- From: Robert Sanderson <azaroth42@gmail.com>
- Date: Mon, 23 Jul 2012 09:38:41 -0600
- To: Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>
- Cc: Leyla Jael García Castro <leylajael@gmail.com>, Reto Bachmann-Gmür <reto@apache.org>, public-openannotation <public-openannotation@w3.org>
On Mon, Jul 23, 2012 at 2:30 AM, Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk> wrote: > The specification describes oa:hasBody as a "Relationship" rather than > a "Property". > So my short answer is no, oa:hasBody must go to a resource, not a literal. Yes, that's correct. oa:hasBody always has a resource as subject, not a literal. This is primarily for two reasons: * Consistency for Modeling. We do not want to have to inspect the object in the triple to figure out what is going on. It should always just be a resource. This makes the modeling much cleaner at the (minimal) expense of two extra triples. * Consistency for Data. As the body resource may not be text (it could be video, audio, image, data, anything!) then the case where it is text should not be a special case, for the sake of saving one triple. > http://www.openannotation.org/spec/core/#Inline species how the > literal (at least string or bytes) can be represented using Content In > RDF. Yep, and also a server should behave if it wants to turn the inline content into a full resource with a dereferencable URI. > I think I have raised earlier that the RDFS definition should > clarifier this with light use of OWL. Sure. I'm happy to replace the current RDFS with whatever is most appropriate, so long as someone (Stian?) will commit to maintaining it :) Rob
Received on Monday, 23 July 2012 15:39:19 UTC