Re: Throughput and storage requirements questions for modern vs old database processing apps

From: Anthony W. Youngman <thewolery_at_nospam.demon.co.uk>
Date: Wed, 19 Dec 2001 22:15:59 +0000
Message-ID: <oXvYmWAfGRI8EwdS_at_thewolery.demon.co.uk>


In article <P03N7.1236$Ao6.146282_at_newsread1.prod.itd.earthlink.net>, John Becich <jbecich_at_nospam.net> writes
>Thanks for your response. I have a few questions that perhaps you might be
>able to help with.
>
>I have been told that our Clipper-based rental system is NOT a relational
>database system. "All other things being equal," would a relational
>database system be faster?

NO!

What a relational database does is add things like integrity, security, etc. This slows things down! and is only acceptable because modern machines are so fast that it doesn't matter too much.

I don't like relational databases for precisely this reason. Relational *DESIGN* is important, and *should* *be* *done*. But all too often the design is then implemented as a relational implementation, and this is quite likely to be sub-optimal.

Relational Theory is basically a variant of mathematical Set Theory. Just like with maths and physics, the fit between theory and practice is sometimes excellent, and sometimes crap. Unfortunately, PHBs seduced by the latest buzzwords have a tendency of demanding the latest theoretical solution, in total denial of the fact that the maths itself declares itself ineligible for the proper solution of the problem at hand.

Cheers,
Wol

-- 
Anthony W. Youngman - wol at thewolery dot demon dot co dot uk
Witches are curious by definition and inquisitive by nature. She moved in. "Let 
me through. I'm a nosey person.", she said, employing both elbows.
Maskerade : (c) 1995 Terry Pratchett
Received on Wed Dec 19 2001 - 23:15:59 CET

Original text of this message