Rachel,
Sorry I am late getting back to you on this. Please see my comments below:
> -----Original Message-----
> From: Rachel Heery [mailto:[log in to unmask]]
> Sent: Thursday, December 06, 2001 9:22 AM
> To: Wagner,Harry
> Cc: [log in to unmask]
> Subject: Re: FW: Problem with our RDF schemas
>
>
> On Thu, 6 Dec 2001, Wagner,Harry wrote:
>
> >
> >
> > I'm not sure who is working on this, but I would like to
> recommend the
> > following changes to our 3 schemas:
> >
>
> Can we start with a sanity check :-) As I understand it we
> want the DCMI
> 'published' schema(s) to contain fullest information needed
> for purposes
> of imparting information about the definition of the DCMI
> terms and the
> relationship between terms. This is because the schemas that
> DCMI produce
> about the terms in DCMI namespaces need to fulfill requirements of
> applications that use those schema whether those applications are
>
> - metadata editors
> - human readable schema registries
> - dictionaries of metadata terms
> - metadata conversion tools etc etc
>
>
> So on that basis I have a couple of comments:
>
> > - all terms have rdfs:label, rdfs:comment and rdfs:isDefinedBy tags
> > - we remove the eor: tags completely. The terms currently use an
> > eor:comment. This text could be moved to (combined with)
> the rdfs:comment
> > tag.
>
> In effect this would mean combining the Description and the
> Comment fields
> that now appear separately within the DCMI reference documents at
>
> Dublin Core Metadata Element Set,
> Version 1.1: Reference Description
> http://www.dublincore.org/documents/dces/
>
> and
>
> Dublin Core Qualifiers
> http://www.dublincore.org/documents/dcmes-qualifiers/
>
> Given there is a qualitative difference between the
> information in these
> 2 fields (i.e. one is definition, the other comment on how
> data element
> should be used) I feel combining the fields is a retrograde step.
>
> For sure, it is somewhat confusing that the RDFS forces a
> 'definition' of
> a term
> into a field tagged 'comment'. I'm not sure if there is
> anything happening
> within the W3C RDFS activity, or in DAML-OIL that might help?
> Are there
> proposals tthat would enable a 'definition' tag as well as a 'comment'
> tag??
If we want to keep a 'description' tag seperate from the 'comment' tag we
should choose a better mechanism than 'eor:comment'. Our registry is going
to infuse schemas from other resource communities, right? I doubt they will
use 'eor:comment" for this purpose.
>
> > - remove the eor:tags used to define the schemas themselves, as in:
> >
>
> I have not gone thro all of the schemas identifying use of
> the eor tag....
> are you suggesting replacing the prefix 'eor' with something
> else or do
> you think all these tags are redundant..... I'm not sure i
> understand the
> motivation.
>
>
> > <eor:Schema rdf:about="&dcns;">
> > <rdf:value>The Dublin Core Element Set v1.1</rdf:value>
> > <dc:title>The Dublin Core Element Set v1.1</dc:title>
> > <dc:publisher>The Dublin Core Metadata Initiative</dc:publisher>
> > <dc:description>The Dublin Core metadata vocabulary is a
> simple vocabulary
> > intended to facilitate discovery of resources.
> </dc:description>
> > <dc:language>English</dc:language>
> > <dc:relation
> rdf:resource="http://dublincore.org/documents/dces/"/>
> > <dc:date>2000-07-02</dc:date>
> > </eor:Schema>
> >
> > This information should be captured by the registry
> application when it is
> > imported (and built into a very similar model), but does
> not need to be part
> > of the schema.
> >
>
> As I see it any expression of a schema would usefully contain
> this sort of
> administrative metadata... such as who wrote it, significant
> dates, name
> of vocabulary etc. Are you suggesting we use another prefix
> for the tag
> or remove all the information?
I don't see a reason to have it. This information is for human consumption
and could be conveyed by the registry without it being part of the schema.
If we do make it part of the schema we should pick a better way identify it
as a schema than 'eor:Schema'. Again, we don't expect anyone else to code
this in their schema do we?
