At 04:27 AM 6/2/97 UT, Gary Acord wrote:
>EXACTLY, i'm certainly not saying that you must change old numbers. my god,
>what an effort!  those with alpha characters in the number may pose an
>intriguing problem depending on how the system is set up to handle the unique
>field.  but nothing that can't be overcome from the db design point of view.
>just to reiterate my opinion, let's get off of relying on those numbers to
>reflect something useful in terms of information.  since we can sort and
>filter and query data in so many ways now in order to make useful information
>out of it, we shouldn't care how the records are individualized.  AS LONG AS
>THEY ARE.

There is a reason not yet mentioned why museums should try to use, whenever
possible, traditional accession numbers as the joining field in relational
databases.  If, heaven forfend, a vital link file gets ruined or some other
digital disaster strikes, at least the accession number can be used to
rebuild the system.  If arbitrary numbers are used relinking can be nearly
impossible.

Accession numbers, however, because of their variety and because they may
need to be processed in parts, can be difficult to use as joins.  In one
database I built the accession number was actually a concatenation of four
fields that were used for selection and sorting.  Links to the accession
number were created by defining the four fields as a cluster of fields and
indexed and linked as a group.  Not every database can do this, however.




===========================
Robert A. Baron
Museum Computer Consultant
mailto:[log in to unmask]