W3C home > Mailing lists > Public > public-esw-thes@w3.org > February 2007

Re: [SKOS] OWL DL compatibility

From: Dave Reynolds <der@hplb.hpl.hp.com>
Date: Wed, 21 Feb 2007 12:27:10 +0000
Message-ID: <45DC3A9E.9020500@hplb.hpl.hp.com>
To: Bernard Vatant <bernard.vatant@mondeca.com>
CC: SKOS <public-esw-thes@w3.org>

Bernard Vatant wrote:
> Hi all
>> I'm still digesting the issue. But one immediate point: I am 
>> uncomfortable with excessive use of OWL's notion of "annotation 
>> properties". They're very much 2nd-class citizens in the OWL world. 
>> I'd rather define extra properties for class etc annotation, and 
>> restrict prefLabel, altLabel to SKOS individuals.
> +1
> The first objective IMO is to be able to integrate SKOS instances in an 
> OWL-DL framework. Using the SKOS vocabulary to label OWL elements is of 
> course appealing, but potentially confusing, and I'm not sure it's such 
> a good idea. OWL-DL has deliberately sacrified the linguistic aspects of 
> naming on the autel of logic, by putting NL labels and languages outside 
> the semantics. I think we (SKOS) should not mess up with that, or try to 
> patch that bug of OWL-DL which is actually a deliberate built-in feature 
> But both OWL and SKOS should be clear on this. Want logic, rules, 
> inference, go the OWL-DL way but forget about dealing with the 
> linguistic aspects (or build your own ad hoc classes and properties).

Sorry if it is not appropriate to interject into this discussion but ...

There is no actual problem with using "logic, rules and inference" and 
keeping the current SKOS approach of refining rdfs:label. It is 
perfectly valid, well-defined, implementable RDFS and OWL/full.

The choice of whether to stick within OWL/DL rather than OWL/full is not 
the same as the choice of whether to use formal inference or not.

Received on Wednesday, 21 February 2007 12:27:33 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 13:32:08 UTC