I guess I like that kind of work, checking all the little perturbations
of authorities to see which headings match which bibs. Probably few
libraries can expect to have it perfect all the time, and maybe we even
need to depend on our patrons to some extent to point out errors.
The drawback to your proposal, as some suggested, is that there would be
a proliferation of $c qualifiers that might not really be unique or be
questionable in other ways. But I'll try to keep my mind open on this.
I've never liked LTI-style authority updates. I guess they go against my
tendency to want to check and correct everything myself. But I suppose
libraries just can't afford that much of the time.
From: Program for Cooperative Cataloging [mailto:[log in to unmask]] On
Behalf Of Lasater, Mary C
Sent: Monday, November 06, 2006 4:47 PM
To: [log in to unmask]
Subject: Re: [PCCLIST] undifferentiated name records
The current situation is that an 'entity' with a generic name can
be established on a non-unique or even uniquely and changed to
non-unique... then moved off the non-unique and eventually all
names on that non-unique can be moved off... except one, which is
then unique. THEN someone can find a date or middle name and add
that info. IF the original library hadn't kept up with all the
changes... their bib, linked to the first 'version' can end up with
a totally unrelated middle name or date. So yes, it is the way we
'do' non-uniques that is the problem for a linked authority system.
Each month I spend hours looking at a report that shows each of the
non-uniques that we have used in our catalog and that have been
changed that month. I see the pair that change and I check to see
if we have the title. If so, I then track down the new heading so I
can keep our headings current. Marcive supplies the updated
authority records and the report showing the changes. We do not
reload all our bibs or want/receive updated bibliographic records.
Much time is spent trying to figure out the new heading. For
example the one I'm looking at now is Spencer, Christopher... there
are four with dates in the authority file and I might have to look
at each one. A little while ago I looked at Schmidt, Sarah... the
version I needed was now Schmidt, Sarah L (Sara Leff), 1934- It is
a real puzzle trying to find changes... and sometimes the 'pair'
has been lost and is no longer on an authority record at all. My
proposal is that we eliminate all this searching and allow linked
authority records to work. The current situation is not efficient.
Hope this makes some sense,
--On Monday, November 06, 2006 1:04 PM -0600 Ted P Gemberling
<[log in to unmask]> wrote:
> Mary Charles,
> Are the undifferentiated name headings really at fault for the
> kind of "false flips" you're referring to here? Could you give an
> example of: "With a linked authority system those changes can be
> really bad with people
> writing books 100s of years before they were born"? My impression
> is that false flips by, say, LTI are more likely a confusion
> between two unique headings than something caused by
> undifferentiated headings. For example, I just looked up "Adams,
> J. T.," who I know as a male singer, and found that if you don't
> add the qualifier "vocalist," LTI would change that to Adams,
> Jeannette T. If "Adams, J. T." were an undifferentiated heading,
> no change could be made.
> My understanding of undifferentiated authority records is that
> they are basically files for recording information about
> occurrences of very generic names, like John Smith. As more is
> learned about particular authors, they are removed from them and
> put in their own unique authorities. So in a way, no date of
> authorship is implied by the use of an undifferentiated heading,
> because no date for the person(s) is known. --Ted Gemberling,
> UAB Lister Hill Library, Birmingham, Ala.
> -----Original Message-----
> From: Program for Cooperative Cataloging [mailto:[log in to unmask]]
> On Behalf Of Lasater, Mary C
> Sent: Saturday, November 04, 2006 10:33 AM
> To: [log in to unmask]
> Subject: Re: [PCCLIST] undifferentiated name records
> You have touched on a topic/problem that I hope we can 'do
> better' under RDA. I would like to see us move toward using
> those phrases that we construct as $c's with the author's name
> and setting these authority records up that way. THEN when we
> find out more about the author, we can change the 'distinct' AR
> instead of the 'non-unique AR if necessary. Several years ago I
> mentioned in a talk at ALA that I spend too much time looking
> for how these have changed and would prefer not to even have the
> non-unique AR. With a linked authority system those changes can
> be really bad with people writing books 100s of years before
> they were born. If instead of constructing non-unique's we
> created individual AR's with the phrases (that we already
> construct for the non-unique authority records) and then changed
> that AR when we have more info, linked authority system changes
> would automatically change the 'correct' authority record, only.
> Much/all of the time spent looking for the changed heading that
> is no longer on the non-unique (Is this the Tom Smith born in
> 1952, or 53, or is it Tom T. Smith or Tom Smith, Ph.D.) would be
> Music catalogers already get to add these phrases and we see this
> type of 'qualification' on various web tools. What are the
> disadvantages? Do they outweigh the benefits?
> Mary Charles Lasater
> --On Friday, November 03, 2006 2:21 PM -0800 "Paul J. Weiss"
> <[log in to unmask]> wrote:
>> I note that the practice of bracketing data in one 670 per person
>> in an undifferentiated name record is not actually given as
>> policy anywhere. The MARC authority format give it as one
>> possibility ("subfield $a may contain a descriptive term for an
>> author enclosed within brackets "). DCM Z1 touches on it in the
>> introduction and at 670. The NACO Participants Manual describes
>> the practice, but our NACO reviewer at LC continues to remind me
>> that the PM does not set policy.
>> Do any of you _not_ follow that practice? If not, what was your
>> thinking behind your decision? Have any of you considered some
>> other practice?
>> UCSD NACO Coordinator
>> Paul J. Weiss
>> Catalog Librarian and NACO Coordinator
>> Metadata Services Department
>> UCSD Libraries
>> [log in to unmask] _______________________________________
> Mary Charles Lasater
> Vanderbilt University
> Email: [log in to unmask]
Mary Charles Lasater
Email: [log in to unmask]