“--the proper solution is to mark the subdivision usage itself as invalid”
Yes, I agree, if the commonly used subject validation program could detect it, for instance, if “600” present, then “$v Commentaries” is invalid. My point is, as thousands of new bib records enter OCLC daily and then pass onto thousands of individual institutions, would it be possible for OCLC/FAST to catch and mark them as questionable or invalid (thus no ‡2 fast ‡0 (OCoLC)fst01423723)? One would think it would be both efficient and economical to do so. Otherwise, the effort would be thousand fold more expensive.
I think one thing that contributes this problem is that people assume a subject heading string is valid if it is controllable in OCLC. The subdivision authority record for Commentaries includes a 073 field relating it to SHM 1188 (Sacred works), but unfortunately the authority records for sacred works don’t have coding to match up with this. OCLC can’t tell what is a heading for a sacred work, so it allows the subdivision to be applied more broadly than is appropriate.
Christopher Thomas, M.L.S.| Electronic Resources and Metadata Librarian
Law Library · University of California · Irvine
The automated FAST generation in OCLC relies on the subject headings being correct. This is not a special issue with the term Commentaries; it is a problem with every subject heading which is incorrect or invalid. As you point out, the system cannot tell whether a wide range of subdivisions are used correctly. There is no good reason to single out the FAST term Commentaries for special attention and work-arounds for correction. If there is a way to detect that a subdivision is used incorrectly, then the proper solution is not to have a work-around fix for the FAST terms--the proper solution is to mark the subdivision usage itself as invalid. If that cannot be done in a practical way, then there is also no practical way to detect it for the FAST generation.
I need to point this out because lately I have been seeing “655 #7 Commentaries. ‡2 fast ‡0 (OCoLC)fst01423723” generated for faulty subject headings everywhere, after names (600), name/titles (600), secular literary works (630), non-literary works (630). Or has there been a policy change on the use of “$v Commentaries” recently that I am unaware of?
Under SHM H1188 [Sacred works], there is an A-Z list of subdivisions that can be assigned. This is the only place we find the form subdivision $v Commentaries. That is to say, “$v Commentaries” can only be used for sacred works entered under the title in the bib (630 0 [Title]). For secular literary works (whether entered under title or author), there is a different list (H1155.8). Neither “$v Commentaries” nor “$x Criticism and interpretation” can be used for a work that is a commentary on such works.
Now, It’s understandable that the machine cannot tell if “630 00 Anacreontea ‡v Commentaries” is valid or not. But what about “600 00 Aristotle $v Commentaries” or “Aristotle. $t Poetics. $v Commentaries”? Would it be too difficult for OCLC to catch this type of mistakes and not to generate “655 #7 Commentaries. ‡2 fast ‡0 (OCoLC)fst01423723” automatically? Because, by definition, this form subdivision applies only to sacred works.
Just an observation.