Re: 3NF question

On 10 Jan 2005 15:46:48 -0800, "vldm10" <vldm10_at_yahoo.com> wrote:

>
>
>This is not necessarily true. There are first three steps in designing
>db applications and solutions: 1. Analysis; 2. Conceptual model; 3.
>Relational model (or other DB models)
>It is not necessary to set column-name rules from Relational Model in
>step 1 as a must (especially not during setting the problems, first
>meetings, interviewing, intended model etc).
>So as the begging of the 1st step we have Elves's interpretation of
>the problem. His use of the name ID in that sense is OK. You jumped
>directly to 3rd step skipping the first two steps and demanding
>Relational Model terminology.
>
>Vladimir Odrljin

Sorry Valdimir, I believe your arguments are weak ;-)

When in 1st step (Analysis) I believe the Analysts should be very careful about introducing their own ID's just to satisfy a potential need in the 3rd step !
The analysis should focus on the users UoD and adapt their nameing of information elements, then analyse, define & standardize based on them.

It's not an option to create a table with names that only "a knowledgeable person of many database details" (understands), as you indicated in anoter parth of this discussion.

Regards Tore Received on Wed Jan 12 2005 - 20:53:46 CET