Re: Implementation of boolean types.

-CELKO- wrote:
> >> I have to say that I buy into Date's viewpoint: relations are predicates, and attributes are the variables of the predicate. <<
>
> I am more inclined toward thinking if them as facts. Facts are always
> "true" in the sense of existing. You do not talk about substitution in
> a set.
>
> Predicates can be true or false (or whatever other logical values your
> deductive system has).

I'm sure you've heard of the closed world hypothesis. I'd say this is exactly what SQL operates under. And it seems to me that calling rows "facts" and calling them "propositions" is entirely compatible. (At least under the CWH.)

> As Dave McGovran points out, SQL has no rule of
> inference, so it is not a logical system.

Natural join meets the definition of inference rule.

> We even call them <search
> conditions> and not predicates.

*Some* people call them predicates. :-)

Marshall Received on Fri Jul 15 2005 - 20:02:15 CEST