Yes, this one really bothers me. A resource about physicians from China and also about veterinarians from Vietnam would get these LCSH headings:
How likely is such a book to exist? It seems an unlikely combination of topics to me, but I guess all things possible will exist eventually. It is definitely a trade off between what is gained and what is lost, however. I actually think a properly reconstructed LCSH can give us the benefits of facetting _and_ still the granularity of pre- coordination. It would take longer to provide examples of how this could be done---and even trickier to figure out the interfaces to support both uses, but I think it's possible. Despite the initial publicity of FAST being about making subject assignment lower skilled or less costly, I think that's a wrong direction which unneccesarily scares people about FAST. The true promise of FAST to me is in _increased access_, making certain types of very useful interfaces possible. On the other hand, if we abandoned pre-coordinated LCSH for FAST, we'd also lose other things. It's possible what we'd gain would be more that what we'd lose---but again, I think it's actually possible to have our cake and eat it too here. More work needs to be done. One of the things holding us back here is in fact that I believe OCLC has in fact done a lot more very interesting work on FAST, on how to use it, on how to automatically generate it from LCSH, in ways that have a lot to contribute to the work to figure out how to 'have our cake and eat it too'---but is holding many of these things back as proprietary information they intend to make into a proprietary product at some point. This is harmful to the community that funds OCLC as a cooperative, holding us back. Research is for sharing, so others can use it and build upon it---this is a core library value. It's a serious pity that not more useful research in these directions comes from our academies, and that the one community institution we have that DOES engage in this very necessary research holds so much of it back as proprietary findings. Jonathan
Physicians--China Veterinarians--Vietnam In FAST, you get: Physicians Veterinarians China Vietnam Can a user figure out the context from the FAST headings? Maybe if the title is explicit, or a summary note is present, or there is a table of contents. Keyword searching will work just as well on both the LCSH strings and the FAST terms, so to exactly what benefit is it to lose the inherent semantic meaning of the precoordinated string other than perhaps to save some cataloging time or enable a less highly trained level of staff assign the terms? Now admitted, as Martha wrote, the precoordinated strings are not perfect, since Physicians--China can mean physicians IN China and/or physicians from China (i.e. Chinese physicians). I agree with Martha that it would be better to develop a way to explicitly code the relationship, e.g. something like: 650 _0 $a Physicians $i in $z China. --Adam Schiff ************************************** * Adam L. Schiff * * Principal Cataloger * * University of Washington Libraries * * Box 352900 * * Seattle, WA 98195-2900 * * (206) 543-8409 * * (206) 685-8782 fax * * [EMAIL PROTECTED] * **************************************
