When I use commercial databases and botch up my search, their fuzzy logic subroutine takes over and starts prompting me with, "Did you mean [XXX]?" statements to let me know something is amiss.  Perhaps LC thinks we will all be fuzzy soon.  I would love it if OCLC and our local system had fuzzy search logic, but I would, of course, want to know when it kicks in, so I'll know I messed up my query.

Is this a valid premise regarding fuzzy logic?


Dean Rogers
Map Cataloger
U.S. Geological Survey Library
Reston, Va.

"Adam L. Schiff" <[log in to unmask]>
Sent by: Program for Cooperative Cataloging <[log in to unmask]>

04/28/2006 01:46 PM
Please respond to
Program for Cooperative Cataloging <[log in to unmask]>

[log in to unmask]
series keywords won't always work

Here's an interesting real example of how relying on just keywords for
series without a series authority record loaded into a catalog is not
going to work well.

The Wildlife Conservation Society publishes a monographic series that has
been established as WCS Working Paper.  On the analytical title page of
the individual monographs, the series title appears variously as WCS
Working Paper or Working Paper.  Some issues have a series title page that
says WCS Working Paper Series.  And many issues have the title WCS Working
Papers on a publisher's listing of all the issues or on the analytical
t.p. verso.

In our OPAC, a keyword search of a string of words first searches for that
string together.  If nothing is found, the system will execute a boolean
"and" search of each word.

A user doing a keyword search on the title WCS Working Papers retrieves
absolutely nothing, because that variant is never found on a source used
for transcription of the series statement.  On the other hand, if the user
did a title browse search of that variant, they would get a referral to
the controlled form of the title used as the series added entry, WCS
Working Paper (the only difference is the singular/plural form of
Paper(s)) because a series authority was created documenting all the
variants and was loaded into our OPAC.

For monographic series that are cataloged as separates (i.e. classed
separately), our current policy is that we do not create a public
series/serial record in our OPAC.  We rely on the series added entries to
collocate the series.  If we stopped tracing controlled series access
points and producing series authority records, I wonder if we might need
to start creating serial records for these resources, since only there
would one find all the variant titles for the series traced in that
bibliographic record, which would provide the keyword and variant title
access that could get people to some resources.

My main point is that keyword access in our (and other) OPACs will only
work when the user searches on keywords present in a bibliographic record.
We need both keyword and controlled access through the presence of series
authorities to get users to some resources.

Adam L. Schiff
Principal Cataloger
University of Washington Libraries
Box 352900
Seattle, WA 98195-2900
(206) 543-8409
(206) 685-8782 fax
[log in to unmask]