David,

What I'm reading into your request is a way to create a MODS record for
the FRBR'd work -- that is, a record with one work and many
manifestations, or a way to "cluster" records for the manifestations.
 From what I understand of what OCLC did and how VTLS does its
FRBR-izing (now THAT's about the ugliest verb in our language!), there
is no "FRBR" record, but there is a FRBR database design. I speculated
on creating a FRBR record in my article in Library Hi Tech
(http://www.kcoyle.net/functional.pdf), but so far FRBR has been an
action taken on standard MARC records without changing those records.
There are logical reasons for having the work/manifestation connection
take place someone other than within the record itself, mainly having to
do with the relative ease of exchanging manifestation-level records as
compared to a composite record with multiple manifestations.

As Barbara Tillett said, some thought has gone into the role of the
authority record as the work-level link that could be used to bring
together records for manifestations, and this approach looks quite
promising. That requires that the link to the authority record be
included in those manifestation records, and at the moment the only
"link" between MARC bibliographic and authority records is that they
contain matching text in the key fields. Some database designs create an
internal link between the headings and the authority records, but the
MARC record design doesn't include that. You could do this in MARCXML or
MODS with an XML link, however. If you experiment in this direction,
please send us your progress reports!

kc

David Mimno wrote:

>Hello,
>
>I am a programmer with the Perseus Digital Library at Tufts University. We are
>in the process of rebuilding our site's catalog based on MODS records. We are
>particularly interested in FRBR because the nature of our materials is already
>extremely hierarchical. In particular, we deal with multiple editions and
>translations of works (like "The Iliad"), as well as commentaries, lexicons and
>other secondary materials that relate to works. The collection is not very
>large, and it is reasonably well organized: we already know which records are
>expressions of which works, and we even have standard, authorized IDs for each
>work (based on the Thesaurus Linguae Grecae and related projects).
>
>What I have not been able to find in the literature is how to represent the
>relation between an expression and a work in MODS. Would there be a record for
>the work "The Iliad" as well as records for a Greek edition and a particular
>English translation? Would the record for an English translation have two
><identifier> elements, one for the expression's ID and one for the work's ID? Is
>there a standard way of declaring which is which? Would a given work be more
>like an authorized uniform title, with a MADS record?
>
>Thank you, and sorry if I have missed anything obvious!
>
>-David Mimno, Perseus Project
>
>
>

--
-----------------------------------
Karen Coyle / Digital Library Consultant
[log in to unmask] http://www.kcoyle.net
ph.: 510-540-7596
fx.: 510-848-3913
mo.: 510-435-8234
------------------------------------