[OAI-implementers] ListMetadataFormats problem

[Steven Bird]

> I'm working on an OAI repository for a group that wants to define it's own
> metadata format. The trick is, this format will be a combination of Dublin
> Core elements and some new elements yet to be defined.

This is just what the Open Language Archives Community (OLAC) wants to do,
but it hasn't been clear to us how best to do it with schemas.  As a
starting point we've made a copy of the DC format with our own namespace
(http://www.language-archives.org/olac-0.1.xsd), and we're planning to
add qualifications in the style of DC Qualifiers.

It will be trivial for our data providers to generate DC from this richer set.
It mostly consists in throwing away the refinements and qualifications,
and mapping from the controlled vocabularies (such as language codes) to
human-readable versions, a mapping which is automatic.

What do people think of this approach?  Keeping things simple for the data
providers seems more important than trying to track what the XML world
currently thinks is the neatest way to do something.  However, if this
multi-namespace approach can work in the OAI framework then we might also
use it.

-Steven