Sorry, I'm catching up with all this a bit late.....

Rachel said:

> 1. Amending treatment of classes in schemas
> --------------------------------------------
> The following UI changes rely on amendment of the way Classes are dealt
> with in
> http://purl.org/dc/terms/
> http://purl.org/dc/dcmitype/
>
> Within http://purl.org/dc/terms/ class is used to associate schemes with
> elements, whereas in
> http://purl.org/dc/dcmitype/ class is used to describe a controlled
> vocabulary for resource. This needs to be reflected by
> - defining a new class 'Scheme' within http://purl.org/dc/terms/ and
> - defining a new class 'ControlledVocabulary' within
> http://purl.org/dc/dcmitype/
>
> Then
>
> - all subClasses within http://purl.org/dc/terms/ are subClasses of
> Scheme, and
> - all subClasses in http://purl.org/dc/dcmitype/ are subClass of
> ControlledVocabulary.
>
> Thus
>
> - a subClass of SubjectScheme would also be a subClass of  Scheme, and
> - a subClass of Resource would be a subClass of ControlledVocabulary.

Before I try to comment on this, I think I need some clarification on an
aspect of RDF/RDFS which I've never quite managed to grasp completely.

In the DCQ in RDF draft,

http://www.mathematik.uni-osnabrueck.de/projects/dcqual/qual21.3.1/qual29-8-
1.html

the schema in Appendix 5.1 includes:

===========================
<rdfs:Class rdf:about="&dcqns;SubjectScheme">
  <rdfs:label>Subject Encoding Schemes</rdfs:label>
  <rdfs:comment>A set of subject encoding schemes
    and/or formats</rdfs:comment>
  <rdfs:isDefinedBy rdf:resource="&dcqns;" />
</rdfs:Class>

<rdfs:Class rdf:about="&dcqns;LCSH">
  <rdfs:label>LCSH</rdfs:label>
  <rdfs:comment>Instances of this class are Library of Congress Subject
    Headings and must provide an rdf:value (with content the code) and
should
    provide an rdfs:label arc (with content the caption).
  </rdfs:comment>
  <rdf:type rdf:resource = "&dcqns;SubjectScheme" />
  <rdfs:isDefinedBy rdf:resource="&dcqns;" />
</rdfs:Class>
========================

i.e. the relationship between the class dcq:SubjectScheme and the class
dcq:LCSH (an individual subject scheme) is indicated by rdf:type.

In the schema at

http://dublincore.org/2001/08/14/dcq#

(which is where the namespace URI http://purl.org/dc/terms/ resolves to). we
have:

<rdfs:Class rdf:about = "&dcqns;SubjectScheme">
  <rdfs:label>Subject Encoding Schemes</rdfs:label>
  <rdfs:comment>A set of subject encoding schemes and/or
formats</rdfs:comment>
  <rdfs:isDefinedBy rdf:resource = "&dcqns;" />
</rdfs:Class>

<rdfs:Class rdf:about = "&dcqns;LCSH">
  <rdfs:label>LCSH</rdfs:label>
  <rdfs:comment>Library of Congress Subject Headings</rdfs:comment>
  <rdfs:subClassOf rdf:resource = "&dcqns;SubjectScheme" />
  <rdfs:isDefinedBy rdf:resource = "&dcqns;" />
</rdfs:Class>

i.e. the relationship between the class dcq:SubjectScheme and the class
dcq:LCSH (an individual subject scheme) is indicated by rdfs:subClassOf

Leaving aside for the moment the different values of rdfs:comment, are these
two documents saying the same thing in terms of the relationship between the
class dcq:LCSH and the class dcq:SubjectScheme?

I might be wrong but I don't think they are saying the same thing.

Based on my reading of sections 2.3.1 and 2.3.2 of RDF Schema, I think the
first example above says:

"the class dcq:LCSH is an _instance_ of the class dcq:SubjectScheme"

This seems a reasonable statement.

The second says

"the class dcq:LCSH is a _subset_ of the class dcq:SubjectScheme,
and therefore:
an instance of class dcq:LCSH is an instance of class dcq:Subject Scheme"

The latter statement does _not_ seem right to me (A LCSH subject heading is
not a subject scheme).

So, to cut a long story short, I'm increasingly unsure about whether
rdf:subClassOf is the appropriate way to describe the relationship between
dcq:LCSH and dcq:SubjectScheme. I think maybe the relationship should be one
of rdf:type, as in the DCQ in RDF document. But I think I need someone much
more versed in RDF/RDFS to make that judgement!

In any case, before thinking harder about Rachel's proposals, which hinge on
the use of subClassOf, I would like to understand/resolve these
discrepancies between the schema at

http://dublincore.org/2001/08/14/dcq

and the one in the DCQ in RDF draft.

Pete
-------
Pete Johnston
Interoperability Research Officer
UKOLN, University of Bath, Bath BA2 7AY, UK
tel: +44 (0)1225 323619    fax: +44 (0)1225 826838
mailto:[log in to unmask]
http://www.ukoln.ac.uk/ukoln/staff/p.johnston/