-----(1)

>From BibleD@DOAKS.ORG Tue Oct  6 15:58:31 1998
Date: Tue, 6 Oct 1998 08:46:41 -0400
From: "Bible, Donna" <BibleD@DOAKS.ORG>
Subject: Re: Killing the Kardex (Susan Scheiberg)

        We switched from Kardex to serials control records on a SIRSI system
approximately one year ago.  The project of creating serials control
records was quite complex.  I am surprised that you were able to
commission this task since each title is quite unique, with regards to
frequency, date of receipt and pattern variations.

        We opted to create MARC holdings records automatically with the receipt
of each new title.  This is one of the biggest virtues of the SIRSI
system, in my opinion.  I would advise that you choose to do the same,
because if the Kardex is to be replaced then you will have to indicate
to the public that this issue has been received.  Using the control
record to generate MARC holdings information is a bit tricky, but it's
well worth the effort.

        One piece of advice is to be cautious about the links between your
serials control, checkins and MARC holdings records.  Each time that the
control records' pattern information is updated, the link of the control
and the corresponding holdings record automatically increases.  If you
have checkin records already in existence then their links will have
become obsolete.  Checking records in at that point will result in lost
data- because the information won't register in the holdings record
since the links no longer match.  So you'll think you've received
something (and the checkin record will appear to have been received) but
the public won't see that the issue was received.   So "tweaking" in
this sense could cause your checkin records to lose their corresponding
link to the MARC holdings records.  Beware, and keep tab, of each link
and each holdings record as you begin to add new information.

        Furthermore, not all of the patterns that we received were supported by
SIRSI.  A few basic patterns did not generate 853s according to the
standard.  Also, indices, supplements and annual reviews are all output
as 866s.  Again, this is something that stumped me for awhile.  My
advice on this note is to not expect the system to do all of the work on
that front for you.  Assume that such issues will have to be manually
entered into the holdings record.  This should not prove problematic,
but it's good to know what is automatic and what is not.

        Good luck!  I loved the project, but it took a lot of patience.

Donna Bible
Library System Administrator
Dumbarton Oaks
1703 32nd Street NW
Washington, DC 20007-2961
e-mail:  bibled@doaks.org
-----(2)

>From leadem@NIEHS.NIH.GOV Tue Oct  6 15:58:31 1998
Date: Tue, 6 Oct 1998 09:19:41 -0400
From: "Leadem.Ellen" <leadem@NIEHS.NIH.GOV>
Subject: Re: Killing the Kardex (Susan Scheiberg)

Here at the National Institute of Environmental Health Sciences, we embarked
on a much smaller though similar project, although we were migrating our
serials control operation from one automated system to another.

Our serials had been checked in  for approximately a decade on Faxon's
Microlinx system. Prior to purchasing our current ILS system, Horizon, we
loaded all of our bibliographic records, and attached a simple summary
holdings statement. When we converted to Horizon, we only needed to create
copy and check-in records.

We prioritised the work according to the most common type of frequency.
Since the vast number of our serial titles were monthly, we focused setting
up the prediction patterns for those first. This simplified the training for
the student worker assisting on this part of the project. We next handled
the second most common frequency, bi-monthlies, etc. We had to work with
under a thousand titles. It took approximately 4 months for a half-time
worker to complete... although this person was also completing other serials
tasks.

We created a single statement of our holdings, prior to live check-in on
Horizon. We then commenced live check-in with first issues for the
subscription year 1997.

Good luck.

-----(3)

>From dilewis@IGSRGLIB01.ER.USGS.GOV Tue Oct  6 15:58:31 1998
Date: Tue, 6 Oct 1998 10:32:35 EDT
From: Diane Lewis <dilewis@IGSRGLIB01.ER.USGS.GOV>
Subject: Re: Killing the Kardex (Susan Scheiberg)

Hi, Susan!

Although your library has twice as many serial titles as ours and we
are on another system (Ameritech's HORIZON), perhaps some things from
our experience may be of interest to you.

The speed with which you transfer check-in records to SIRSI may to a
large extent be determined by the complexity of the prediction,
holdings, and copy records in that system as well as the makeup of
your incoming titles.  Are the predictions, etc. easy to set up, do
you have a lot of titles with unusual chronologies or enumerations,
will you be following the NISO standard Z39.44 MARC Format for
Holdings, and if you are, what level will you follow?  These are the
kinds of questions that will determine your approach to this project.

Some practical suggestions:

1.  Save the irregularly or erratically publised items for last.
2.  Set up templates for the various prediction, copy, and holdings
patterns.
3.  DO NOT throw out or dissect the Kardex until ALL of the
information in it is transferred to the new system.  Even then, you
might consider retaining (for a while) the old, unified Kardex for
quick lookups and times when the system is down.
4.  With a large title base such as yours, the Kardex is probably
replete with notes and records that tell the check-in staff important
stuff.  How does SIRSI handle records for titles that are not
kept?  We create bibliographic records in HORIZON for these and then
restrict them to staff view.
5.  We had not been able to claim regularly for years.  As a result,
we came across many inactive titles.  Will you claim these and then
profile them as the current issues arrive, or will you just set up
records for them anyway?