Since late 1987, a small group of librarians at Cornell has been trying
to plan for the implementation of MARC holdings on NOTIS.  At that time,
NOTIS was planning to begin implementation in 1988, but of course that
never happened.  We spent the time writing internal guidelines for
holdings statements on the VHLD record, writing specifications for the
conversion of VHLD to MHLD, and being the court-of-last-resort for
sticky holdings problems.

As MHLD seemed to come closer to reality (announced for 5.0 and 5.1), we
geared up again to look at our 1987/88 decisions and prepare to train
our staff to use tagged holdings fields, which we had determined to use
as soon as possible.  Imagine our surprise then to see in the April 1991
NOTISes a "reminder" that 85x/86x fields would not be used any time soon
for display in MHLD records.  This knocked us for a loop, needless to
say, since to our way of thinking, MARC Holdings meant full content
designation, not a mere substitution of 866/7/8 numeric tags for
numbered lines on a VHLD screen.

We went through "channels" and requested clarification of this through
our Systems Office, and were told that we should look at the
demonstration of the new check-in system at ALA, since the 85x/86x
fields were being implemented in 5.1.  Huh?  After some questions to
various people who we though might have a good bead on this issue, we
have gleaned (we think) reasonably hard information, which has not
comforted us much.

It seems that the "MARC Holdings implementation" that has been billed
for 5.0-5.1 is a very strange animal indeed.  NOTIS' traditional
non-relationship between check-in and holdings is unchanged, and a lot
of hoopla is going into the "re-write" of the check-in system for 5.1.
Apparently, a "template" (supposed to be based on 85x/86x fields and
subfields) will be available, allowing predictive check-in.  This nicely
insulates the check-in folks from having to deal with MARC tagging--a
dubious value at best.  In some manner, unclear to us at this point,
pattern and enumeration/chronology information will be stored as 85x/86x
fields in the acquisitions record, viewable in tagged form in staff mode
and displayed from these same 85x/86x fields in the OPAC.

The bad news is that somewhere along the line, information from these
85x/86x fields will need to be moved (manually, of course) to the new
"MARC holdings" screen.  There it will be translated BACK into 866/7/8,
unsubfielded form.  Unless, of course, you choose to maintain two sets
of data, one for display only, the other for communication and practice
in tagging.  This is progress?

It seems to us that NOTIS needs a firm message from their customers that
this state of affairs is not acceptable.  At the very least we need to
put a "10" or "9" beside the Serials SIG enhancement ballot item which
states: Implement full functionality for the paired sets 85x/86x in the
MHLD record.  This is galling, certainly, when such functionality SHOULD
have been integral to the original implementation.  And how many
libraries will skip over that item, thinking it is already accomplished,
or not recognizing its importance?  An additional message can be given
by speaking directly on this issue to Jane Burke and other NOTIS staff
members at ALA, as well as responding positively to this message.

Diane I. Hillmann
Cornell Law Library
Chair, Task Force for the Implementation of MARC Holdings, Cornell U.
Lib.
dh5@cornellc.bitnet