I disagree. Good experience and hard work can teach you why denormalization is a bad idea. However, it's not universally a bad idea. It has to be applied to what you are trying to accomplish. Formal education helps you to understand these problems, however, it is not essential.
The best data modeller I ever worked with started off spending 20 years working for the forestry service, running around in the woods.
- Original message --------------
> Hi Stphen,
>
>
> >=20
> > I think there has to be a balance between producing pretty diagrams
> > and producing a system that will actually fulfil the spec and do what
> > it's supposed to do.
>
> or perhaps we can say that a sound description of the data model in
> mathematical terms is necessary BUT is not enough. For example if=20
--
http://www.freelists.org/webpage/oracle-l
Received on Tue Jun 14 2005 - 14:11:36 CDT
