It all depends, First off all, you've got to identify whether this database is going to be used by different applications (from different vendors).

Well, these things often can't be predicted.

Third, I don't think that the excercise we did is so difficult that it would save a lot of effort or avoids a lot of duplication when this was done seperately if neccesary.

Sorry, but I just can't agree with that one at all. That's a maintenance pain and is risky of backfiring, especially on anything other than small teams.

As you can see in my example (which is far more generic)...

??? How's that? How is your example far more generic?