RDF Issues

I would like to see an improvement to the RDF syntax, and would also be
happy to offer time and effort to see it through.

I whole-heartedly agree with all aspects of Brian McBrides response to your
call for contributions.  There should be a move to separate model & syntax,
identify and resolve model issues, then design a "simper, more regular"
syntax in that order.

To this end I suggest it would be useful to put together a document
describing the RDF model aspects of the M&S spec without any reference to
syntax.  Not only would this be useful in its own right, it can become the
focal point for resolving model issues and subsequently driving out the
design of the new syntax (or else for evaulating other web data graph syntax
for the job).

Work on it could begin right now and be done in parallel to other current
tasks such as collating issues with the current syntax.  Does anyone agree
this would be a good starting point?  If asked to, I would be happy to kick
it off.

I would also suggest that further issues be posted to the list in the format
of the issues document.  If I were in Dan's shoes, content would be
preferable to points alone and from the RDF IG's point of view, distributing
work in this manner would help speed things along for all of us.

I will post other issues as I come accross them. In the meantime, I have
fleshed out one of the other www-rdf-comment issues that help was requested
for:

1) rdf:resource vs resource
===========================

RDFMS-???: Misapplication of namespace semantics to RDF attributes

Raised Wed, 26 Apr 2000 by mailto:connolly@w3.org

Summary: unqualified RDF attributes on element types in the RDF namespace
are _not_ equivalent to attributes with the RDF prefix.

see also: http://www.w3.org/TR/1999/REC-xml-names-19990114#uniqAttrs,
http://www.xml.com/pub/2000/03/08/namespaces/myth1.html#myth4, and
http://lists.w3.org/Archives/Public/www-rdf-comments/2000AprJun/0019.html


Analysis: According to (the non-normative) Appendix A.2 in the 'Namespaces
in XML' spec, attributes with a prefix are in the 'Global Attribute
Partition' wheras attributes without a prefix are in the 'Per-Element-Type
Partition'.  Hence rdf:resource and resource may share a localpart.  However
they are entirely distinct entities (at least syntactically).

Examples in the RDF spec interchange the qualified and unqualified
attributes at different points.  Specifically 'rdf:about', 'rdf:type',
'rdf:resource', and 'rdf:value'.  The tendancy in the spec is to use
unqualified attributes for basic RDF syntax examples and qualified
attributes for second and third RDF abbreviated form examples - in these
cases the element type is (usually) not in the RDF namespace, so the
attribute is given the RDF prefix.

A suggested solution is to use global (qualified) attributes throughout.  In
order to make the syntax slightly more forgiving, parsers should treat any
per-element-type attributes on RDF elements the same as their global
counterparts.

Currently: for discussion

Received on Friday, 8 September 2000 06:32:22 UTC