Re: The problem with denormalization.
Date: Tue, 28 Sep 2004 07:57:47 -0400
Message-ID: <iNadnWhl3bpOzMTcRVn-vA_at_comcast.com>
"Tony Andrews" <andrewst_at_onetel.com> wrote in message news:1096367125.055253.224010_at_k17g2000odb.googlegroups.com...
> Nor me: normalisation is formalised common sense. My table designs
> just ARE normalised, without me having to apply any normalisation
> algorithms.
>
Agreed.
I should have added earlier that I never worried about normalisation beyond 3NF. My attitude was that I would worry about it when the time came.
Where normalisation really helped me was in the analysis of given application files. The problems caused by lack of normalisation are a real easy way of learning the FDs, if you aren't budgeted for a formal analysis project.
In fact, my introduction to 1NF, 2NF, and 3NF was in a magazine article that explained how you can use normalisation to your benefit in the design of indexed files. At that point in time, those were precisely the tools I had to work with. I was building indexed files with RMS (a layer of the VMS file system). The article really helped formalise my common sense. Received on Tue Sep 28 2004 - 13:57:47 CEST