Re: Evaluations of Test Cases 46-56

Manu Sporny wrote:
> TC48: SEMANTICS OK, SPARQL OK
>       I've stared at this one for quite some time and thoroughly
>       confused myself. Somebody that is more familiar with the
>       processing rules will have to explain why this is right or
>       wrong. It passes the latest pyRDFa and SPARQL implementations.

Here's how I think about it:

If you had the following markup:

<div about="http://www.example.org"
     rel="foaf:knows">
  <p property="foaf:name">John Doe</p>
</div>

Then by the chaining rule of @rel, you get:

<http://www.example.org> foaf:knows _:x
_:x foaf:name "John Doe" .

where the bnode _:x "corresponds" to the DIV and is the chaining node.

@instanceof, when added, applies to this corresponding bnode, the
chaining node, so, when you add

instanceof="foaf:Person", you get the same triples as above plus:

_:x rdf:type foaf:Person .


-Ben