On 27/02/2012 05:02, Kelley McGrath wrote:
<snip>This is a very interesting project, and demonstrates once again, that if the emphasis is focused on the users, who are assumed to need to do the user tasks as defined in FRBR (i.e. the functional requirements), then those functions can be achieved through modern indexing tools instead of requiring new cataloging rules and new structures. These tools are found right now in Worldcat, Koha, the Extensible Catalog, and this one that you discuss, too. Probably others, too. Each of those tools can be improved, and probably relatively simply. Such a wonderful development should be seen as good and positive, especially during these difficult economic times.
There has been some discussion about the relationship between the FRBR entities (especially group 1) and end-user display or underlying data structure.
I think OLAC's FRBR-based prototype for moving image materials (http://blazing-sunset-24.heroku.com/) is a good example of an interface where the underlying logic is based on FRBR, but the user display is not centered on the FRBR group 1 entities. We only display two levels: the work+primary/original expression and the expression-in-hand+manifestation+item-location. A brief overview of the prototype and its aims can be found in the first section of http://pages.uoregon.edu/kelleym/publications/FRBR&Facets_C4L2012.pdf. We also don't limit the user to top-down access to the FRBR group 1 entities. Many systems force users to start with the Work and move down, but through facets, we allow users equal flexibility to start from the bottom up with item location or manifestation format (e.g., DVD, Blu-ray). This is described in our JCDL short paper at http://pages.uoregon.edu/kelleym/publications/JCDL_OLAC_FRBR_prototype.pdf.
The data in the prototype came from a RDMS. However, the tables weren't strictly based on the FRBR entities (although they could have been).
I think this shows that you can build something that is conceptually based on FRBR where neither the display nor the underlying data structure maps 1:1 to the FRBR entities.
We should push the computer systems to their utmost, instead of pushing the catalogers to do even more work that is simultaneously more complex. After all, that is one of the primary reasons for introducing "technological innovations".
But, if the purpose of FRBR now is to enter the "Linked Data" universe, (which had not yet been foreseen when FRBR was first published) then that does indeed go beyond the functional requirements, and certainly goes far beyond matters of simple metadata. There are many, many, many ways of entering the Linked Data universe.