Jim,

This may be where we disagree. I don't think it is enough to say "more than one candiate key..." I think there has to be more than one concatenated candidate key - each sharing a common data element. If you don't have this, the BCNF does not apply. And, no single field can have the ability to serve as a PK. Therefore, if you have a surrogate key, it by defintion is sufficient to be the PK. Therefore, if you have a surrogate PK, the BCNF does not apply. You just need to get the the 3rd. NF, and perhaps the 4th. NF if M:M relations apply.

< jvp >




>Martin,
>
>>I absolutely agree with Codd's statement. What I don't understand is why you said "making 3rd NF not quite good enough". Can you elaborate on this?
>
>As I said in my reply to , if there is more than one candidate for the primary key then the Boyce-Codd NF applies. It says that once a table is in 3rd NF for the chosen PK it must also be in 3rd NF for all feasible alternate keys.