Tom,
> In my opinion, the point is not so much which terms are used
> -- whether they are owned and maintained by DCMI or not --
> but what underlying grammatical typology (e.g., elements
> versus encoding schemes) or Abstract Model is being used.
>
> The CEN Guidelines for Dublin Core Application Profiles ([1],
> to be discussed in a separate posting) say: "By definition, a
> DCAP is based in part on Dublin Core and follows DCMI
> Grammatical Principles [2]."
>
> I don't see how one could be more specific than "based in
> part" (we don't want to go down the path of deciding whether
> it takes at least 1, or 5, or 15...). In my opinion, the
> emphasis on the model is what is important here.
So... taking one of Diane's specific examples....
The DC Libraries AP borrows an "element" called "location"
http://dublincore.org/documents/2002/09/24/library-application-profile/#
Location
from the Metadata Object Description Schema (MODS)
http://www.loc.gov/standards/mods/
MODS is defined as an XML schema. It defines an XML tree structure, and
its "elements" are (I think) XML elements. As far as I can tell, MODS
has no other explicit data model outside the tree structure, it makes no
references to properties of resources or to URIrefs (and certainly not
to elements-in-the-DCMI-sense and encoding schemes), only to XML
elements and XML attributes (and their "structural" relationships in the
tree) and to XML QNames.
I do wonder what exactly is being borrowed here? A "unit of meaning"?
Maybe, but to what extent is that "unit of meaning" defined by MODS
"reusable" in this way if it is defined as part of a MODS tree
structure?
Apologies, DC-Lib people, I'm not singling out DC-Lib AP for criticism,
and I'm sure I could find other examples - I'm not even sure I can
provide clear answers to my own questions. And maybe this is a
non-problem, but these are the sort of issues where I feel we are not
entirely on firm ground.
Pete
|