Re: Need help with twisty design problem

From: Amer Neely <>
Date: Mon, 13 May 2002 17:58:00 -0400
Message-ID: <>

--CELKO-- wrote:
> >> The columns beginning with 'User...' are the respondent data. The
> columns beginning 'DM...' are what the respondent submits as the
> person meeting the criteria in 'KeyDM'. However, if the respondent is
> also the person meeting the criteria, I don't want to populate
> the'DM..' columns with the same data. <<
> Why not? They are playing a different role as a "Qualified surveyee"
> than they played as a "first-time surveyee" -- and they did in fact
> qualify themselves. There is a real logical difference here.
> You might want to prune the relationship table down to user_id's and
> have a table of all user's data rather than repeated the attributes in
> two places in the relationship table. Maybe add a few columns about
> when this happened, etc.

I guess I was balking at the duplication, and didn't think of it as 2 logically different sets of data. I'll look at implementing your suggestion about pruning. Thanks much for the insight.

/* All outgoing email scanned by Norton Antivirus 2002 */
Amer Neely, Softouch Information Services
V: 519.438.5887
Perl | PHP | MySQL | CGI programming for shopping carts, data entry
"We make web sites work!"
Received on Mon May 13 2002 - 23:58:00 CEST

Original text of this message