[OAI-implementers] OAI Identifiers

Simeon Warner simeon@cs.cornell.edu
Fri, 22 Feb 2002 09:17:20 -0500 (EST)

On Fri, 22 Feb 2002, Alan Kent wrote:
> > Note that the
> > schema says:
> >     <!-- * A "repositoryIdentifier" that is a unique identifier for a -->
> >     <!-- repository.  Its uniqueness will be tested as part of a -->
> >     <!-- registration process.  The repositoryIdentifier is case -->
> >     <!-- sensitive and may contain the characters [a-z]|[A-Z]|[0-9] -->
> >     <!-- e.g. arXiv, VTETD. -->
> If you allowed '.' in the name then sites (and maybe '-'?), then you could
> at least use domain names. oai:mds.rmit.edu.au:1234. It would not be
> mandated, but people would be reasonably confident of having a unique
> name without collision. They can always tack something on the start
> if they have to repositories on one site. But I am really suggesting
> this as something a site could choose to do if it does not want to
> go to the global registry.
> Digitial Libraries might want to keep OAI in-house. The idea of central
> registries just seems a pain if the protocol ever took off. Reusing
> domain names registry processes would be an easy 'convention' to 
> follow without having to go to the extreme of full URLs.

I don't see harm in allowing also '-' and '.'. (I wouldn't want to make it
case insensitive.) However, without some enforceable policy about naming
(avoiding the need for OAI registration which currently solves the
uniqueness problem) does this really buy us anything? After all, I could
use identifiers "http://arXiv.org/abs/hep-th/9901001" and such for arXiv
but I choose to use the simpler oai scheme "oai:arXiv:hep-th/9901001".
I don't see that there is any collision problem for in-house use. Instead
of a global registry you have one local-one (likely a person coordinating
the effort).

> > Note also that OAI-PMH does not mandate this scheme and not all registered
> > repositories have chosen to use it.
> Yes, I noticed 'mit.ethese', 'UKOLN-ejournals', and 'NSDL-DEV-CU'.
> Should I then fail to harvest these sites? :-)

You could _choose_ to harvest only sites that use oai identifiers but it
was never the intention to force this scheme on everyone. These sites can
still be 'OAI compliant'.