Pete,

Looking again at these, I'm wondering why you're not using the xmlns
attributes for declaring namespaces. I know you don't want to use the
CURIE specification directly, but even RDFa does use the xmlns
attributes, see for example

http://www.w3.org/TR/xhtml-rdfa-primer/#id2261651

Using that would decrease the complexity somewhat.

Also, there are still cases of dc-title and similar in the examples
(should be dc:title).

/Mikael

tor 2007-07-19 klockan 11:08 +0100 skrev Pete Johnston:
> I've created an updated version of the XML format which I was calling
> DC-XML-Full to reflect the recent revision of the DCMI Abstract Model
> (2007-06-04) [1]:
> 
> http://dublincore.org/architecturewiki/DCXMLRevision/DCXMLFGuidelines/20
> 07-06-19
> 
> The DC-XML-Full format supports the complete DCAM description model.
> 
> The main changes are:
> 
> (i) introducing the distinction between literal value surrogates and
> non-literal value surrogates. In DC-XML-Full this is reflected in
> differences in both the attributes and the child elements of the
> Statement Element. 
> (ii) removing support for rich representations.
> (iii) revising support for value strings to distinguish plain and typed
> literals, and to include support for XML Literals as typed literals.
> 
> These changes have required a fairly substantial re-structuring of the
> latter parts of the document.
> 
> I also made a minor change to the convention for abbreviating URIs to
> use a ":" instead of a "-". I've labelled this a "Prefixed Name" and
> referred to the SPARQL specification for the syntax, but the description
> of how to map to a URI is defined in ths document.
> 
> Also:
> 
> A W3C XML Schema. See
> 
> http://dublincore.org/architecturewiki/DCXMLRevision/DCXMLFBaseSchemas/2
> 007-06-19
> 
> A (first cut at a) RELAX NG Schema.  See
> 
> http://dublincore.org/architecturewiki/DCXMLRevision/DCXMLFBaseSchemas/2
> 007-06-19
> 
> An XSLT transform to map DC-XML-Full instances to RDF/XML (i.e. to
> support GRDDL [2] for the DC-XML-Full format) using the DCAM-RDF mapping
> described by [3]. See
> 
> http://dublincore.org/architecturewiki/DCXMLRevision/DCXMLFXSLT/2007-06-
> 19
> 
> Some example instances
> 
> http://dublincore.org/architecturewiki/DCXMLRevision/DCXMLFInstances/200
> 7-06-19
> 
> On the topic of GRDDL: that page includes links which enable you to see
> the results of applying the GRDDL transform. There are two links: one
> based on instances which reference the transform directly; and one based
> on discovering the transform via a "namespace document". For the latter
> I had to use a parallel set of examples with a different XML namespace
> name, so bear in mind when you're looking at the examples that that
> isn't the proposed URI. But, hopefully, the second approach will be the
> mechanism used to support GRDDL for the DC-XML-Full format i.e.
> implementers won't need to reference the transform in each instance: in
> fact they never need to worry about the transform process; rather it
> will be implemented by a GRDDL-aware processor "following its nose" to
> find the transform from from that single "namespace document". 
> 
> Some notes:
> 
> 1. The format implements the full DCAM description model
> 
> 2. This format "foregrounds" the DCAM constructs but is very verbose (I
> suppose it occupies a similar point on the landscape as does the
> core/unabbreviated form of TriX [4] for RDF)
> 
> 3. Allowing URIs to be encoded both as URIs and as DC-XML Prefixed Names
> makes it easier for a human to read/write but more complex for an
> application to process.
> 
> 4. The structure of a DC-XML-Full instance is more or less the same as
> that of a DC-Text instance.
>  
> 5. CURIEs: This format doesn't use CURIEs [5]. My understanding is that
> there is still considerable discussion around the CURIE proposal and it
> is by no means certain that it will be deployed in the short term, even
> within RDF/A. See e.g. [6, 7, 8].
> 
> 6. Trying to maintain XML docs as attachments to Wiki pages was a pain,
> so all the examples, test schemas etc are currently available using
> non-DCMI URIs for which I make no promises of their longer term
> persistence! If this is ever given any formal status by DCMI, then
> relevant components will be assigned DCMI-owned URIs.
> 
> Comments welcome! ;-)
> 
> Cheers
> 
> Pete
> 
> [1] http://dublincore.org/documents/2007/06/04/abstract-model/
> [2] http://www.w3.org/TR/2007/PR-grddl-20070716/
> [3] http://dublincore.org/documents/2007/06/04/dc-rdf/
> [4]
> http://www.mulberrytech.com/Extreme/Proceedings/html/2004/Stickler01/EML
> 2004Stickler01.html
> [5] http://www.w3.org/TR/2007/WD-curie-20070307/
> [6]
> http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0705&L=DC-ARCHITECTURE&
> P=R1577
> [7]
> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2007Jun/0054.
> html
> [8]
> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2007Jul/0102.
> html
> 
> ---
> Pete Johnston
> Technical Researcher, Eduserv Foundation
> Web: http://www.eduserv.org.uk/foundation/people/petejohnston/
> Weblog: http://efoundations.typepad.com/efoundations/
> Email: [log in to unmask] 
> Tel: +44 (0)1225 474323
> 
-- 
<[log in to unmask]>

Plus ça change, plus c'est la même chose