Re: Date of Birth as an attribute

> I agree. I also much prefer the OO method of analysis over the ER.

Taking it one step further. A good OOA model can be used to derive an ER model in a matter of hours.
It's just a projection onto the data requirements view.

> IMHO OO has all but been destroyed by the OO Programming types who believe
> that OO is synonymous with hacking. Extreme Programming and Agile have
> institutionalized this.

I said as much when I said that the problem is not that they are object oriented. The problem is that they are programmers.

> There is no difference. ERD just refers to the diagram, as opposed to the
> model that the diagram describes.

OK. The reason I ask, is that, to me, ER modeling is incomplete without the attributes and the domains of the attributes.

From there, A Relational model is just a matter of hours, if that. And a physical model a few more hours.
And an SQL build script just a few more hours.

And you should have a database ready by the time the programmers are ready for it.

But it all depends on analysis. (Or maybe "discovery" would be a better term.) Received on Thu Apr 08 2004 - 03:58:38 CEST