I meant to add that also, for any non-RDF/XML XML format used, we plan to provide a GRDDL transform to generate RDF/XML, so an RDF version of the description set will always be available.
Pete
-----Original Message-----
From: DCMI Architecture Group on behalf of Pete Johnston
Sent: Tue 9/26/2006 1:09 AM
To: [log in to unmask]
Subject: Re: Proposal for a minimal description model subset
Hi Bruce,
Thanks for the comments. I think Mikael did make a similar point way back when we initially discussed the idea of supporting a subset of the DCAM description model, and I must admit it had slipped my mind when we resuscitated the subset notion.
At first I was going to agree with you. However, although the two examples you give are structurally similar, I don't think we intend the two structures to be semantically equivalent.
In your RDF/XML example, the literal content of the Property Elements represents RDF literal objects e.g. (I'm choosing dc:subject rather than dc:title as it illustrates the point better, I think), the RDF/XML fragment
<rdf:Description rdf:about="http://dublincore.org/pages/home">
<dc:subject>History</dc:subject>
</rdf:Description>
represents an RDF triple with a plain literal object:
<http://dublincore.org/pages/home> dc:subject "History" .
The resource has-as-subject the string "History".
But in the DC-XML-Min example the intent is that the fragment
<dcxm:description dcxm:resourceURI="http://dublincore.org/pages/home">
<dc:subject>History</dc:subject>
</dcxm:description>
represents a DC statement that says the resource has-as-subject is a resource of unspecified type, represented by the string "History", which in RDF would be represented by a blank node.
So I was about to agree with you and say we should replace
dcxm:description -> rdf:Description
@dcxm:resourceURI -> @rdf:about
@dcxm:valueURI -> @rdf:resource
@dcxm:vocabEncSchemeURI -> @dcrdf:inScheme or @dcrdf:isMemberOf (or whatever we decide the property for that relationship is)
@dcxm:valueClassURI -> @rdf:type
@dcxm:descriptionId -> @rdf:nodeid
@dcxm:descriptionRef -> @rdf:nodeid
and end up with a profile of RDF/XML.
But I'm not convinced that would be sufficient, because the RDF/XML Property Element content generates triples with literal objects.
I think we'd have to use something like
<rdf:Description rdf:about="http://dublincore.org/pages/home">
<dc:subject rdf:parseType="Resource"><dcrdf:valueString>History</dcrdf:valueString></dc:subject>
</rdf:Description>
to get the blank node, but that reintroduces the child element that people objected to.
Or alternatively using RDF/XML Propert Attributes, then
<rdf:Description rdf:about="http://dublincore.org/pages/home">
<dc:subject dcrdf:valueString="History" />
</rdf:Description>
generates the same triples.
But using this RDF/XML pattern we can't associate either a language or a datatype with the value string as attribute value, and I suspect that further reduced subset may be a cut too far.
Pete
-----Original Message-----
From: DCMI Architecture Group on behalf of Bruce D'Arcus
Sent: Mon 9/25/2006 11:31 PM
To: [log in to unmask]
Subject: Re: Proposal for a minimal description model subset
Hi,
I'm new to the list (more on that later), and so may be asking a
question whose answer has a history I'm not aware of, but why the "try
as hard as possible to not be RDF but still retain the model" format?
What value does this ...
<dcxm:descriptionSet
xmlns:my="http://my.example.org/terms/"
xmlns:dcxm="http://dublincore.org/xml/dc-xml-min/2006/09/18/"
xmlns:dc="http://purl.org/dc/elements/1.1/">
<dcxm:description
dcxm:resourceURI="http://dublincore.org/pages/home">
<dc:title>DCMI Home Page</dc:title>
<!-- Reference to related description using label -->
<dc:publisher
dcxm:descriptionRef="DCMI"/>
</dcxm:description>
... have over this?
<rdf:RDF
xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:dc="http://purl.org/dc/elements/1.1/">
<rdf:Description
rdf:about="http://dublincore.org/pages/home">
<dc:title>DCMI Home Page</dc:title>
<!-- Reference to related description using label -->
<dc:publisher rdf:resource="DCMI"/>
</dcxm:description>
They are, after all, structurally the same, and creating an XML schema
to validate one is no more difficult than the other (notwithstanding
XML Schema's lack of support for attribute-based validation and
unordered content models; limitations that RELAX NG does not have).
Why, in other words, not just create an RDF/XML profile that achieves
the same effect (easily processed with XML tools)?
Bruce
|