Cataloging Spam (Was: Re: Horrible record)

The comment made by a reference librarian to an Autocat cataloger that 
multiple records for a title in OCLC's WorldCat constituted Spam (or is it 
properly spam?) was a telling one.  It's always instructive for us T.S. 
types to get a glimpse of the P.S. types' reactions to cataloging.  The 
multiplicity of records for a title are kind of like spam in the sense that 
they clutter up our list of choices (in the case mentioned, the choice of 
which WorldCat record to turn to for ILL purposes). even they are not (as 
Janet Hill pointed out)created as intentional irritants.  But, of course, 
many factors enter into the answer to this apparent problem.

First, there are cases where binding, imprint, or other differences between 
manifestations of a work justify multiple records, differences that may not 
be apparent in a simple WorldCat search.  Second, there are times (as 
evidenced by Autocat questions) when a cataloger is uncertain about what 
constitutes a legitimate reason to create a new record (but does so 
anyway).  There are instances of records from machine or tape loads that 
generate unchecked-for duplicates. And, of course, there are the many 
duplicate records that have historically plagued OCLC's database, despite 
OCLC's heroic efforts at duplicate detection and resolution 
(DDR).  (Remember when OCLC gave a monetary incentive for libraries to add 
new records to its database?) I think it's safe to state that an enterprise 
as large as OCLC's catalog, into which almost any level of cataloging staff 
can contribute almost any kind of record, cannot be made 
duplicate-record-free.  As I have noted before, WorldCat is kind of 
Wikipedia-like in that sense.

But, as many posters have noted in past discussions, we can do things 
collectively to reduce such duplications, such as building on an extant 
record (even the dreaded Lvl3's) instead of generating a new one, editing 
more records in OCLC (instead of downloading them to local systems and