-------- Original Message --------
Subject:        OCLC QC Tip of the Month / July 2010
Date:   Tue, 6 Jul 2010 12:12:17 -0400
From:   Goodson,Luanne <[log in to unmask]>
Reply-To:       askqc <[log in to unmask]>
To:     <[log in to unmask]>



This message is being widely cross-posted.



*************************************************************************************************************



OCLC thanks everyone who adds new records to WorldCat.  OCLC has spent
the past several years working on a re-implementation of its Duplicate
Detection and Resolution (DDR) software in the Connexion environment and
to expand its capabilities to deal with all types of bibliographic
records.  Between May 2009 and January 2010, OCLC ran small subsets of
WorldCat against the live database in order to fine tune its algorithms,
examining each resulting merge and learning from both the successes and
the failures.



The new DDR software is now in full operation.  DDR began running
through the full WorldCat database (beginning with OCLC #1) on Feb. 2,
2010.  In addition, a separate process that examines selected new
records and replaced records from a day's journal files began running
Jan. 26, 2010.  As of the end of June 2010, 2,919,942 duplicate records
have been removed out of 67,179,212 records processed.



DDR processing will continue for a number of months.  As a result, you
will notice fewer duplicates, particularly for printed music, sound
recordings, and audiovisual materials since the original DDR software
only dealt with records for books.



Like all automated processes, this new DDR will make occasional errors
in spite of our best efforts to minimize such cases.  Thank you for
reporting erroneous merges.  OCLC staff will examine the records in
question, reverse any merge deemed to have been inappropriate, and try
to assure that such incorrect merges do not occur again.  One or more of
the records may be edited so that our algorithms are better able to
identify important differences.  Additionally, we will determine if we
can learn something more general from such instances and further refine
our algorithms to reduce such errors in the future.



We urge users to help us make this process more effective by
re-searching your title immediately before entering a new record into
WorldCat, especially if the record was previously in an Online or Local
Save file.  Duplicates reduce the efficiency of the database, so please
always verify that a record has not already been added by another library.



For information on When to input a new record please see Bibliographic
formats and standards Ch. 4
http://www.oclc.org/bibformats/en/input/default.shtm



For more information about DDR please see
http://www.oclc.org/worldcat/catalog/quality/ddr/default.htm



Duplicates and reports on possible erroneous merges can be reported to
[log in to unmask], or by using the Action Menu--Report Error function
while viewing a bibliographic record in Connexion.  This function opens
a window which is free-text, allows users to have a copy sent to their
own email address, and includes a snapshot of the record as it appeared
when the function was invoked.  Additionally there is a webform
specifically for reporting duplicates found here under Forms:
http://www.oclc.org/us/en/toolbox/default.htm





Please send any questions or concerns to: [log in to unmask]
<mailto:[log in to unmask]>







Luanne Goodson

Consulting Database Specialist

OCLC Quality Control Section

6565 Kilgour Place    MC 139

Dublin, Ohio, USA 43017-3395