- From: Matthias Samwald <samwald@gmx.at>
- Date: Wed, 9 Jul 2008 12:14:29 +0200
- To: "semantic-web at W3C" <semantic-web@w3c.org>, <public-lod@w3.org>
> In general, I would encourage people to work on developing a more > sophisticated alignment vocabulary. At the moment the only such vocabulary that is on the way towards (at least potential) widespread acceptance is the mapping vocabulary of SKOS [1]. Maybe this vocabulary could become of importance for the LOD community when it turns out that the quality of the owl:sameAs links between datasets is too questionable, and that rdfs:seeAlso is too loose? Of course, the domain and range of these properties as defined in the SKOS core schema are restricted to skos:Concept at the moment, which discourages use of these properties for other entities. On the other hand, I already observed in some other projects that useful RDF properties from SKOS (such as skos:prefLabel) are also used for entities other than skos:Concepts... probably not an elegant solution, but at least something to keep in mind. In the case of classes and properties: what is wrong with owl:equivalentClass and owl:equivalentProperty for gluing together the LOD cloud? They don't 'smush' entities together as owl:sameAs does, but the recent discussion gives the impression that this (meaning: less aggressive 'smushing' of resources) might actually be a positive behaviour. [1] http://www.w3.org/TR/2008/WD-skos-reference-20080609/#mapping Cheers, Matthias Samwald Semantic Web Company, Austria // DERI Galway, Ireland
Received on Wednesday, 9 July 2008 10:15:17 UTC