On Thu, Sep 11, 2003 at 04:25:19PM -0400, Diane Hillmann wrote:
> At 04:46 PM 9/11/2003 +0100, Rachel Heery wrote:
> >As I understand it there is a fundamental mis-match in the metadata model
> >for IEEE LOM and DC. I am suggesting we mandate that the DQ model insists
> >all data elements must adhere to the DC model as regards the sorts of
> >relationship between properties. (The data elements need not be properties
> >with DCMI namespaces) This would exclude IEEE LOM data elements.... it
> >would include 'novel' DC-like data elements. Not that I dislike IEEE LOM
> >data elements, just that they are covered by another model.
>
> Oh, oh. I think this would be problematic. What about the DC-Libraries
> Application Profile that uses the MARC Holdings field for location for
> physical objects? Or DC-Ed, which has designated three IEEE LOM elements
> (all of which are included in the latest version of the NSDL qualified DC
> schema, by the way).
>
> I think we're well down that road, and I don't see it as a particular
> problem--what we're using of those elements is the definition, not the data
> model they come from. [snip]
I agree with Diane here. DCMI, MARC21, and IEEE/LOM all
signed the CORES Resolution, which starts off with the shared
understanding that "our metadata standards have 'elements'
-- units of meaning comparable and mappable to elements of
other standards" [1].
The D-Lib Magazine article [1] summarizes a very interesting
discussion in the IEEE/LOM community on whether a given
"element" is really the same (i.e., should be identified
with the same URI) when nested in different containers,
or whether their meanings are context-specific enough to
require assigning to them different URIs. As of press time
in July, this issue had not been fully resolved.
Borrowing "units of meaning" from vocabularies associated
with other underlying data models and using them outside their
original context seems like a natural and somehow inevitable
development -- a bit like English borrowing the word "tofu"
even though Chinese has a different grammar. We could try
to keep the boundaries neat by insisting that each particular
grammar have its own, non-overlapping vocabulary, but as Diane
says, this would most likely just create alot of confusion.
One would of course need to take care not to infer a contextual
data model from an element alone:
-- the presence of http://purl.org/dc/elements/1.1/title in a
record that consisted otherwise of 50 LOM elements would not
allow one to infer that the record follows DCMI principles,
and
-- the presence of http://ltsc.ieee.org/LOM/v1.0/MetadataIdentifier
in a record would not in itself allow one to infer the
existence of a "Meta-Metadata" containment structure.
Tom
[1] http://www.dlib.org/dlib/july03/baker/07baker.html
--
Dr. Thomas Baker [log in to unmask]
Institutszentrum Schloss Birlinghoven mobile +49-160-9664-2129
Fraunhofer-Gesellschaft work +49-30-8109-9027
53754 Sankt Augustin, Germany fax +49-2241-144-2352
Personal email: [log in to unmask]
|