
your resourceBase, propertyBase and valueBase attributes. Maybe it is 
even a solution to the character entity short-name problem.

  Paul Prescod

Received on Monday, 11 November 2002 18:25:20 UTC