Regards,
Harry
>
> I'm not sure I understand why you would want to leave such information
> out of the schema?... the DCMI schema might be imported by a registry,
> for example, so that that particular schema could be
> navigated along with
> several other schema. In which case the registry might want
> to provide an
> export functionality whereby that schema is exported. Or the registry
> might more efficiently point at schemas in distributed way
> and index them
> centrally.
>
> Whichever way the information in the schema needs to include
> the admin metadata??
>
>
> > Please let me know what you think about these
> recommendations. If no one is
> > already working on cleaning up our schemas I would be happy
> to mock up a
> > draft on our test site and distribute it for this groups approval.
>
>
> As part of this I think we need to reach a definitive view as to
>
> 'should schemes be associated with particular terms?'
>
> and
>
> 'should element refinements be associated with elements?'
>
> As I see it we should retain these relationships and furthermore these
> relationships should be expressed by the RDF schema.
>
> I guess these last 2 issues need a wider debate whether than
> just within
> dc-architecture WG, perhaps including Usage Board or perhaps using the
> 'recommendation process'. If that debate does start within
> this group I
> guess we just let others know that its happening.
>
> rachel
>
>
>
>
>
>
>
> >
> > Regards,
> > Harry
> >
> > > -----Original Message-----
> > > From: Rachel Heery [mailto:[log in to unmask]]
> > > Sent: Monday, November 26, 2001 1:24 PM
> > > To: Thomas Baker
> > > Cc: Wagner,Harry; 'Stuart Sutton'; Rachel Heery; DCMI Executive
> > > Committee; Eric Miller; Diane I. Hillmann ([log in to unmask])
> > > Subject: RE: Problem with our RDF schemas
> > >
> > >
> > > On Mon, 26 Nov 2001, Thomas Baker wrote:
> > >
> > > > As I see it, we would then get the sixteen eyes of the
> > > Usage Board to
> > > > pore over the details, agree on the introductory words,
> and approve
> > > > the result. Then we have a canonical version of Dublin
> > > Core semantics
> > > > in HTML that can be used for a revision of the RDF schema.
> > >
> > > Good!
> > >
> > > Keep in mind that the HTML version can be expressed in RDF in
> > > different ways, all of which may be 'well-formed RDF' but the
> > > different
> > > forms of expression may be more or less expressive of the
> > > structure of the
> > > DCMI vocabularies.
> > >
> > > So 'revising the RDF schemas' needs to take account of the
> > > structure DCMI
> > > which to express in the RDF schema. This is particularly true as
> > > recommendations for the 'instance metadata' (the syntax for
> > > DC matadata
> > > records) suggested in recent mails will not distinguish
> elements from
> > > qualifiers.
> > >
> > > So translating the HTML schema to RDF schema will need
> some careful
> > > thought...
> > >
> > > Rachel
> > >
> > > >
> > > > Tom
> > > >
> > > >
> > > ______________________________________________________________
> > > _________________
> > > > Dr. Thomas Baker
> > > [log in to unmask]
> > > > Birlinghoven Library, Fraunhofer-Gesellschaft
> > > > Institutszentrum Schloss Birlinghoven
> > > +49-2241-14-2352
> > > > 53754 Sankt Augustin, Germany
> > > fax +49-2241-14-2619
> > > >
> > > >
> > > >
> > >
> > >
> > > --------------------------------------------------------------
> > > -------------
> > > Rachel Heery
> > > UKOLN
> > > University of Bath tel: +44
> > > (0)1225 826724
> > > Bath, BA2 7AY, UK fax: +44
> > > (0)1225 826838
> > > http://www.ukoln.ac.uk/
> > >
> > >
> >
>
>
> --------------------------------------------------------------
> -------------
> Rachel Heery
> UKOLN
> University of Bath tel: +44
> (0)1225 826724
> Bath, BA2 7AY, UK fax: +44
> (0)1225 826838
> http://www.ukoln.ac.uk/
>
>
>
|