Turns out I completely misunderstood large parts of the Open Repositories talk on pf-dspace. With his permission, I quote from James Rutherford’s email correcting me:
In your comment, you mention that all nodes have to be DSpace, but this isn’t true; one of the main purposes of pf-dspace is to be able to allow nodes to harvest from heterogeneous sources (anything that can provide an OAI-PMH data provider interface).
The “pure DSpace” aspect wasn’t intended to exclude Fedora et al, more to highlight that no additional software is *required* (eg registries). Thus, if someone builds support for this kind of federation into Fedora, or Eprints, etc, then there’s no reason why they couldn’t actively participate in the federation. Of course, even if they don’t actively support it, they can still passively participate by allowing their content to be harvested.
I’ll strike out the mistaken bits of the earlier post, and add a link to this one. Mea culpa, and I’m now extremely interested in the project!
