[OAI-implementers] Re: Reconsidering mandatory DC in OAI-PMH

[Andy Powell]

On Sun, 10 Aug 2003, Gordon Paynter wrote:

> I've come late to this discussion, but I want to record my strong
> preference for retaining the oai_dc requirement.
>
> I've found the arguments made in favour of retaining the oai_dc very
> convincing.  Most objections seem to misunderstand the reason oai_dc is
> there, IMHO, which is to provide basic interoperability.
>
> Many people have pointed out that they use OAI-PMH internally (as do I)
> with their own schemas (as do I) and have no use for oai_dc (nor do I); or
> that oai_dc does not accurately represent their resources; or that it
> misrepresents their work.  I think these are reasonable objections, but
> not worth sacrificing the inherent interoperability that oai_dc gives us
> for.
>
> In cases like these, I'd suggest returning empty or near-empty records to
> oai_dc requersts.  The fields are optional, after all. If a client
> requires more complex information, they can request a different metadata
> format.

This message tends to sum up my views as well - though, at this stage, I
would probably indicate a 'preference' rather than a 'strong preference'
in favour of keeping oai_dc.

There is no doubt in my mind that mandating oai_dc has been hugely
beneficial to the uptake and success of the OAI-PMH.  My gut feeling is
that this will continue to be true in the future as well (though I must
confess that I am much less certain about this).

As Gordon says, in cases where oai_dc is inappropriate, one can always
return a minimal oai_dc record containing, say, only dc:identifier.

Andy
--
Distributed Systems, UKOLN, University of Bath, Bath, BA2 7AY, UK
http://www.ukoln.ac.uk/ukoln/staff/a.powell       +44 1225 383933
Resource Discovery Network http://www.rdn.ac.uk/