mmm... please excuse a slightly peripheral comment, drawing
on the esoterica of library cataloguing.
Library cataloguing distinguishes between an 'author' and a
'statement of responsibility'. The latter is what is
inscibed on the item itself, typically, in this case, on
the title page. It is, to use historic DC parlance,
'intrinsic' to the resource, and for that reason appears in
the 'bibliographic description' (that part of the
catalogue record which draws data from the item in hand).
The 'author' on the other hand is a 'heading' (those parts
of the record which are 'normalised' to facilitate
common access as part of a fabric of authors, subjects and
works). The (putative) role of Authority files is to
manifest that fabric in shared, usable ways. So, I think
the notion of identifier is germane there (and other
members of the list will be closer than I am to initiatives
elsewhere to develop creator identifiers).
Lorcan
(Without going on :-) it is for reasons like this that I am
always puzzled to see the expressed view that DC draws
on/is based on library cataloguing practice. In my view,
one can mount a critique of DC from cataloguing theory and
practice which would share aspects of the critiques of
Carl, Godfrey Rust, and Keith Jeffrey. FRBR is not a fresh
start: it expresses an existing model in new terms with a
view to progressing stuff to a more rational - explicitly
modeled - basis.)
On Wed, 25 Oct 2000 08:38:58 +0800 Simon Cox
<[log in to unmask]> wrote:
> Andy Powell wrote:
> >
> ...
> > However...
> >
> > The name of the creator is *not* a property of the resource being
> > described. It is a property of an 'agent' (a person or organisation) that
> > is related to the resource.
>
> I prefer a slightly different reading: the name of the creator is an *identifier*
> of the related resource. We even find that names usually follow some encoding
> rules, although they are typically not as formal as other identifiers, such as URL's,
> that we commonly use (except in France, perhaps :-).
>
> Now it certainly can be argued that identifiers are just another property,
> but we usually consider identifiers to be a rather special property, and
> give them special treatment in most models, as well as in most of the encodings.
>
> After DC-7 I wrote up a short discussion [1] re-analysing DCMES to find out
> for which of the elements the "range" was actually another resource, possibly
> in-scope for a DCMES description. For these elements we would expect the
> *value* to be /precisely/ an *identifier*. These identifiers can take various
> forms, of course, and "name" is certainly one.
>
> I was also careful when drafting the BOX, POINT and PERIOD specs [2] to be
> consistent to this notion by refering to the values encoded in these ways
> as "identifiers" of a place or time-interval, as appropriate. (Sneaky?)
>
>
> [1] http://www.ned.dem.csiro.au/research/visualisation/metadata/dc-id/
> this was mentioned on the dc-datamodel list
> http://www.mailbase.ac.uk/lists/dc-datamodel/2000-01/0000.html
> [2] http://purl.org/dc/documents/recommendations.htm
>
> --
> [log in to unmask] CSIRO Exploration & Mining
> T:+61(8) 9284 8443 F:+61(8) 9389 1906 M:0403 302 672
> http://www.ned.dem.csiro.au/research/visualisation/
----------------------
Lorcan Dempsey
[log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|