While it is generally possible to change and modify numbers in an automated
database to accommodate sorting conventions, this is not the ideal solution.
The ID Number is and should be the one constant link between an object and
its documentation, both manual and automated.  It should be EXACTLY the same
on the object and in the documentation.

What seems obvious to people modifying a system (i.e., everyone knows 85.3
means an object from 1885 and *85.3 means 1995) isn't obvious to the people
who will come after us.  How many of us are now hassling with numbering
issues that make no sense to us and cause confusion and needless time and
energy expenditure but were crystal clear to the people who established the
numbering formats?

A better way of handling the numbering sorting dilemma than changing the
number in the automated system and thus destryong the sacrosanct link
between object and documentation is to establish a separate sort field
associated with the ID Number that sets forth the correct numbering
sequence.  This can be a series of sort fields, one sort field for each
segment or component of a number.  Or, can be a zero-fill field.  The series
of sort fields is far more versatile and flexible but is more work than the
single sort field solution.  So, your automated database would have the
following set of fields:

METHOD ONE:  SERIES OF SORT FIELDS

     ID NUMBER      87.3.5a
     SORT ONE        001987
     SORT TWO        000003
     SORT THREE    000005
     SORT FOUR           A
     SORT FIVE
     SORT SIX

METHOD TWO:  SINGLE SORT FIELD

     ID NUMBER      87.3.5a
     SORT                 01987000003000005     A

You do have to make sure that the alphabetical segments sort correctly
(i.e., A before AA or AAA) and this requires knowing what the character sort
sequence is for your operating system or database management system.  Some
systems sort blanks before characters, others characters before blanks.
Similarly, some systems cluster lower and upper case and others don't.
That's why the alphabeticals above are upshifted.

Also keep in mind that the sort fields should not be construed to have
meaning in and of themselves.  Doing that leads to heartbreak.  They are
simply there to induce the correct sort order, nothing more.

Speaking of 2000, I heard an ugly rumor that some museums are hosting
millennium parties on December 31, 1999.  Can this be true?  Since the
millennium doesn't begin until January 1, 2001, this is awful!!!  Certainly
museums should be digging their heels in on this issue and educating the
masses instead of joining in with the great unwashed on the
misinterpretation of what is and isn't an end-date for a millennium
(although looking at the earliest and latest dates assigned to centuries in
some automated systems -- incorrectly 1800-1899 instead of 1801-1900, maybe
I'm assuming too much).   Let's take a stand on this!!!!  Let's rise up as
one!!!  Why did they think the movie was called 2001??

*****

Date:    Fri, 11 Jul 1997 12:15:54 -0500
From:    Emily Nedell <[log in to unmask]>
Subject: Help w new access numbers, yr. 2000

Because of the problem with the date change in the year 2000, we are having
to redo our
accession numbering. We've been accessioning by the last two digits of the
date and then
the order in which it is received, i.e., 97-1, 97-2, etc. 00-1 posses a
problem for our
system. What are other automated museums doing?

Thanks for your help.

Emily Nedell
Museum of Printing History