Re: Another view on analysis and ER
From: paul c <toledobythesea_at_ooyah.ac>
Date: Wed, 05 Dec 2007 20:03:17 GMT
Message-ID: <9uD5j.1478$sg.563_at_pd7urf1no>
>
> I agree with that, but it doesn't address the issue of whether you design
> an unnormalized schema and then normalize it on the one hand or on the other
> hand start with something from which you can design a schema that will
> already be normalized.
Date: Wed, 05 Dec 2007 20:03:17 GMT
Message-ID: <9uD5j.1478$sg.563_at_pd7urf1no>
David Cressey wrote:
> "paul c" <toledobythesea_at_ooyah.ac> wrote in message
> news:0LB5j.6$iU.1_at_pd7urf2no...
...
>> I suspect that two or more equally "correct" normalizations, correct in >> terms of theory, often suggest themselves.
>
> I agree with that, but it doesn't address the issue of whether you design
> an unnormalized schema and then normalize it on the one hand or on the other
> hand start with something from which you can design a schema that will
> already be normalized.
Why should anybody except possibly recipe-followers, conclude that, eg., a top-down, or bottom-up, or indeed stepwise or iterative style dictates any such order, eg., anything but re-visiting motives such as normalization during implementation, as necessary? Also I would think "as necessary" has more to do with the motives behind normalization.
The notion of "point-in-time" process analysis/design steps has more to do with collective/group project management and organizational dogmas than with skilled interpretation aimed at implementation of those steps. Received on Wed Dec 05 2007 - 21:03:17 CET