Re: USMARC and postgresql?

On Tue, May 09, 2000 at 08:44:16AM -0400, Andrew Sullivan wrote:

> The commercial vendors of SQL-based library automation systems do this with
> BLOBs.

Sorry to respond to myself, but I happen to be in the middle of looking
demos for a library automation system, and I just found out that _not_
everyone uses BLOBs for this. SIRSI Corp's Unicorn system does not. It's
built on top of Informix or Oracle (you get a real SQL database iff you pick
the Oracle option) and it does not store the MARC records as BLOBs. What it
does use I don't know, although the sales reps told me they'd be happy to
get me in touch with their database guru. (Anything for a sale, I guess.)

I'm guessing that the other suggestions here are probably going in the
right direction. MARC is a nasty pain from the point of view of database
design (although it's got nothing on the foolishness that is the punctuation
standard!).

There are, by the way, some other things to consider when thinking about
MARC. Don't forget that the ALA standardised on the ANSEL character set, so
diacritic support in the MARC records you have (or might get easily) will
probably not be what you expect. I know that the Library of Congress has
uploaded some data with ISO character sets instead; at least one other
company (DRA) does this with Unicode (DRA is also using OOSQL, so the
buzzwords start flying fast and heavy).

Just thought I'd correct myself, for the sake of accuracy.

--
Andrew Sullivan Computer Services
<sullivana(at)bpl(dot)on(dot)ca> Burlington Public Library
+1 905 639 3611 x158 2331 New Street
Burlington, Ontario, Canada L7R 1J4