[AccessD] OT: The Great Primary Debate

>"Go take a data-modeling course." 


ROTFL

That's like asking a lawyer if tort reform is a good idea!

The kids coming out of Med school still have to do the "36 hours on
duty" thing.  Everyone knows it is just a dangerous thing to do, for
their health, for the patient's health.  But "I had to do it, so you do
too!"

I too remain unpersuaded.

John W. Colby 
www.ColbyConsulting.com

-----Original Message-----
From: accessd-bounces at databaseadvisors.com
[mailto:accessd-bounces at databaseadvisors.com] On Behalf Of Arthur Fuller
Sent: Thursday, June 03, 2004 9:50 PM
To: 'Access Developers discussion and problem solving'
Subject: RE: [AccessD] OT: The Great Primary Debate


It's actually more complicated than that, and while I fall on the side
of ANPKs, I do respect the opinions of such luminaries as Joe Celko, who
despises ANPKs. His reason? They model nothing in the real world; his
premise -- any data model that artificially maps real-world entitities
is by defnition flawed.

He works in different environments than I. Typically I think of the
egg-carton, and question the significance of which two eggs I choose to
make my morning omelet. He thinks in terms of "this alternator was
manufactured by XYZ for use in Ford models T, U and V. It was
manufactured on a day in a factory and signed off by employee ABC as
valid and working. Or to put his argument another way, I am a vendor of
antiques and I have precisely one Queen Anne male chair (the difference
between male and female being the armrests or lack thereof), which was
hand-crafted by Hortense Witherspoon circa 18whatever. 

If I read him correctly (which, given my rapidly increasing senior
moments, may be a faulty assumption), these cases (as opposed to the
eggs in a carton) can all be given primary keys which derive from the
data themselves, and do not falsify the picture by introducing an ANPK.

I don't mean to resurrect the PK debate, nor to say that I agree with
Joe. Concerning the latter, I emphatically do not agree. But he is a
luminary that has written some brilliant books and probably makes twice
the money that any three of us on this list do. Oracle and DB/2 seem to
be his favorite turf, but I could be wrong about that part.

Anyway, I stand clearly on the side of ANPKs and have found none of
Joe's arguments on this subject persuasive. But as he wrote to me in an
email a while back, "Go take a data-modeling course." Well, I took his
advice and took a data-modeling course, and I remain unpersuaded. 

Arthur