Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!mit-eddie!uw-beaver!sumax!polari!6sceng!6sigma!zuker
From: zuker@6sigma.UUCP (Hunter Zuker)
Newsgroups: comp.databases
Subject: Re: Balancing Normalization with Performance
Message-ID: <355@6sigma.UUCP>
Date: 26 Nov 90 19:44:32 GMT
References: <1092@cortex.med.jhu.edu> <1990Nov20.190315.13505@informix.com> <33265@netnews.upenn.edu>
Distribution: usa
Organization: Six Sigma CASE, Inc.
Lines: 29

Aaron Watters writes:

>ALTERNATIVELY there is the phenomenon of the UNJUSTIFIED
>NORMALIZATION.  Where we make use of an assumption such as
>   Persons only have one address.
>   Husband's last name determines Wife's last name.
>etcetera.  In this case the normalization may DISALLOW information
>that a user may later want to insert into the database.  This
>is another issue.			

I would argue that these aren't "UNJUSTIFIED NORMALIZATIONS".  I would
say that the data was not modeled correctly.  

There was probably a time that it would have been an accurate model
to say that a wife's first name is an attribute of the husband.  Of course
now this is not accurate.  Exactly how this should be modeled will
depend on what you are building the database for.  

In many databases it is sufficient to build the data model to have a
person have only one address.  In others it may not be sufficient.  Once
this decision has been made, then normalization is used to determine
what the tables look like to optimize your data model.

You should normalize after you have modeled your database.

Hunter
-- 
Hunter Zuker		Six Sigma CASE, Inc.	 13456 SE 27, Suite 210
zuker@6sigma.UUCP       (206) 643-6911           Bellevue, WA 98005-4211