W3C home > Mailing lists > Public > public-esw-thes@w3.org > October 2005

RE: SKOS 'concept schemes' and DCMI 'vocabulary encoding schemes'

From: Miles, AJ \(Alistair\) <A.J.Miles@rl.ac.uk>
Date: Mon, 10 Oct 2005 17:12:28 +0100
Message-ID: <677CE4DD24B12C4B9FA138534E29FB1D0ACDC1@exchange11.fed.cclrc.ac.uk>
To: "Pete Johnston" <p.johnston@UKOLN.AC.UK>, <DC-RDF-TASKFORCE@JISCMAIL.AC.UK>
Cc: <public-esw-thes@w3.org>

Hi Pete,

> Without taking a position one way or the other ;-) I'd just note that
> "pattern 2" is embedded in the DCAM
> 
> http://dublincore.org/documents/abstract-model/
> 
> > Each resource may be a member of one or more  classes. Note 
> that where
> the resource is a value, the class is referred to as a vocabulary
> encoding scheme.
> 
> so it's not just the mapping to RDF which would have to change.
>  

Yup :)  

> Also I don't think that the examples of "pattern 1" in the 
> draft DCQ-RDF
> doc ever say that the object of the rdfs:isDefinedBy triple is the URI
> of the scheme. 
> 
> Section 2.3.2 suggests that "URI1" in that example is _not_ the URI of
> the scheme: the URI of the Class is the URI of the scheme 
> ("If there is
> some RDF Class given to identify the Scheme....").

Quoting from [1] section 2.3 ...

'In the case where there is a URI, specifying the object we want to use in it's relation with the scheme, we could make an rdfs:isDefinedBy arc pointing to that URI. Such a triple of RDF(S) properties hanging off a resource is what one may call: Poor Man's Structured Values'

?

Oh, another thought, the only way I can see 'Vocabulary Encoding Scheme' mapping to RDF as is, without changing the AM, is to model them as RDF datatypes.  I.e.:

<http://www.example.com/somedoc> dc:subject 'D08.586.682.075.400'^^dcterms:MESH.

Although I certainly can't claim to understand the finer points of how an RDF datatype maps a set of literal values to resources, the notion of a datatype in RDF seems to fit best with the notion of 'encoding' as it is described in the DCMI AM.

Would it be possible to allow both e.g. ...

<http://www.example.com/somedoc> dc:subject 'D08.586.682.075.400'^^dcterms:MESH.

... and e.g. ...

<http://www.example.com/somedoc> dc:subject <http://purl.org/dc/terms/MESH/D08.586.682.075.400>.

... where there is also ...

<http://purl.org/dc/terms/MESH/D08.586.682.075.400>
  rdfs:label 'Formate Dehydrogenases';
  rdf:value 'D08.586.682.075.400';
  rdfs:isDefinedBy dcterms:MESH;
.

... to live side-by-side?  In this case, the last bit of RDF could be used to define the literal-to-resource mapping via the rdf:value (or some other) property?  And it would seem reasonable to then also state things like...

<http://purl.org/dc/terms/MESH/D08.586.682.075.400>
  a skos:Concept;
  skos:prefLabel 'Formate Dehydrogenases';
  skos:altLabel 'Formate Hydrogenlyases';
  skos:altLabel 'NAD-Formate Dehydrogenase'; 
  skos:scopeNote 'Flavoproteins that catalyze reversibly the reduction of carbon dioxide to formate. Many compounds can act as acceptors, but...'; 
  skos:historyNote '91(80); was see under ALDEHYDE OXIDOREDUCTASES 1980-90';
  skos:broader <http://purl.org/dc/terms/MESH/D08.811.682.075>;
.

...?

Also, the pattern:

<http://www.example.com/somedoc> dc:subject 'D08.586.682.075.400'^^dcterms:MESH.

... seems to also match closely the XML encoding of qualified DC, using xsi:type.

That's all I have for now, cheers,

Al.

[1] http://dublincore.org/documents/2002/05/15/dcq-rdf-xml/

>  
> Pete
> 
Received on Monday, 10 October 2005 16:14:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:54 GMT