Yes, I have applied the technique. The point is, data and relational theory is abstracted until you get to the data model. At that point, you take your best shot at using the equivalents to build the physical model. If you have to fudge a bit to normalize, so be it. I recall that in Jacobson's OOSE book it basically says to do it this way and it was no big deal if you had to break with the object model to get a realistic physical data model I believe his conclusions were, in the end, do what works.

>Hiya Gonz,
>I've tried to wrap my brain around such methodologies, even sat down and tried to design a dbc for a home app project with such an approach. I got as far as having 4 rough table structures scribbled out on paper and seen great potential for braking 2nd and 3rd normal forms. The result of this little exercise was I sat and stared at my rough design for approximately 2 corona's, and decided it scared the hell out of me (and happily filed it in the trash can halfway thru the third corona lol). So my question for you is have you tried to apply such a methodology??