> With no intented hair-splitting, in which document is it formally explicited?
I’m looking at
From: DCMI Architecture Forum [mailto:[log in to unmask]] On Behalf Of Bernard Vatant
Sent: 02 September 2009 11:18
To: [log in to unmask]
Subject: Re: rdfs:range for dcterms:subject and other "open-range" properties
Hi Andy
2009/9/2 Andy Powell <[log in to unmask]>
> I understand therefore your take on this as being a non-issue, and it's up to applications to deal with the absence of range.
That is my understanding (except that I disagree with your assertion that the range is missing – it’s not missing, it is explicitly set to rdfs:Resource).
With no intented hair-splitting, in which document is it formally explicited? It's neither in http://dublincore.org/documents/dcmi-terms/ nor in http://dublincore.org/2008/01/14/dcterms.rdf as far as I can see.> OTOH, as an informative note, could DC suggest/recommend as a best practice to use some classes, such as skos:Concept?
I tend to agree, and think there’s probably room for this kind of guidance in
Guidelines for Dublin Core Application Profiles (Working Draft)
http://dublincore.org/usage/meetings/2008/09/berlin/dcap-guidelines/
I guess SKOS folks would be very happy with that.
(Note: I’m making the assumption here that this is indeed recommended best practice!).
It's indeed recommended or strongly suggested from the SKOS viewpoint. See http://www.w3.org/TR/skos-primer/#secindexing
"While the SKOS vocabulary itself does not include a mechanism for associating an arbitrary resource with askos:Concept
, implementors can turn to other vocabularies. Dublin Core, for instance, provides adct:subject
property ..."
Actually as a reminder there used to be in the first SKOS versions a skos:subject property which was deprecated under the rationale that it was already defined by DC.
Bernard