Re: Thinking about MINUS
Date: Sun, 07 Jan 2007 15:48:06 GMT
Message-ID: <WC8oh.566516$R63.102758_at_pd7urf1no>
J M Davitt wrote:
...
> All this "boolean operations on sets" has me scratching my head. There
> is, underneath it all, the presumption that one really means "is an
> element of," right? I mean, what could the meaning of "NOT (strawberry
> OR apple) AND grape" be?
> ...
That's a pie I don't want to eat. I would like a database I could give to people who are always giving me pie I don't like.
>>>> NOT operator. NOT(A) = U MINUS A.
>
>
> of something that requires significant presumptions that I don't think,
> if I may invent a design on the fly, can be extended to expressions like
> "NOT product" or "NOT order INTERSECT customer" or "product INTERSECT
> NOT order INTERSECT customer."
>
> Sure, we can think of JOIN as an analog to AND and UNION as an analog to
> OR, but they are very different operations.
> ...
Wait a minute, I thought Codd did exactly that - he invented some presumptions in order to extend logic to expressions that are useful for databases to support but remain as susceptible to proof as ordinary boolean algebra. I think analogues are useful only for teaching, just to help focus during the early going, but they are imprecise as far as an implementation is concerned.
p Received on Sun Jan 07 2007 - 16:48:06 CET