 |
 |
Hi Tom, On 8 March 2012 13:37, Thomas Baker <[log in to unmask]> wrote: > On Thu, Mar 08, 2012 at 08:57:22AM +0000, Pete Johnston wrote: >> And from >> >> http://dublincore.org/documents/dcmi-terms/#H8 >> >> A DCAM VES is >> >> > An enumerated set of resources >> >> i.e. a DCAM VES is a set/aggregation of things of any type (concepts, >> people, places, documents) whereas a SKOS concept scheme is an >> aggregation of concepts/conceptual things. > > Ah, but SKOS Concepts are not formally disjoint with non-conceptual things. > Put another way, by saying that something is a SKOS Concept one is not actually > saying that it is a "conceptual thing". If a SKOS Concept is informally > understood to be a conceptual thing, that is fine, but the SKOS data model does > not make a formal distinction. OK :) But I think in practice we start to run into some peculiar situations if we end up saying that a resource like http://id.loc.gov/authorities/subjects/sh85102557 has a "has-broader-concept"/"is-narrower-concept-than" relation with http://id.loc.gov/authorities/subjects/sh85103405 and that same thing also has a "is-located-in" relation with http://dbpedia.org/resource/Paris (especially if people start applying "dates of creation" etc (which I appreciate the LoC example does not)) I don't object to the suggested change - I think it's what we should have done in the first place (despite my probably having argued to the contrary at the time!) But for myself, I would shy away from using the construct where it might result in these sorts of mental gymnastics! Pete -- Pete Johnston