Re: Do Data Models Need to built on a Mathematical Concept?

"Neo" <neo55592_at_hotmail.com> wrote in message news:4b45d3ad.0305031919.438d71d2_at_posting.google.com...
> > > > What are the flaws and limitations you had in mind?
> > >
> > > 2. That all values in a column be of the same type.
> >
> > This is not generally regarded as a limitation. It is like
> > saying that addition is limited because it cannot operate
> > on dates.
>
> Just as in pure set theory, I want the flexibility that the elements
> can be anything.

Sure; the model totally allows this.

> But, in the rdb model, that flexibility cannot be
> fully realized because the type of values in a domain is restricted to
> some hardware dependent type (ie int, long, date, 50 chars, etc).

That might be a limitation on, say, Oracle, but it's not a limitation of the relational model. Domains can be anything; the relational model doesn't say anything about them at all, other than to say they are sets of values.

> If I
> think of a set of arbitrary thing in my mind, I don't think about
> their type to decide if they can on cannot be included in that set. If
> the things in a set are of different types, I probably would not be
> performing operations such as add or average on them.

Ah, but the things in a set are *not* different in at least one aspect: they are all members of the set!

To the extent that you can say things (loosely speaking) about those members of that set, you can say it within the relational model.

Marshall Received on Sun May 04 2003 - 06:43:35 CEST