> A value results when an element or element refinement is used
> to describe a resource.
This definition -- and the related definitions of "value URI"
and "value string" -- feel like a weak point in the draft.
It doesn't really say what a value _is_ but merely alludes
to how one comes about. (This is arguably a problem with
the definition for "encoding scheme" as well, not just in
your draft but in the approved Principles document on which
it is based.)
I suppose this vagueness has its uses, but it is also
potentially confusing, especially for readers unfamiliar with
this sort of jargon.
One way to address this might be to back up one step and
show how these pieces fit into a grammar of "statements"
or "assertions". For example, my D-Lib article of October
2000 defined Dublin Core as "a small language for making
a particular class of statements about resources" .
It described the structure of such statements by using
a deliberately bone-simple, "grammar-school" variety of
sentence diagram with an implied subject ("the resource"),
an implied verb ("has"), one of fifteen properties, and a
"property value (an 'appropriate literal')", with optional
qualifiers hanging below. The Appendix of the paper explained
how this mapped onto the RDF jargon of (explicit) subject,
predicate, and object.
In the current draft of the Abstract Model I find myself
missing the basic (and I think helpful) point that we are
talking here about assertions about resources, and those
assertions have a fundamental structure recognizably similar
to sentences we all speak.
Defining "value" as part of a statement structure would help
say what it actually _is_, and this would help put the other
facets of value ("value URI" and "value string") into context.
Judging from the definitions alone, for example, I found
myself wondering whether a "value string" was intended to be
identified by a "value URI" (on the logic that if a value
string "represents" a value, a value URI is "identifying"
the value's "representative"...).
Providing a statement model as an overall frame could perhaps
help avoid such confusion. Of course, that statement model
would now maps much more completely onto the RDF model than
the literal-value-based model of the D-Lib paper.
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]