Re: object algebra
Date: Sat, 28 Feb 2004 06:09:18 GMT
Message-ID: <igW%b.136141$jk2.563938_at_attbi_s53>
"Neo" <neo55592_at_hotmail.com> wrote in message news:4b45d3ad.0402271658.1d05b22_at_posting.google.com...
> >
> > I am against nulls. I'm not sure, but I think perhaps the only relationship
> > I recognize is set membership. I do not currently see the need for different
> > kinds of relationships.
>
> Applying your logic, that set elements are only determined by mere
> insertion, then in your db the following sets would be acceptable?
>
> Color = {red, green, banana, 8}
> Fruit = {apple, grape, brown, 7}
> Number = {5, 6, blue, peach}
In fact, what you've done is make three sets, and give them names. The names are not the sets; the names are not part of the set; the names are merely names. Every place the set's name appears, we could substitute the set itself, and everyything would still work.
(I would also say that this is definition and not "logic", and that this is set membership and not "insertion.")
> While there are sets where the relationship between it and its
> elements is very, very general, ...
There is no "relationship" between a set and its elements other than the identity relationship. A set IS its elements.
I think you are getting set mathematics and semantics confused. You seem to think that there is some kind of embedded semantics in the set, and that some sets are therefor "better" than others. Sets and their semantics are distinct; the semantics of sets is assigned by humans in their own head; set theory knows nothing of what Date calls the "external predicate." In the same way, the numbers one, two and three have no *innate* connection to apples one, two and three in a basket, even if we use arithmetic to figure out how many apples there will be if we remove one.
Marshall Received on Sat Feb 28 2004 - 07:09:18 CET