- From: Bijan Parsia <bparsia@cs.man.ac.uk>
- Date: Wed, 12 Dec 2007 14:45:52 +0000
- To: Jeremy Carroll <jjc@hpl.hp.com>
- Cc: Semantic Web <semantic-web@w3.org>
Is there any real advantage to this over using xml literals? Or even string literals? (We did try xml literals for OWL-S but tool support was weak...) I don't see how this helps with, e.g., axiom annotations esp. as it stuffs all the triples into the parent file as well. (Which I take it is for backwards compat with current parsers? Warn that there's this funky attribute and then drop the named graphs?) Decent thing about the literal approach is that it can be layered ontop of existing parsers. Also, things in named graphs can be spread out throughout a file instead of clumped. (And wouldn't it be nice to add property elements with a generic rdf:predicate rdf:predName="" bit of syntax so that we can capture all RDF graphs and have relatively XML friendly serializations and....) (Oops, was that the sound of a can of triples^H^H^H^H^H^Hworms opening? ;)) Cheers, Bijan.
Received on Wednesday, 12 December 2007 14:44:23 UTC