Exactly what is the difference between "guerrilla DC" and
"extensibility," which is supposed to be one of the major features of
DC?
>Well, I didn't think "extensibility" meant that anyone could
(mis)appropriate the DC prefix. DC is not a synonym for any metadata - it's
a precise set of 15 elements (at least in its 1.0 incarnation). Local
extensions to DC should start from these elements. For example, if there is
a need for "substance", this should either be expressed as
"DC.Subject.Substance" or as something like "IC.substance" or
"IC.chemsubstance", or whatever (where "IC" = Imperial College).
>I don't see how the IC set can be called "extensions to the Dublin Core"
when they bear no relationship to the DC elements.
>I would say that the common practice amongst implementors to date has been
to use the "dot kludge" to add local subelements when the DC set suffices
at the element level, and to use local prefixes when the DC set just
doesn't seem to fit. (For example, if it's still not clear where "audience"
goes [with arguments to date suggesting DC.Subject or DC.Coverage or, if in
doubt, good old DC.Description], some implementors have used their own
prefix for a specific "audience" element tag. What they haven't done is
called it something like "DC-GEM.audience".)
>I guess without trademarking "DC", there's nothing we can do to stop
implementors using "DC" in their own tags. I could have a set called DC-CM
if I wanted. When it comes to software interpreting what I mean by my DC-CM
tags, I guess that as long as I refer to a DC-CM namespace, that will be
all right, will it?
>Cliff
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|