On Thu, 27 Jul 1995 Suzanne Quigley <[log in to unmask]> said:

>When I was at the Detroit Institute of Arts, we used an X number system to

>indicate the date the piece was located, thus X1989.1 was the first object

>bearing no identifiable number located in the 1989 inventory.  Such
numbers
>are easily reconcilable against object 'not located' in the same inventory

>- provided full descriptions are on line or at least available.  The use
of
>an X prefix was easily handled because we parsed out the component parts
>of the accession numbers in the database anyway.  A lot depends on the
>sophistication of the software.

The history of museums and collections is often complex.  Careful design of
how accession numbers are entered into a database can produce useful
collection management benefits.

Suzanne's note brings to mind one museum I worked for that was composed of
multiple collections assembled from varying sources, each with their own
set of accession numbers.  The only practical way to create an integrated
collection management system for these collections was to add collection
identification prefixes (letters) to distinguish one collection from
another.  This would insure that there were no duplicating numbers in the
system.  Adding a prefix is always better than changing numbers.

Instead of putting each accession number into a single field, the database
was constructed with four or five fields each holding one segment of the
accession number, i.e. one for prefix, one for date, one for sequence, one
for parts, etc.  This made it easy to sort the collection by any accession
number criteria and made it easy to format lists so that accession number
parts appeared in structured columns.  The database used had a feature that
allowed any set of fields to be treated as one field for purposes of
indexing and relational joins.  Thus the four accession number fields could
be indexed as a unity for uniqueness and could be joined as a group to
other files based on accession numbers.  A separate "lot" number identified
related objects that did not necessarily have sequential numbers -- as when
engravings from a portfolio had disparate accession numbers because they
were donated in yearly batches.

Such a structure would permit several interesting usages:  A multi-part
item could be described part by part while an accession number without a
part indicator could be used for a general description.  Each accessory,
boxes, frames, etc. could receive a number that easily linked to the items
to which they belonged.

The most interesting application of the four-field/unique indexing
structure came during data entry.  The first data entry task was to obtain
a complete list of accession numbers and align these numbers to the filing
categories used for the accession cards.  Using a simple macro program
(Superkey, I believe) it was possible for a single operator typically to
enter over 1000 records per day.  In a little over a month all 33,000
museum objects were represented by a brief record.  At the same time,
because many errors had crept into the manual cards, the database,
insisting on unique accession numbers, caught most of these errors.  They
could be fixed on the spot or flagged for future correction.

While this project was in full swing the museum began its new-wing
expansion, and every object from storage and on display had to be moved and
listed.  A similar database was set up to record the accession numbers and
destination of each moved object.  The intention was to compare the two
lists: the moving list and the accession card list.  By subtracting list A
from B and B from A it would have been possible to develop a list of
accessioned objects not accounted for in the physical inventory and a list
of found object that had no accession card.  Found objects with no
accession number received a temporary identification.  These would form the
pool of objects to be compared to the list of accessioned, but missing
objects.

Museum-l readers of the "Ex Post Facto Accession Numbers" thread should be
aware that the structure and control of accession number registration
schemes may provide one of the most important tools a registrar can have
for collection control.  Indiscriminate manipulation of accession number
forms and content can be quite destructive.  Databases can be set up to add
significant value to the object numbering and registration system already
in place.

Readers of the above note may be interested in a short article I wrote on
computerized accession ledgers.  I'd be happy to e-mail a copy to anyone
who requests it or to post it to museum-l if enough requests come in.
______________________________________

Robert A. Baron
Museum Computer Consultant
P.O. Box 93, Larchmont, NY 10538
[log in to unmask]