The statement on p. 7 that RDF specifications indicate that identifying more than one domain for an element is to be interpreted using a Boolean AND (i.e. that all instances of the element must be members of both domains) seems to me to hint that the FRBR tables are logically incompatible with RDF (and logically incorrect). The element analysis currently lists only a few elements with no domain in column D, but, following FRBR nearly all of the other elements are identified with manifestations, not with expressions or works. Thus, according to this table, expressions and works do not have titles, only manifestations do. Expressions are not identified by edition statements such as 2nd ed. rev., only manifestations are. Expressions are not identified by statements of subsidiary authorship (or, in RDA terms, statements of responsibility pertaining to contributors), only manifestations are. In my opinion, many more elements should lack a domain in column D.
Im concerned that RDA will analyze the cataloging record into elements that cannot be reassembled by systems people into a coherent display that is understandable to catalog users. We have taken the watch apart and strewn the pieces over the table. Can systems people be trusted to learn the complexities of the bibliographic universe well enough to reassemble all of these elements properly, linking them effectively and demonstrating the complex relationships? Past experience would seem to indicate that perhaps they cannot be trusted to do so. In my opinion, the cataloging rules should be designing the displays, not the system designers. So far, it appears to me that we have extracted and thrown out all of the babies, while carefully conserving the bathwater. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Martha M. Yee Cataloging Supervisor UCLA Film and Television Archive 1015 N. Cahuenga Blvd. Los Angeles, CA 90038-2616 323-462-4921 x27 323-469-9055 (fax) [EMAIL PROTECTED] !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!