On Tue, Sep 16, 2003 at 01:50:48PM +0100, Pete Johnston wrote:
> > 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.
> I'm not opposed to the "borrowing" of "units of meaning" and their reuse
> in new contexts with other "units of meaning" from other sources.
> But I do think that when we do this, we have to be very careful and very
> clear about what it is we are borrowing (and yes, I am worried about
> some of the "borrowings" that I see appearing in "Application
> Roland made the points the other day
> that we _must_ _not_ confuse an XML element and an RDF property, and
> that we _must_ take care in how we interpret the conventions used in an
> arbitrary XML encoding.
I appreciate the nervousness about promiscuous borrowing, but
I would assert that the problem lies not with the "semantics"
per se (assuming they are defined in generic ways that do not
somehow inextricably bind them into a nesting construct of some
data model), but with the model within which they are embedded.
One of the important functions of the Abstract Model is to
serve as a common reference point for many different kinds
of encoding syntaxes and practices -- some of which will be
able to support the full Model and others of which (like HTML)
will inevitably fall short.
I am assuming that an XML structure which arbitrarily nests
things here and there without making the entity distinctions
explicit in the way called for by the Abstract Model would
fail the conformance test.
Bottom line: it is not the borrowing of semantics per se
that we need to worry about, but the extent to which ad-hoc
metadata constructs fall short of the Abstract Model. A DCAP
could perhaps serve as a "corrective lens" to such ad-hoc
models by functioning, in effect, as a "view", or "mapping",
from the ad-hoc-modeled metadata to a clean projection which
conforms to the Abstract Model.
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]