>Peter,
>
>For each example I have given you you have decided to change the domain definition so the example is meaningless. "My" rules as I stated them is the domain definition of the data and if you choose to discuss some other domain then we are talking about two different things and will never come to an agreement.
>
>Also, my example of the projects was intended to violate BCNF but not 3rdNF until you change the PK. The rules are the domain definition so if you changes those then the whole discussion that follows is meaningless.
>
>You will never get me to agree that surrogate kayes are useless as E.F. Codd himself has said "Intelligent keys aren't". Meaning that a PK that has meaning other than being the PK is a potential problem. Based on this assumption (which I hold to be true based on study and experience) every entity should have a meaningless generated PK. This certainly indicates that many entities, which have a potential Key in their domain definition, will necessarily have more than one candidate for the PK.

Jim, before getting really sad overhere, could you explain "domain definition" please ? I can translate it but just can't place it to the subject.