[DCRM-L] FW: Slides and Recording from Nov. 1 "When to Input a New Record" Webinar

Dooley,Jackie dooleyj at oclc.org
Wed Nov 3 12:27:53 MDT 2010


Let me assure you, Richard, that it was an ironic "hmm." This is indeed
a big problem, and one that RDA-if adopted-will make significantly
worse. I'm pretty deficient in my RDA knowledge and was not happy to
learn on Monday about its implications for edition statements.

 

Quite right that OCLC has to be conservative about establishing its own
guidelines that depart from cataloging rules.

 

--Jackie

 

From: dcrm-l-bounces at lib.byu.edu [mailto:dcrm-l-bounces at lib.byu.edu] On
Behalf Of Noble, Richard
Sent: Wednesday, November 03, 2010 7:33 AM
To: DCRM Revision Group List
Subject: Re: [DCRM-L] FW: Slides and Recording from Nov. 1 "When to
Input a New Record" Webinar

 

Thanks, Jackie. But it's more than "Hmm". It is most peculiar that
DCRM(B), concerned as it is with the direct representation of evidence
that, among other things, distinguishes editions; and now RDA, concerned
(at least as its creators profess it to be) with the usefulness of
metadata across the boundaries of various activities, should so utterly
frustrate our ability--anybody's ability--to ensure that recognition of
fundamental entity distinctions should be easily programmable in so
basic a database process as de-duping.

 

Jay said up front that the best way to protect a record against false
de-duping is to add a 250.* Dupes are a scourge--the body of pre-1801
records is now hopelessly infected with them--but at least they are for
the most part available for inspection in the result set of a
sufficiently broad search. False merges are infinitely worse:
unrecognizable without an "item in hand" (and, even with one, seldom
recognizable) and all but impossible to reverse. (How does one re-assign
correct holdings to a record or records resurrected from the limbo of
019?)

 

If the makers of RDA prove resistant to further change, is there not a
role for OCLC to provide for effective "defensive cataloging" by way of
its own bibliographic standards? It seems to matter to the folks in
Quality Control, and I don't see why they should have to throw up their
hands and say "Sorry--you're defenseless."

___________________________________

*This is not a MARC-specific problem--in general terms, it's an instance
of functionally inadequate data definition. It is, by the same token,
too bad that the 503 was killed--at least a programmer could have worked
with that, absent the 250. There are, however, unassigned 25X values,
one of which (251?) could be established, if only in the OCLC database,
as the distinctive tag for a supplied edition statement backed up by a
note. Obviously OCLC is reluctant to depart from the "rules"; but it
could be said that there are such things as rules that depart from
acceptable standards.

 

Obiter dictum: I have very dark feelings about the mind-set that clearly
regards the minimizing of expertise as the governing imperative for the
creation of metadata. "Cataloger judgment" is what we are supposed to
"get over", along with ourselves. But we are, in this instance at least,
talking about the integrity of the database as an accurate record of
bibliographical entities. What else is it for? Retrospective quality
control of metadata that turns out to be inadequate is a lot harder than
quality control in its creation--i.e. ("in english"): Doing it over is
more expensive (and often less possible) than doing it right in the
first place.

 

RICHARD NOBLE : RARE BOOKS CATALOGER : JOHN HAY LIBRARY : BROWN
UNIVERSITY

PROVIDENCE, RI 02912 : 401-863-1187/FAX 863-3384 :
RICHARD_NOBLE at BROWN.EDU 

On Tue, Nov 2, 2010 at 5:44 PM, dooleyj <dooleyj at oclc.org> wrote:

Some of you no doubt attended this OCLC webinar yesterday and have
already received this message. For those who didn't, Jay Weitz gave an
overview of how the guidelines for when to input a new record intersect
with OCLC's latest de-duping algorithms. While much of the information
was quite basic, he did touch on some issues particular to rare
materials during the Q&A-including verifying that no records for
pre-1800 imprints are ever de-duped.

Lots of questions about supplying edition statements in various
contexts, in response to which Jay repeatedly noted that RDA doesn't
allow catalogers to add ed. Statements in the 250 if they don't appear
on the piece. Putting them in a 5xx field won't get picked up by the
de-duping algorithms. Hmm.

-- 
Jackie Dooley
Program Officer
OCLC Research and the RLG Partnership

949.492.5060 (work/home) -- Pacific Time
949.295.1529 (mobile)

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://listserver.lib.byu.edu/pipermail/dcrm-l/attachments/20101103/ce57eed5/attachment.htm 


More information about the DCRM-L mailing list