> I define a record as "some structured metadata about a
> resource, comprising one or more properties and their
> associated values". It doesn't seem to me that my definition
> is much different to Pete's, which I think you are agreeing
> with?? Perhaps I'm missing something?
I certainly didn't intend to suggest a difference between my description
and yours ;-)
Rather, I was trying to clarify the relationship (in my mind at least)
between the "metadata record" view and Stu's "collection of assertions"
Taking it a bit further, I'd see a QDC/SDC metadata record as an
instance of a specific type/class of metadata record (a set of
assertions) in which the assertions made were limited to the use of
specific properties - which (as you say) is typically how many metadata
applications have co-ordinated the sharing of data.
Maybe (I'm not sure) that "typing" of metadata records has less of a
role within a framework which doesn't depend on that sort of
pre-coordination, so maybe (again, as you suggest) it needs to be a
presented as a "refinement" (eek) of some more generic level. But I
think I'd be reluctant to lose it completely.