[OAI-implementers] on the OAI-PMH data model

[Caroline Arms]

Fabio,

LC, in providing access to digital reproductions of content in its
collections does find the 3-level distinction useful -- even if the
OAI terminology is not intuitive to my colleagues.

In LC's lcoa1 data provider, most of the "items" are essentially master
descriptive records in the MARC format.  The OAI "records" are generated
on the fly from those master records.  Recently, we changed the MARC to DC
transformation.  We changed the datestamp on all the virtual oai_dc
records when we changed the transformation, but the marc21 and mods
records have earlier datestamps.  The marc21 datestamps will correspond to
the last time the underlying MARC record was changed in the OAI
repository.  A mods record timestamp may correspond to the last update to
the underlying "item" or the date we last changed the MARC to MODS
transformation.  We maintain datestamps for OAI records even though the
records do not actually exist.

These 'items' (master descriptive records) are maintained independently of
the resources they describe.  In the oai_dc records, the significant
attributes (dc:date, dc:creator, etc.) usually relate to a physical
original which hasn't changed since it was created (usually before 1923
for copyright reasons and maybe many centuries before that) and the fact
that it was digitized by the Library of Congress is of secondary
importance for discovery.

That may provide an example of why the distinction is made for OAI.  
It certainly doesn't mean that every implementation needs to make use of
the distinction.  If the master metadata for your OAI implmentation is
embedded into the resources, for example, it would make little sense to
distinguish item from resource.

   Caroline Arms                                  caar at loc.gov
   Library of Congress

    


On Thu, 17 Mar 2005, Fabio Simeoni wrote:

> > I think the idea with this distinction is that there may be multiple items
> that disseminate different assertions (but in the same metadataFormat) about
> the same resource.   
> >  Example: Bob's item about the resource  <http://w3.org/> http://w3.org/
> says 
> >  the dc:title is "The W3C", and Cindy's items says the dc:title is "The
> homepage of the W3C".
>  
> that makes sense, of course. thanks. is there anyone 'using' the distinction
> in that sense? since we're at it, how common is it for datestamps to be
> actually attached to records, as per specs?  iow, who's keeping persistent
> records in different formats and allow the latter to change independently
> from items and each other? Aggregators? What class of changes are record-
> but not item-dependent, btw? 
>  
> thanks,
>  
> f 
> 
> ##############################################
> Fabio Simeoni
> Research Fellow
> Department of Computer & Information Sciences
> University of Strathclyde, Glasgow
> 
> TEL: +44 141 548 (3590)
> FAX: +44 141 548 (4523)
> 
> 
> _______________________________________________
> OAI-implementers mailing list
> List information, archives, preferences and to unsubscribe:
>  <http://www.openarchives.org/mailman/listinfo/oai-implementers>
> http://www.openarchives.org/mailman/listinfo/oai-implementers
> 
> 
> 
>