Since I was the metadata "person" on the OpenURL committee, I guess I
should speak to that. The OpenURL in version 1.0 is a profile-based
standard that can be expanded by anyone wishing to register their
profile and metadata elements in the OpenURL registry. So the metadata
elements and the "genres" included in them are there really to
demonstrate the registry, not to be definitive in any way. As soon as we
put them out there we started getting requests for new ones, of course.
OpenURL has this concept of a "community" which I think is a useful way
of thinking about metadata. The community of film archivists is going to
have different metadata needs from the community of book publishers, and
very different from the community of chemists or auto parts dealers.
Rather than trying to accommodate them all with a single metadata
format, OpenURL allows each to create one or more profiles and to
register whatever metadata formats, identifiers, etc. that they need.
The community that has so far been actively using the OpenURL is being
called the "scholarly community" -- and it is mainly being used to
deliver services (full text, ILL, etc.) relating to citations in
abstracting and indexing databases. It is not intended to be metadata
that definitively describes those citations, it is a way to transport
enough metadata to either locate the item online or in a library's
This is kind of what I mean by needing to understand the purpose behind
the metadata and the genres in order to know where your list should end.
The OpenURL profile for the scholarly community would not include
personal communications because 1) those aren't included in the
abstracting databases 2) there are no bibliographic services to
provide. In theory the metadata could be expanded to include it, but it
would mean that someone has a web service they wish to provide. OpenURL
is not a metadata format for the database storage of citation metadata;
OpenURL metadata is a transient stage in a web service.
If we want the genre element in MODS to include any genre from any
metadata, then it should be a text field, not a controlled list. A
controlled list only makes sense if you have some reason to keep
control, i.e. there's some particular meaning to the list that you don't
want to see violated.
On Thu, 2003-12-04 at 16:22, Bruce D'Arcus wrote:
> On Dec 4, 2003, at 5:56 PM, Rebecca S. Guenther wrote:
> > As I suggested before, we could establish a list of values for
> > citations
> > to be used in MODS <genre> if Bruce wants to provide them.
> > Alternatively
> > we might consider the genres listed in the proposed OpenURL standard
> > ...
> If I look at the OpenURL genre list, it does add a few things that I'm
> looking for. It leaves out some things that are common in citation
> formats, though, like personal communications (which could cover
> letters, memos, emails, etc.), thesis and hearings.
> One possibility I think I've discussed with Rebecca is coming up with a
> list that covers all the genres currently covered in existing
> bibliographic formats like Endnote, Reference Manager, BibTeX, etc.
> That would in essence be a superset of the OpenURL list. I'm happy to
> do that, but will probably need some help.
> This still leaves the question of whether some of these genre terms
> ought to make it into the MARC list.
> I guess this gets back to Karen's questions. How would you propose to
> deal with this Karen?
Digital Library Specialist
Ph: 510-540-7596 Fax: 510-848-3913