Below is LC's response to Adam Schiff's original message and Robert
Bremer's response on behalf of OCLC:


On Thu, 3 Sep 1998, Bremer,Robert wrote:

> Adam Schiff wrote:
>
> > We can't export OCLC #39477127 because the Elvl is set to value E.
> > Apparently, from looking at this record in the LC database, it was
> > originally a PCC record from the University of Chicago, which LC copy
> > cataloged.  Can someone at OCLC explain what the error is that caused
> > the
> > Elvl to get changed to E?  I'm wondering if this is a unique case or
> > whether there are going to more of these we find.
> >
> <snip>
>
> My response:
>
> <snip>
>
> Record #39477127 appeared on our E-level listings as containing
> incorrect data in field 040.  The batchload software used to process
> records received from the Library of Congress looks for a valid USMARC
> library (NUC) symbol in 040 subfield $c.  Because of the OCLC system
> design, that symbol must match up to a corresponding OCLC symbol.  Field
> 040 subfield $a is left as is.  In this particular case, the record
> received from LC did not contain the recognizable NUC symbol "ICU" for
> University of Chicago in 040 subfield $c, but instead contained the OCLC
> symbol "CGU" which the batchload software could not match up against our
> tables of NUC symbols.  Consequently, the 040 subfield $c was changed to
> a default "DLC" symbol.  We periodically sent these error reports back
> to the Library of Congress so that they can change and redistribute
> their records as needed.
>
> Robert Bremer
> [log in to unmask]


LC Perspective on Symbols Used in 040 Fields

     At the time LC began to import records from OCLC in an offline mode,
the OCLC internal forms of the symbols used in field 040 (Cataloging
source) were converted to the USMARC Code List Symbols for Organizations
(UCLO) form before the records were transmitted to LC.  (The UCLO form was
formerly Symbols of American Libraries and originally NUC symbols.)  When
LC began importing records from OCLC through the Z39.50 protocol, the
symbols used in the 040 field were not converted to the UCLO form.  LC met
with OCLC staff and were informed that it was not possible to convert the
symbols when import was executed in an online mode.

     LC determined that use of already existing records for copy
cataloging was a prevailing concern and made the decision to continue to
import records from OCLC even though the symbols in the 040 field would
not be converted to the UCLO form.  LC made the further decision not to
require LC staff to provide the UCLO form at the time of cataloging,
although staff are encouraged to change the symbols to that form if it
readily known.  Concomitant with this decision, LC decided not to change
the symbols when they are reported as errors to LC.  At the time the
decision was made to use the OCLC form of symbols imported in the online
mode, OCLC did not indicate the consequences of this procedure on their
batchload software.

     Sally McCallum, Chief of LC's Network Development and MARC Standards
Office, provided the following statement to the USMARC electronic mail
list on June 5, 1997:

     "The Network Development and MARC Standards Office has great sympathy
with all the comments on the 040 field.  We would prefer that all folks
who use the format use it according to specification, and have tried to
come as close to that ideal as possible ourselves -- although LC is not
perfect either.  The reason for following the specifications are obvious
-- it is the key to efficient interchange of records.  OCLC, RLIN and
others also have a commitment to that ideal but have a few more
differences largely left over from the distant past.  But very few
differences, taking the whole record into account.  They have resource
problems like the rest of us and determine partly with their customers
what they spend the resources on.  If all of us gave them a high priority
to following the letter of the format all the way, they probably would do
that.  For most distribution from OCLC to LC, OCLC uses a special batch
exit program that gives us pretty clean USMARC with the correct symbols in
the 040 -- but it is available to us only for batch files.

     LC is trying to make its operations more efficient as we meet the
funds crunch.  We have the opportunity to have our catalogers carry out a
Z39.50 search from our input system and then instantly incorporate the
record, finish the cataloging and send the item along.  OCLC's Z39.50
access was not initially designed to support cataloging and it simply
pulls the record as in the PRISM database and does not carry out the
conversions that we get with the batch files.  When we first began to work
on this with OCLC the records were coming from First Search and were
missing other crucial items. OCLC worked with us to solve all the problems
except this one, which they could not do in a limited time span.  At that
point we realized that many of you see OCLC records all the time, with
OCLC symbols in the 040 and perhaps the small number of our records that
contain the same would not be too strange to see.  We were unhappy but
realistic.  We are not clear why OCLC allows OCLC symbols in the record
proper, as opposed to the holdings files but it is a fact.

     There is the possibility that OCLC could use the USMARC Organization
codes (formerly called the NUC symbols) in the fields of their records,
although as Rich Greene points out, in some places this would cost a great
deal of programming time at OCLC to condition those processes that use
those codes. Over the last few years LC has worked with OCLC to get
organization code assignments made for all the OCLC cataloging libraries.
I believe that we are up to date on that and have good continuing
cooperation on assignments.  As you know, OCLC publishes the corresponding
USMARC Organization codes in their code directory.

     LC is not currently in a position to fix those codes ourselves,
although with a new system there might be potential.  But it is generally
the responsibility of sending organizations to make records conform to
community developed and approved standards, and if the original sender
does not do this than many to whom the original sender distributes have to
take on duplicative work.  But we all have to agree that that is a
priority.

     As the systems become more global, this will only get worse.  ISO has
just started to work on a 'standard for library identifiers'.  According
to the draft that will have its first ballot in a few months, it would
allow the USMARC Organization codes to be used for US libraries if a
country code for "US" is added before each.  This may be useful
ultimately."

Kay Guiles
Cataloging Policy and Support Office
>
posted by:

Ana Cristan
Acting Team Leader/BIBCO Coordinator
Cooperative Cataloging
Regional and Cooperative Division
Library of Congress, LM537
Washington, DC 20540-4382
(202) 707-7921 (fax): (202) 707-2824
e-mail: [log in to unmask]