- From: Pat Hayes <phayes@ihmc.us>
- Date: Fri, 29 Nov 2013 12:04:58 -0600
- To: Dan Brickley <danbri@google.com>
- Cc: Guus Schreiber <guus.schreiber@vu.nl>, RDF WG <public-rdf-wg@w3.org>
On Nov 28, 2013, at 4:49 AM, Dan Brickley <danbri@google.com> wrote: > On 28 November 2013 10:12, Guus Schreiber <guus.schreiber@vu.nl> wrote: >> >> >> On 28-11-13 06:00, Pat Hayes wrote: > >>>> RDF Schema provides a simple data-modelling vocabulary for RDF data. >>>> Other publications, including SKOS [ ] and the W3C Recommendation OWL2 [ ], >>>> define more elaborate data models which extend RDFS in various ways. >> >> >> I prefer to leave "simple" out. RDFS is not simple, mainly because it has >> the notion of sub-property, which is a powerful mechanism not present in most >> data-modelling languages (UML has the notion of "Association Class", but that >> is less flexible). > > It might still be 'simple'; this just shows some awkwardness in > calling it a 'simple data modeling' language, perhaps? e.g. It might > be a simple language for data modeling, even if not 'simple' by > data-modeling language standards? > >> Also, "simple" has a subjective connotation. > > Clearly :) > > As with RDFa there is quite a different experience for implementors > versus publishers/authors. If I'm just writing claims using RDF/S, > it's straightforward enough to understand that something like > dc:creator refines the more general super-property dc:contributor, and > that if I say that someone is the dc:creator of something it wouldn't > make much sense to simultaneously deny that they were also a > dc:contributor. Authors will need help understanding how domain/range > interact here, but the basic idea is straightforward. > > For someone writing an inference engine or trying to optimize database > indices for common query patterns, they might plausibly object to the > use of 'simple' :) Well actually it was thinking about the inferences needed, and mentally contrasting that with what is needed for OWL2, that suggested the word "simple" to me. You can implement a complete RDFS reasoner using the entailment patterns in the spec, applied blindly to exhaustion using a simple one-way unifier, and it will run quite effectively on most published RDFS, but you need be a description-logic specialist to even start implementing a semantic tableau OWL reasoner. But I won't defend the use of "simple" if Guus wants to remove it. Pat > Similarly, the publisher experience of RDFa and > Microdata is pretty similar, whereas the parser writers have > significantly more pain with RDFa, to support its various additional > idioms. > > Dan > ------------------------------------------------------------ IHMC (850)434 8903 home 40 South Alcaniz St. (850)202 4416 office Pensacola (850)202 4440 fax FL 32502 (850)291 0667 mobile (preferred) phayes@ihmc.us http://www.ihmc.us/users/phayes
Received on Friday, 29 November 2013 18:05:34 UTC