Re: DB clasical structure violation

From: Jan.Hidders <hidders_at_hcoss.uia.ac.be>
Date: 9 Jul 2002 15:47:55 +0200
Message-ID: <3d2ae98b$1_at_news.uia.ac.be>


In article <2c63b9ee.0207090028.667b45a7_at_posting.google.com>, Anthony Youngman <anthony.youngman_at_eca-international.com> wrote:
>
>What do you mean? Newton's theory of Gravity is presumably also a
>tautological abstract reasoning.

The theory of gravity is much more direct in making predictions about what we see in the real world. In set theory itself the abstract concept 'set' is studied and nothing is said about how it relates to reality. Can you tell what predictions are derived from the fact that there are more real numbers than integers? Are there any measurements we can do to verify or falsify this?

However, it will be quite clear that set theory is based on our tendency to classify things (some of them in the real world) and to reason about these classifications. Since almost all reasoning can be regarded as some kind of classification it follows that set theory has some relevance here.

>But I then have to map this to the observed world, and to my mind all too
>often set theory doesn't map comfortably with what I observe.

In that sense the relational model is not based upon set theory. It's semantics may be defined in set theory, but that is also the case for OO data models, MV data models, and almost all other data models.

It is much more correct to say that the relational model is based on *first-order logic* because there the data is presumed to exist of things and predicates over these things with the restriction that these predicates cannot themselves also be arguments of predicates. Do you see the correspondence between "Columns of relations cannot contain relations" and "Arguments of predicates cannot be predicates"?

Higher-order logics, for example, are not so picky and also allow predicates over predicates, which gives you MV data models, or the nested relational model, or whatever you want to call it. But also the semantics of higher-order logics are stil defined in set theory.

So next time, please blame the right theory. :-)

  • Jan Hidders
Received on Tue Jul 09 2002 - 15:47:55 CEST

Original text of this message