Systems can associate data from different parts of a given source in
different ways. Differentiated authorities should always contain more
identifying data than just the name, and that data can be joined to
the name in an indexed string, or presented in cluster, or presented
as facets derived from a given set of authority records, or shown in a
spreadsheet with the users choice of column categories, or ... There
are lots of ways systems could make it relatively easy to choose among
authorities with identical names in the 1XX. I'll agree our systems
aren't there yet, but compared with what we could have to assist our
choices among authority records, the long list of meagerly
differentiated names we usually work with now hardly seems like a
paragon of user friendly design.
Not differentiating persons means that the user has to do the work. It
may be more convenient for the cataloger to toss another bib access
point on an undifferentiated pile, but that's less convenient for the
user trying to find books by a given author. The advantage of our
doing the work of associating a title with a particular author rather
than an undifferentiated heading is that there's lots of downstream
benefit from that work, and that each act of differentiation further
refines the information in the system specifying that author. When
each user has to do the sorting for him or herself, no one else
As for automatic controlling of headings, I agree it's a generally
unwise thing to do for unqualified headings; but that's the case now.
I've spent lots of time unmatching bib headings in our Aleph system
from the unqualified heading that the system matched them to and
adding the right qualifier to enable the correct match. I'd prefer to
have the system distinguish between those matches which are either
cataloger-attested or which include sufficient qualifying heading
information to satisfy some confidence threshold for automated
matching, like OCLC controlled headings, or simply report the headings
as unauthorized. Having a name heading linked to an authority record
which represents just a name and nobody in particular is just
pointless, and adds to the problem we already have with common names
rather than ameliorating it.
On Fri, Sep 2, 2011 at 4:27 PM, Robert Maxwell <[log in to unmask]> wrote:
> This sounds ingenious and good in theory, but I do wonder about it in practice-with a very common name there could be dozens (hundreds?) of these records to plough through, to say nothing of all the records for the same name that have dates or other qualifiers, that the cataloger will have to check before deciding either that the person he/she needs for the current item in hand is already represented on one of the undifferentiated (or differentiated) records; or that the person is not, and needs to have a new undifferentiated authority record created.
> Perhaps the benefits of the proposed system outweigh this, but it is a fairly easy proposition to eyeball the 670s in an undifferentiated record (even if there are a lot); it will not be "fairly easy" if we have to go through dozens or more, record by record, deciding which one should link to the bibliographic record we're working on.
> I'm also imagining how the linking process might work-after we've determined that one of the many undifferentiated records is the right one, will we again be presented with a list of (presumably unsorted) records to choose from for linking (for example, in OCLC's controlled heading window)? In any case, this would of necessity be a manual linking process, given multiple records with the same 1XX field.
> And the question of how local systems will handle this is not an insignificant question. In our system, for example (Sirsi/Dynix), the authority records link to matching access points automatically; if the database contains more than one authority record with the same 1XX field, there's no way to manually tell the system which one to link the bib record to. And since the 1XX's would all be identical I can't imagine that this linking wouldn't be manual, bib record by bib record.
> So either we do manual processing up front (in the proposed system) or we do manual processing later (in the current system) in the event that names get removed from the undifferentiated record. Since it's only sometimes that names get removed from undifferentiated records in the current system, but manual linking would be needed in every case in the proposed system, it seems to me that the current system might in fact be more efficient.
> Robert L. Maxwell
> Special Collections and Ancient Languages Catalog Librarian
> Genre/Form Authorities Librarian
> 6728 Harold B. Lee Library
> Brigham Young University
> Provo, UT 84602
> "We should set an example for all the world, rather than confine ourselves to the course which has been heretofore pursued"--Eliza R. Snow, 1842.
> -----Original Message-----
> From: Program for Cooperative Cataloging [mailto:[log in to unmask]] On Behalf Of Adam L. Schiff
> Sent: Friday, September 02, 2011 2:50 PM
> To: [log in to unmask]
> Subject: Undifferentiated personal names
> I personally don't see why we need to wait for RDA to implement a new
> policy regarding undifferentiated personal names. Simply change the
> policy to create a separate NAR for each different person. If the access
> point on two personal name NARs is identical, code the 008/32 on both "b".
> NACO normalization rules would need to be changed to allow two NARs to
> have the same 100 field.
> One of the beauties of the suggestion to establish separate name
> authorities for persons with the same access point is that the separate
> authorities could then be linked to the proper bibliographic records (in
> Connexion, the control heading function could be used) and when the access
> point is changed to differentiate it from other NARs, the right headings
> in bib records would automatically get flipped in systems that used the
> link. How this would work in many local ILSs is another question, but
> would probably be no less a problem than what we currently have when
> headings are pulled off of undifferentiated authority records.
> Why wait to make this change in another 15 months or longer?
> Adam L. Schiff
> Principal Cataloger
> University of Washington Libraries
> Box 352900
> Seattle, WA 98195-2900
> (206) 543-8409
> (206) 685-8782 fax
> [log in to unmask]
Stephen Hearn, Metadata Strategist
Technical Services, University Libraries
University of Minnesota
160 Wilson Library
309 19th Avenue South
Minneapolis, MN 55455