Re: Proposal: 6NF

vc wrote:
>
> Strictly speaking, Z is not a subset of R or even of Q (some stucture
> is acquired some is lost when one moves between R, Q and Z), but the
> subtlety is rarely important so one can say that Z is a subset of R (or
> Q). A more correct way would be to say that Z is isomorphic to a
> certain subset of R (or Z is ebedded in R).

Hmm, wouldn't it be more correct to say that most mathematical definitions of these concepts either define them up to isomorphism (so the question doesn't really make sense) or define them explicitly as subsets (Z is the subset of R for which ...)?

> > I think
> > the concept is called "specialization by constraint."
>
> What's that ?

Defining a subclass A of B such that it is exactly the subset of B for which a certain predicate holds. It usually means that objects in A have no extra attributes except those already specified for objects in B.

• Jan Hidders