[OAI-implementers] Universal Identifiers

Young,Jeff jyoung@oclc.org
Fri, 9 Feb 2001 16:32:05 -0500


The answer to your question about unique identifiers for non-bibliographic
items isn't necessarily simple. 

If there is a simple answer, this is it. The Library of Congress has a MARC
name authority database for people and institutions, and each records has a
widely recognized unique identifier called the LCCN. The coverage, however,
is hardly exhaustive. For details about the format, consult
http://lcweb.loc.gov/marc/authority/.

Closer to home, the Networked Digital Library of Theses and Dissertations
(NDLTD) project (see http://www.ndltd.org/), may be of interest to you since
it's dealing with this very issue. NDLTD is working to create OAI
repositories around the world containing thesis and dissertation records
from participating universities and libraries. OCLC has been assigned the
task of developing a scheme for allowing links to be stored in the records
in place of personal names. (I'm oversimplifying a little here.) For
example, where you might otherwise see <dc:creator>Twain, Mark</dc:creator>,
I expect that you will instead see some sort of URL that will return a rich
set of XML data about that person. In an important sense, then, this URL is
a unique identifier for the person, and a useful one to boot.

I should note, however, that NDLTD has chosen to accomplish this feat by
defining their own metadata format based on the Resource Description
Framework (RDF). Since OAI hasn't chosen to use RDF to encode their oai_dc
format, such linking isn't appropriate from their default format. If you
plan to store any kind of link in your records, you too will be forced to
define your own format just as NDLTD will do.

To continue, and this is all still preliminary thinking, the URL I just
mentioned might very well resolve to an OAI GetRecord command to an OAI
repository containing name authority records. The implications of using the
OAI protocols are that the authority database could be distributed around
the world and freely shared. To show the potential, the LC authority
database I mentioned above could easily be mounted as an OAI repository and
the MARC content distributed in XML format for convenient use. One caveat
here, though is that there is currently no Dublin Core standard for name
authority information. The NDLTD application, however, depends on such a
scheme, so it will happen, even if it's ad hoc.

An even bigger caveat, however, is that this is just the tip of the
logistical iceberg for making such a scheme work. I've worked on authority
control-related projects for a number of years here at OCLC, and can attest
to the fact that it's a horribly complicated problem. The internet and tools
such as OAI, XML, Dublin Core, and perhaps even PURLs afford us an
opportunity to reexamine the way we can address this problem. I'm pretty
excited about the possibilities.

Jeff Young

---
Jeffrey A. Young
Senior Consulting Systems Analyst
Office of Research, Mail Code 710
OCLC Online Computer Library Center, Inc.
6565 Frantz Road
Dublin, OH   43017-3395
www.oclc.org

Voice:	614-764-4342
Fax:		614-764-2344
Email:	jyoung@oclc.org




> -----Original Message-----
> From: ePrints Support [mailto:support@eprints.org]
> Sent: Friday, February 09, 2001 3:26 PM
> To: 'OAI-implementers'
> Subject: [OAI-implementers] Universal Identifiers
> 
> 
> Is there a plan for registering archive identifiers yet? I'd like
> to mention it in the eprints.org documentation.
> 
> Does anyone know of a system for assigning unqique identifiers to 
> other cromulent entities such as people (creators), institutions
> and/or anything else you might want to uniquely identify in an
> archives metadata.
> 
> At least journals and books have ISSN and ISBN.
> 
> -- 
> 
> Christopher Gutteridge 
> ePrints Technical Support
> support@eprints.org
> 
> _______________________________________________
> OAI-implementers mailing list
> OAI-implementers@oaisrv.nsdl.cornell.edu
> http://oaisrv.nsdl.cornell.edu/mailman/listinfo/oai-implementers
>