Brenndorfer, Thomas
Fri, 03 Sep 2010 07:20:52 -0700
> Also, analysis of the FRBR user tasks often stops after the > find/identify/select/obtain part, which really is almost totally > speculative since those are the things people do completely on their own, > and what they *really and genuinely* do is extremely difficult to know. In > any case, what should be of primary concern for catalogers right now are > the rest of the tasks, since that is what we are proposing to build and > spend our resources on, i.e. creating the > "works/expressions/manifestations/items finding them by their > authors/titles/subjects". > I think people want to find/identify/select/obtain on additional attributes and relationships. But to get started on any project, one has to define a scope http://www.rda-jsc.org/docs/5rda-scoperev4.pdf ... and in this case the scope is largely based upon existing data and conventions, which makes learning RDA hardly a backbreaking exercise for experienced catalogers. One can argue that FRBR and RDA have picked the wrong entities, the wrong attributes, and the wrong relationships. Perhaps one can argue even specifying these categories is wrong. But one can't built a house with Google searches. One builds a house by finding, identifying, selecting and obtaining the material and skilled tradespeople (i.e., the resources) to build the house. These are tasks that need to be done, and they are not acts akin to keyword searching continuously for needles in a haystack. To think otherwise would prove Nicholas Carr's thesis in "Is Google Making us Stupid?" The Panizzi-Carlyle disagreement is a case in point. http://pi.library.yorku.ca/dspace/bitstream/handle/10315/1250/denton-frbr-and-the-history-of-cataloging.pdf?sequence=1 Carlyle saw no entities beyond the physical book. A book need not be related to any other entity, and it could be made findable by no other tool than a title list. Today, we might say, nothing but a Google search on the title and the full text. Panizzi saw that this was not appropriate for people's actual needs. The content in books (and now in so many other resources) needs to be addressed in ways that meet people's actual needs. People want and have a right to see the resources of a library organized by relationships and around entities that they themselves think about-- authors, versions, topics, genres, formats. If anything, the argument has been to open up the silo of this metadata in libraries so it can connect better with comparable metadata elsewhere to accomplish comparable tasks (i.e. not just reduced to full text Google searches). RDA takes the existing conventions of AACR2 and incorporates a more consistent data model. If we stick with AACR2 then we're boxed in with a data model that is difficult to extend and shot through with internal inconsistencies around content/carrier and collocation problems, as well as card catalog-centric language. If we move to an entirely new data model independent of the past and built upon vague ideas of what users want and overhyped promises of specific bits of technology, then we have the headaches of backwards compatibility and a likely fracturing of our data into more silos, not fewer. Thomas Brenndorfer Guelph Public Library