<font size="4"><font face="garamond,serif">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.</font></font><div>
<font size="4"><font face="garamond,serif"><br></font></font></div><div><font size="4"><font face="garamond,serif">Jay said up front that the <i>best</i> 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?)</font></font></div>
<div><font size="4"><font face="garamond,serif"><br></font></font></div><div><font size="4"><font face="garamond,serif">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."</font></font></div>
<div><span class="Apple-style-span" style="font-family: 'courier new', monospace; ">___________________________________</span></div><div><font class="Apple-style-span" face="garamond, serif"><span class="Apple-style-span" style="font-size: large;">*</span></font><span class="Apple-style-span" style="font-family: garamond, serif; font-size: large; ">This is not a MARC-specific problem--in general terms, it's an instance of functionally inadequate data definition. </span><font class="Apple-style-span" face="garamond, serif"><span class="Apple-style-span" style="font-size: large;">It </span></font><span class="Apple-style-span" style="font-family: garamond, serif; font-size: large; ">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.</span></div>
<div><span class="Apple-style-span" style="font-family: garamond, serif; font-size: large; "><br></span></div><div><span class="Apple-style-span" style="font-family: garamond, serif; font-size: large; "><i>Obiter dictum</i>: 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.</span></div>
<div><br></div><div><div><span class="Apple-style-span" style="font-family: 'courier new', monospace; ">RICHARD NOBLE : RARE BOOKS CATALOGER : JOHN HAY LIBRARY : BROWN UNIVERSITY</span></div><div><font face="'courier new', monospace">PROVIDENCE, RI 02912 : 401-863-1187/FAX 863-3384 : <a href="mailto:RICHARD_NOBLE@BROWN.EDU" target="_blank">RICHARD_NOBLE@BROWN.EDU</a> </font><br>
<br></div><div class="gmail_quote">On Tue, Nov 2, 2010 at 5:44 PM, dooleyj <span dir="ltr"><<a href="mailto:dooleyj@oclc.org">dooleyj@oclc.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div>
<font face="Calibri, Verdana, Helvetica, Arial"><span style="font-size:10pt">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.<br>
<br>
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.<br>
<br>
</span><font size="4"><span style="font-size:10.5pt">-- <br>
Jackie Dooley<br>
Program Officer<br>
OCLC Research and the RLG Partnership<br>
<br>
949.492.5060 (work/home) -- Pacific Time<br>
949.295.1529 (mobile)<br><br></span></font></font></div></blockquote></div></div>