Frank A. Roos wrote:
>
> It seems to me that DC.Type will always be present in a DC "record"
Yes - if it is a "complete" record.
But there are no mandatory fields in DC.
> To generalize this thought, but maybe a bit off topic, although worthwhile I
> think: it seems to me that each different type, is accompanied by a different
> combination of the other DC fields. Maybe some fields are even mandatory to
> make a type belong to a certain type (see for instance the "refer" and
> "BiBTeX" bibliographic tools, where you have first to define type before
> adding the (mandatory) fields to a record). So, type triggers a couple of
> mandatory DC fields to be added to the DC record, that can be identified
> beforehand as typical for that type. "Event" may at least need DC.coverage (in
> most cases; does anybody know an example where an "event" does not have a
> temporal or spatial aspect?), but undoubtedly also other fields.
>
> One may even want to categorize the needed fields for each type.
I suspect that it would be possible to deduce a set of
rules along these lines. That amounts to modelling the
classes corresponding to the universe of "resources".
Interesting, but almost impossible to do exhaustively.
Hence, an early decision was made in DC to allow all elements
to be present between zero and many times in all instances.
--
__________________________________________________
Dr Simon Cox - Australian Geodynamics Cooperative Research Centre
CSIRO Exploration & Mining, PO Box 437, Nedlands, WA 6009 Australia
T: +61 8 9389 8421 F: +61 8 9389 1906 [log in to unmask]
http://www.ned.dem.csiro.au/SimonCox/
|