RE: Linking RDF

Brian McBride [] wrote:

>Hi Lee,

Hi Brian!

>> Discussions on these issues seem to have died down, yet the issues have
>> been resolved and the new RDFCore working group are not even going to
>> address them.  
>There is an issue on the RDF Issues list:
>Does this, together with the link to Perry's message cover the issues
>you'd like to see resolved, or are there other aspects you'd like to
>draw the WG's attention to?

The link to Perry's message covers one aspect, "weird, unwieldy namespaces
with different semantics to
other XML namespaces" - i.e. the deviation of namespace treatment in RDF
from its use in other XML technologies and the implied intent of the XML
Namespace spec (as stated in one of the appendices [1], IOW it was made
clearer *and* normative).

The two other issues were:

* "open grammar, which is harder to validate simply (and nigh on impossible
to do properly with DTDs)" - Syntax validation within the context of RDF
embedded in other XML grammars would be easier if the RDF syntax were only
of the 'Fixed-Schema' variety, see [2].  Currently, the propertyElt
construct, and abbreviated forms of RDF are of the 'Schema-follows-data'

* "resolution of RDF schemas clashes with resolution of XML schemas" - both
XML Schema and RDF Schema utilise namespace URIs to locate markup that
describes XML syntax rules and RDF model rules, respectively.  I am now
thinking this is not so much of an issue, as you could always use
content-negotiation to retrieve the type of schema you are looking for, and
there is another mechanism within RDF to state where to find a schema
description for a given resource (i.e. rdfs:isDefinedBy).



PS, I for one find it very reassuring that the issues list is now being
actively maintained, keep up the good work!

Received on Monday, 23 April 2001 08:04:44 UTC