- From: <Simon.Cox@csiro.au>
- Date: Thu, 2 Feb 2017 22:38:13 +0000
- To: <L.Svensson@dnb.de>, <maxime.lefrancois@emse.fr>, <ghislain.atemezing@mondeca.com>, <public-sdw-wg@w3.org>
Folks - I won't make this a veto issue, and would be OK adding these triples if that really is the policy. But I have a residual uneasiness. The convention to use rdfs:isDefinedBy to point from ontology resources to the ontology (URI) is by now quite well established. But it does seem to somehow hijack the rdfs:isDefinedBy predicate for a rather narrow use, in the context of formal ontologies. There is a risk of implying that all other uses, which were allowed by the original RDFS spec, are somehow invalid. I can see that there is a gap in the OWL design (ontology membership/ownership predicate) and I see where this convention came from. But I would suggest that some clarification in W3C or RDFS is probably warranted. Simon -----Original Message----- From: Svensson, Lars [mailto:L.Svensson@dnb.de] Sent: Thursday, 2 February, 2017 23:24 To: Maxime Lefrançois <maxime.lefrancois@emse.fr>; Ghislain Atemezing <ghislain.atemezing@mondeca.com>; public-sdw-wg@w3.org Subject: RE: OWL-Time - questions on Turtle file On Thursday, February 02, 2017 1:04 PM, Maxime Lefrançois [mailto:maxime.lefrancois@emse.fr] wrote: > Le 01/02/2017 à 18:03, Francois Daoust a écrit : > 2. There are no "rdfs:isDefinedBy" statements. More specifically, I > was expecting to see "rdfs:isDefinedBy <http://www.w3.org/2006/time>" > for each class and property in the ontology. Aren't they needed (or > recommended) to link back classes and properties back to the ontology? > +0. There is no recommendation of using rdfs:isDefinedBy for each > +class and > property. IMHO, knowing that all the ontologies used in the vocab are > dereferenceable par their URIs, I would not see any need to add that extra triples. > > +1 in favour of rdfs:isDefinedBy axioms. > > My arguments are : > > - this is in the set of Metadata Recommendations For Linked Open Data > Vocabularies, promoted by the Linked Open Vocabulary website [1] . See > section "Vocabulary Elements (Classes and Properties)" > - This is consistent with the practice applied in most of the > vocabularies published on the W3C website. > - This practice will also be applied in the ontologies published by the SSN sub-group. > There should be some consitency at least within the SDW group. +1 for rdfs:isDefinedBy from me, too, for the very same reasons (that I couldn't have phrased so well). Best, Lars
Received on Thursday, 2 February 2017 22:39:22 UTC