On Wed, May 21, 2008 at 08:37:17PM +0100, Andy Powell wrote:
> > > On that basis dcterms:description is *not* limited to being
> > a literal.
> >
> > OK, but then surely dcterms:abstract should be a literal?
>
> Possibly. I guess that is a question for the Usage Board. I don't
> recall our discussions about dcterms:abstract, though I'm sure we had
> them! :-)
This issue has arisen in an analogous way for Simple Knowledge
Organization System (SKOS), an RDF vocabulary and model for
describing concept schemes such as thesauri.
The SKOS working group (W3C Semantic Web Deployment)
is defining the SKOS vocabulary on the basis of OWL Full.
The assumption (or hope) is that the further development of
OWL should, over time, take care of OWL-DL-unfriendly issues.
The SKOS "documentation properties" -- skos:note and
subproperties skos:scopeNote and skos:definition -- have a
range of rdfs:Resource [3]. As the Primer explains [2], these
properties can be used with plain literals, but other patterns
are possible. This was captured as a formal issue [4]:
SKOS currently has 7 "documentation"/"note" properties:
skos:note, skos:scopeNote, skos:defition, skos:historyNote,
skos:editorialNote, skos:changeNote, skos:example.
In the SKOS Core Guide
<http://www.w3.org/TR/2005/WD-swbp-skos-core-guide-20051102>,
3 different design patterns are allowed for these
properties -- "documentation as an RDF literal",
"documentation as a related resource description", and
"documentation as a document reference".
Do we continue to allow these different design
patterns? If we do, we have to accept a complex range
for these properties. Is that OK? How should we formally
specify that?
The issue has recently been resolved with the decision that
SKOS will explicitly allow all three patterns for documentation
properties [5].
The range of dcterms:title is currently on the agenda of
the Usage Board, and the case for dcterms:abstract should be
reconsidered. By leaving the ranges of such terms unspecified,
however, the Usage Board is at least in good company.
This is clearly an issue for RDF vocabularies generally,
where more input and implementation experience are needed.
Tom
[1] http://www.w3.org/TR/2008/WD-skos-reference-20080125/#L1917
[2] http://www.w3.org/TR/2008/WD-skos-primer-20080221/#secdocumentation
[3] http://www.w3.org/TR/2008/WD-skos-reference-20080125/#note
[4] http://www.w3.org/2006/07/SWD/track/issues/64
[5] http://www.w3.org/2008/05/07-swd-minutes.html#item13
--
Dr. Thomas Baker <[log in to unmask]>
Director, Specifications and Documentation
Dublin Core Metadata Initiative
|