I'm going to cc this reply to a few places, in the hopes that someone
will have an answer.

On Dec 12, 2004, at 9:54 PM, James Howison wrote:

> Trevor has written a python script that seems to be able to insert XMP
> (which is RDF) into arbitrary PDF files.  That's great news and we
> should now be able to push ahead on getting metadata into the PDF
> files.
> You are the person I know who I'm hoping can explain where I can learn
> about these new fangled schema (I'm a bibtex dinosaur when it comes to
> data models I'm afraid).  Is there an easy way to represent MODS in
> RDF and then an easy way to convert MODS to bibtex (because that's
> still my rubber meets the processing tools format?

I'd be interested in an RDF representation that would map well to and
from MODS.  I've not seen one, however.  Does anyone have any ideas?

Perhaps PRISM?

It's not as rich as MODS, but sufficient for journal articles
certainly.  And it should be easy to write XSLT stylesheets to go back
and forth because the structure is broadly similar.

As for MODS to BibTeX, that's easy: