Folks:

An important design feature for the next generation bibliographic framework 
will be "catalog code neutrality".  Discussons about "main entry, yes or 
no" should probably be left to the cataloging folks; the new framework 
should theoretically support any approach.

Attempting to restrict the new framework to a reduced element set will 
probably be a non-starter a) since Dublin Core already exists, and b) since 
each and every data element in MARC has been championed over time and 
declared vital by some group or other.  Rather than setting ourselves the 
task of winnowing up front, it would be wise to explore a data dictionary 
approach and a robust element relationship framing / grouping 
mechanism.   Within the new framework various interest groups should be 
free to define "profiles" (as has been the case, e.g., with METS, although 
ideally more persuasively) that represent different catalog codes, best 
practices, consortial agreements, etc.

BTW - I understand that W3C and others have recognized that the RDF triples 
approach in fact lacks two important parameters that will need to be 
defined before we go much further, namely namespace and provenance. So 
we'll  need "quintuples" instead of triples ;}  (I'd like to hear more 
about this from W3C.)  David Weinberger in his keynote at the recent DLF 
Forum expressed the view that the two most important innovations in our 
field recently are linked data and namespaces, the latter because, he 
believes, we can no longer impose single, broad-based schemas on the 
information world and must give in to the idea that we will increasingly 
and always have multiple, competing schemas and metadata standards.  The 
challenge then will be to build and maintain effective crosswalks.
___________________________________________
Stephen Paul Davis
Director, Libraries Digital Program
Columbia University Libraries
535 W. 114th Street, New York, NY  10027
email: [log in to unmask]  phone:  (212) 854-8584
___________________________________________