Re: The MySQL/PHP pair
From: Kenneth Downs <firstinit.lastname_at_lastnameplusfam.net>
Date: Thu, 11 Nov 2004 07:52:34 -0500
Message-ID: <it3c62-o82.ln1_at_pluto.downsfam.net>
> support
>
> I really don't want to quibble about this 1NF stuff. You've noted the
> problems. They exist. We, as developers, deal with them every day. I'll
> bet you do too. Normalization is a rule that needs to be broken for a
> variety of reasons. So many reasons, in fact, that one questions the
> rule.
> :-)
>
Date: Thu, 11 Nov 2004 07:52:34 -0500
Message-ID: <it3c62-o82.ln1_at_pluto.downsfam.net>
"Bill H" <wphaskettatTHISISMUNGEDadvantosdotnet> wrote:
>> >> So my original question was about 1NF. You stated that 1NF will not
> support
>> auditable systems, but it seems to me that we will violate 3NF, not 1NF. >> Did I mistake your meaning?
>
> I really don't want to quibble about this 1NF stuff. You've noted the
> problems. They exist. We, as developers, deal with them every day. I'll
> bet you do too. Normalization is a rule that needs to be broken for a
> variety of reasons. So many reasons, in fact, that one questions the
> rule.
> :-)
>
For my own sanity I have been able to come up with a rule on normalization that I have yet to break.
Ruthlessly Normalize all user-entered data (even if the user is an EDI routine supplying the information). The only denormalization is the audit/automation information which is built on top of the normalized columns. This requires the general constraint that a direct INSERT or UPDATE statement cannot affect audit/automated columns, but after that you have the best of both worlds.
-- Kenneth Downs Use first initial plus last name at last name plus literal "fam.net" to email meReceived on Thu Nov 11 2004 - 13:52:34 CET