From: Robert Paulsen <robert@paulsenonline.net>
To: pgsql-sql@postgresql.org
Subject: Re: Question re: relational technique
Date: Mon, 13 Mar 2006 06:15:17 -0600
User-Agent: KMail/1.9.1
Cc: Richard Huxton <dev@archonet.com>,
	chester c young <chestercyoung@yahoo.com>
References: <20060312172903.6034.qmail@web54305.mail.yahoo.com>
	<200603121215.46141.robert@paulsenonline.net>
	<4415355F.2020001@archonet.com>
In-Reply-To: <4415355F.2020001@archonet.com>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200603130615.17335.robert@paulsenonline.net>

On Monday 13 March 2006 03:03, Richard Huxton wrote:
> Robert Paulsen wrote:
> > This still requires me to modify the overall database structure but not
> > the original item table. As my reward :) I get to use any type I choose
> > for each new attribute.
>
> The whole point of the database structure is to accurately reflect the
> requirements of your data. If you don't want your change your structure
> to keep track of the real world, why bother to structure it in the first
> place? Just stick it all in text documents and let htdig free-text
> search against it.

Requirements change and differ from one application of the datbase to another. 
The database structure is maintained by others and is used by several diverse 
locations. It is an effort to incorporate and coordinate changes. The 
database already uses the name-value technique in one place, probably for 
this very reason. I was suspicious of the technique so posted my original 
question. The answer given is a resonable compromise. I can have my own table 
whose structure I control, even though the fields in the table "really" 
belong in another table.

Bob