Starting from Elaine Svenonius' set-theoretic interpretation of bibliographical records, let’s redefine the Class 1 FRBR entities in the following way:
- An item is a unique physical embodiment of a work (i.e., a singleton set).
- A work is a set of items with the same intellectual content.
- An expression is a set of items with the same realization of intellectual content.
- A manifestation is a set of items with the same production history.
This set of definitions allows us to shift the burden of effort in creating a FRBRized catalog from defining specific entities of the four classes with explicit relationships from one level to the other (which, I believe is the assumed work process) to simple entry of values for facets associated with production, realization and intellectual property metadata on a per-item basis.
The difference in the level of effort by catalogers would be profound. Instead of having to hand-craft entities at the work, expression and manifestation levels as well as that of the item, they emerge bottom-up out of the item data, and can be driven by researcher usage rather than forcing catalogers to take on this additional burden. Since the practice of cataloging is traditionally bottom-up from the item in hand, it would ease the adoption of FRBR by reducing the cost of adoption. Evidence that this would be the case comes from the similar reduction in the level of effort seen by contrasting creating and maintaining monolithic taxonomies versus a handful of flat and/or hierarchical facets, a point made frequently by Joseph Busch and Ron Daniel of Taxonomy Strategies in their discussions of the benefits of faceted classification in information access and content management.
It also avoids the hit-or-miss nature of automatic clustering manifested in the work-set algorithms like that of OCLC and Karen Coyle’s proposal for the Open Library. Like their counterparts in search results clustering in systems such as Vivisimo, clustering algorithms tend to be, using an analogy I heard first from Alan Kay, like puppies; a lot of the time they’re fun to have around but sometimes they make a big mess. Having people craft their own faceted descriptions at the work, expression and manifestation levels is a more reliable manner to provide useful groupings fitted to a given resource discovery task.
This approach preserves the FRBR model, but moves the focus from entity to attribute. Given the work done in creating preliminary RDF vocabularies for FRBR (and their upcoming successors from the DCAM/RDA/RDF crew), we have the ability to build this type of system today.
OK, someone else must have thought of this. Anyone out there know who?