"Smith, Allison" wrote:
>
> I am very taken by the idea of collapsing all "agents" into a single field,
> It seems to me that, if DC Creator, Contributor and Publisher get collapsed
> into a singe field, DC would not have 15 elements anymore, but would instead
> have 13. Does this mean we have room to add one or two more fields? This
> is a joke, however, in a similar fashion to what Bernhard has proposed
> below, I would propose the elements this way:
>
> DC.Name (Name of the Creator/Contributor/Publisher/Manufacturer/Firm, etc.,
> in standard AACR2 format, one per tag - repeatable)
> DC.Role (free text - concatonation of the above name, and following it with
> a term chosen from a list of acceptable "Agent terms", such as Creator,
> Contrubutor, Publisher, Manufacturer, Firm.....)
>
> It would look like this:
>
> <DC.Name> Adams, Ansel
> <DC.Role> Ansel Adams, Creator
>
> Since the elements are repeatable, you could further describe Contributors,
> Surrogate creators, etc.
>
> Please explain to me if this is not do-able, and why, because it
> seems like such a simple solution to me.
I believe doing this would create something that is not compatible
with the Dublin Core Element Set (DC 1.0) as defined in RFC2413.
The DC 1.0 schema is a list of properties, all of which describe
to a common resource. Each property is a tuple, consisting of
a Property Name (e.g. "DC.Creator") and a Value (e.g. "Ansel Adams").
As far as I am able to tell, the way you propose to handle the so-called
"agent terms" will break this schema, and put in its place a different
schema where two named properties (DC.Name and DC.Role) are linked
through some common value element to form a triple (consisting of the
pseudo-property "agent", Value (e.g. "Ansel Adams") and Role
(e.g. "Creator").
Unfortunately, this would, among other things, utterly confuse any
software that has been written to process DC 1.0 metadata.
I understand that there is a need to extend DC 1.0 to include such
things as roles -- but it must be done in a manner that is compatible
with DC 1.0. I.e. you may hang additional stuff onto the data model
to qualify elements with things such as roles, but the basic tuple
must be preserved so that software who doesn't understand (or don't
want to bother with) qualifiers can extract the DC 1.0 and ignore
the qualifiers.
Fortunately, the syntax for type and value qualifiers currently
proposed by the DC data model working group satisfies this
criteria, so all is not hopeless :-) .
To me, at least, the entire field of extending DC beyond the basic
data model implicit in RFC2413 seem to be a very complex area -- one
I would like to see it pursued with extreme caution.
--
- gisle hannemyr ( [log in to unmask] - http://home.sol.no/home/gisle/ )
------------------------------------------------------------------------
"Use the Source, Luke. Use the Source." -- apologies to Obi-Wan Kenobi
------------------------------------------------------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|