|
Tech.
Services >
Stanford University Libraries Redesign Report
Process Change 3
Monographs: Vendor-Assisted Processing
Process Change 3Use
automated batch search services to repeat
bibliographic searches when fuller cataloging
copy is needed.
Assumptions
- Approximately 17% of approval, firm order,
and standing order monographs will have
complete bibliographic records available for
electronic transmission following automated
search of bibliographic utilities.
Activity / Transactions Eliminated
- Cataloging
-
Automate bibliographic search and record
update for items with incomplete records
Full record after automated search: 13,000 transactions
New Process Description
Fuller cataloging copy may be needed for some
titles acquired through the qualifying vendors as
well as from any other source (e.g., gifts,
government documents, non-qualifying vendors). In
the proposed system, with the help of
bibliographic utilities, we will track
bibliographic records for which upgrades are
necessary; the utilities will automatically
re-search for an acceptable record. The range of
"acceptable" for different categories of
languages, formats, genres, etc. will be defined
as part of implementation. While the cycling for
copy is in process, the books themselves are in
the shelving locations as they are now.
If an incoming bibliographic record from a
qualifying vendor is not acceptable (e.g., not
upgraded CIP) the record is loaded to our local
system as is. A local program saves that record
or some fields of it and puts it in a queue for
later re-searching. Materials, including
government documents, acquired from other
vendors, may be coded in the local system so that
they also are put in the queue for
re-searching.
Stanford provides periodic datastream of new
records in queue (full record or selected fields)
to bibliographic utilities for titles to be
searched; our local system number must be
included and stored by the utility to facilitate
eventual overlaying of the preliminary record in
the local system with the fuller cataloging found
in the utility. The utility maintains a profile
which defines acceptable results for its
electronic matching service and which describes
the frequency of searching and the date for
retiring searches. Since the accuracy of an
overlay is questionable in the best of
circumstances, insightful quality control checks
must be in place to maintain record
integrity.
"No-hits" are recycled automatically on a
predetermined schedule until acceptable records
are found or the time for re-searching expires.
The report of expired "no-hits" forms the basis
for original cataloging. Multiple hits reports
are transmitted for staff decision on cataloging
copy.
In the current acquisition-to-access process,
titles are searched individually at multiple
points to download records from the utilities
title by title to our local system. This occurs
at the point of receipt and at the point items
arrive for copy cataloging. These searches repeat
that used for the original order. A number of
months pass between these points, which increases
the hit rate for copy. If no records are found,
the items are placed in a processing queue (PUB)
to be searched again in 24 to 36 months. There is
an extremely high hit rate for copy after this
period.
This new process attempts to capture and store
the information needed for a batch process to
re-search many of these titles. Turnaround time
for receiving acceptable records is greatly
improved by this process, as items are more
frequently searched by the automated batch
process than they can be in the current manual
item-by-item searching that occurs only at the 24
or 36 month PUB recall point. Fuller catalog
records may be used to overlay
less-than-acceptable ones without manual
intervention for over 13,000 monographs which did
not have adequate cataloging upon receipt.
OCLC described its PromptCat service when it
met with Stanford earlier this year and when the
Team visited Ohio in the fall. PromptCat
currently falls far short of what we have
proposed; however, there is growing interest at
OCLC to support the need to recycle searching in
batch mode and transmit records for hits. RLG has
a service on the drawing boards which approaches
what we have conceived. Since the need is not
Stanford-specific, it is quite possible that one
or both utilities will develop services to meet
the market need.
- Partner actively with utilities as they
develop batch re-search services in 1995
- Define acceptable record levels, "matching"
criteria, and search schedules
- Develop mechanisms for exporting/importing
data to/from vendor(s) and utilities
- Define overlaying strategy, exception
reports, quality checkpoints
- Develop workflow for end-processing
materials after new cataloging is received (PUB
books that get a classification number)
 On
to next section
Back
to Table of Contents
Back to
Tech. Services Redesign Page
Last modified:
August 12, 2005 |
