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

[Naomi Dushay]

I don't have any problem with Jose's suggestions, but:

1.  Jose, do you know that the current OAI-PMH allows any number of XML
schema validating formats, *in addition* to oai_dc?  From your message,
I wasn't sure, and we've encountered many, many, MANY folks who think
that OAI-PMH only allows oai_dc.   While a crosswalk repository may be
useful, it sounded like you were hoping to use crosswalks to
interoperate, rather than using an additional, richer metadata format.

2.  Validating OAI-PMH served metadata to its format, structurally, is
exactly equivalent to XML schema validation of OAI-PMH responses.
OAI-PMH requires an XML schema for each format served.  The XML schema
is the way that computer recognizable structural information about the
metadata is stored.


As for my stance on the oai_dc requirement -- I'm wondering about the
impact on my current employment, the National Science Digital Library.
If NSDL sites have trouble providing oai_dc, and that requirement is
removed, then what?  Can the documents be written in a way that
encourages folks to use oai_dc, particularly if they don't have another
format in mind, or if they're somewhat baffled by OAI-PMH?  Many of our
sites fall into those categories, and losing the basic metadata
interoperability provided by oai_dc could have negative ramifications on
the NSDL.  On the other hand, perhaps the NSDL could provide more
documentation for its contributors.

Lastly, I do believe that the OAI protocol itself could well be "format
neutral", and that would free it up to be used in other ways --
primarly, to exchange any schema validating XML that can be sensibly
broken up into "records."

- Naomi Dushay