Re: plz normalize this table for me

From: paul c <>
Date: Sun, 02 Dec 2007 15:18:18 GMT
Message-ID: <_0A4j.14708$UQ1.1954_at_pd7urf1no>

Jan Hidders wrote:
> This group is not a home-work service. But some help will be provided
> if you show that you have made serious effort to solve the problem
> yourself and can make clear and explicit why you are now stuck. You
> could start by saying which functional dependencies you think hold,
> which normal form you would like to achieve, what you think the steps
> of the normalization proces are and which of these steps you have
> difficulties with.

Right you are. Also, I remember being criticized by some pedant for including a "post code" in the same base relation as the address. Often the "business" purpose is sometimes more important to decide which FD's to choose. Some people think the use of codes should always involve an appeal to external authority, even for the simplest of systems. (I usually read that technocrats as meaning higher authority. Ralston-Saul calls technocrats the descendants of reason.)

The question looked like it had to do with Canada, where some postal codes don't determine address, eg., H0H H0H determines a person (or should I say entity?) named Santa Claus! Such a letter will get delivered even to him even if his North Pole address is omitted. Even if that weren't so I think it depends on the app whether one wants to use FD's for integrity checking as opposed to structural normalization.

This might help demonstrate why it is always harder to explain a design! Received on Sun Dec 02 2007 - 16:18:18 CET

Original text of this message