Re: Normalize until neat, Automate until Complete

From: Kenneth Downs <firstinit.lastname_at_lastnameplusfam.net>
Date: Wed, 24 Nov 2004 22:15:28 -0500
Message-ID: <hbvf72-eg5.ln1_at_pluto.downsfam.net>


Laconic2 wrote:

> 
> "Kenneth Downs" <firstinit.lastname_at_lastnameplusfam.net> wrote in message
> news:o5ve72-r03.ln1_at_pluto.downsfam.net...

>> We hear the term "Denormalizing for performance" very often, and
>> depending
>> upon who is doing it, there might be an exasperated "et tu brute?" along
>> with it.
>>
>> The informal slogan that serves as a nominal rationalization for
>> denormalization is often given as "Normalize until it hurts, denormalize
>> until it works." This slogan, even though it almost rhymes, does not
>> give
>> us much useful guidance. In this slogan normalization is a bad thing,
>> but
>> actually is given no redeeming value. On the contrary, the denormalized
>> system has the advantage that it *works*. So why normalize? Why not
>> just make a system that "works"?
> 
> 
> I've said it before, and I'll say it again.  The problem with
> "denormalizing" is that it doesn't tell you which way to go.
> 
> "Do you know the way to San Jose?"
> "Don't go West."
> 
> Yeah, but...
> 
> When I'm designing a star schema,  it's not that I'm "denormalizing". 
> It's that I'm following a different discipline.

agreed. well put.

-- 
Kenneth Downs
<?php $sig_block="Variable scope? What's that?";?>
Received on Thu Nov 25 2004 - 04:15:28 CET

Original text of this message