Re: does a table always need a PK?

Heikki Tuuri wrote:
>
> [..] My view is that
> there are several relational models, and to be precise we should refer to
> them by the author and the paper.

That depends on whether such precision is required or not. I certainly agree that "the relational model" is in some sense a vague concept (otherwise Date could not have refined it as he claims he has done) and some of Codd's rules certainly lack mathematical precision, but the point is that for most purposes it is precise enough and that is what matters.

• Jan Hidders