Picking up Roland's point here....
> Without a given literal to value map the datatype encoding
> style produces a serious problem to dc-dumbdown.
>
> A dc:subject content 061(40) in my view is not to be
> recommended as a dumbdown from dc:subject scheme="UDC"
> content="061(40)".
>
> This is NOT a syntax issue, as you seem to imply.
>
> The xml guideline suffers from the same problem.
I think I understand the point Roland is making here (which I agree is
syntax independent): in
> dc:subject content 061(40)
the literal string "061(40)" is the value of the attribute.
In
> dc:subject scheme="UDC" content="061(40)"
the value of the attribute is the result of a transformation process
performed on the literal "061(40)", where the transformation is defined
by the scheme/datatype.
This became particularly apparent in the (syntax-specific) discussion of
the xs:anyURI datatype and escaped characters and so on
http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0210&L=dc-architecture&T=
0&F=&S=&P=7120
and follow-ups (when I undertook to explore some examples which I fear I
still haven't had time to do).
But I suspect that many implementations of dumb-down in DC metadata have
not taken this "datatype-aware" approach, and _have_ simply discarded
the "scheme" information. This certainly seems to be the behavior
suggested by
http://dublincore.org/documents/usageguide/#dumbdown
Are we saying this blunter, less sophisticated approach to dumb-down is
incorrect?
Pete
|