More on empty property elements, ID/resource and the associated productions.

Dave B's just pointed this out to me (I lost the coin toss so I get to
write this :-/ )

In brief:
- does anyone out there actually _use_ the last clause of production
  [6.12] in the grammar?

namely:

propertyElt ::= '<' propName idRefAttr? bagIdAttr? propAttr* '/>'

The interpretation of this is given by paragraphs 229-243
(see http://ioctl.org/rdf/ms/rdfms for numbers)

...but an example is simpler. I've added some additional test cases.

test1 shows this paragraph in action. Test2 demonstrates how an
anonymous resource is created even without an rdf:resource attribute.
Test3 shows something similar; this isn't currently allowed by the
grammar but is indistinguishable from test2.

Finally, error1 shows that the value of the empty property in this case
(where propAttrs exist) is constrainted to be a resource; it's an error
to try doing otherwise.

I've left out rdf:ID examples from this because (as Jos pointed out)
this needs to be nailed down. If rdf:ID were used to create a reified
statement, it would have to be made clear _which_ of the statements
created* by the production had their reification identified by its
attribute.

jan

* if created smacks of process too much, use "implied" or something
equivalent.

-- 
jan grant, ILRT, University of Bristol. http://www.ilrt.bris.ac.uk/
Tel +44(0)117 9287163 Fax +44 (0)117 9287112 RFC822 jan.grant@bris.ac.uk
HP-unix: Open Sauce product, available in 57 distributions.