- From: Sandro Hawke <sandro@w3.org>
- Date: Fri, 25 Mar 2011 15:20:14 -0400
- To: Jiří Procházka <ojirio@gmail.com>
- Cc: Steve Harris <steve.harris@garlik.com>, semantic-web@w3.org
On Fri, 2011-03-25 at 18:10 +0100, Jiří Procházka wrote:
>
> Strawman 5: make rdf:Bnode/rdf:ExistentialVariable and use it as usual
> RDF class (+ being a semantic extension)
> <http://example.com/our_bnodes/12343245435> rdf:type rdf:Bnode .
>
> Why complicate things even further?
That doesn't work, because RDF type information would be about the thing
denoted by the node, not the node itself.
For example, with FOAF data, we might find this in one file:
_:sandro foaf:mbox <mailto:sandro@hawke.org> .
which your proposal might turn into:
<http://example.com/bnode/10> foaf:mbox <mailto:sandro@hawke.org> .
<http://example.com/bnode/10> rdf:type rdf:Bnode .
and this in another:
<http://www.w3.org/People/Sandro/data#Sandro_Hawke> foaf:mbox <mailto:sandro@hawke.org> .
Now, a little inference (foaf:mbox is an inverse functional property)
allows us to conclude:
<http://www.w3.org/People/Sandro/data#Sandro_Hawke> rdf:type rdf:Bnode .
... which isn't what we wanted at all.
-- Sandro
Received on Friday, 25 March 2011 19:20:23 UTC