Re: Pizza Example

"Dawn M. Wolthuis" <dwolt_at_tincat-group.com> wrote in message news:c4so4g$dle$1_at_news.netins.net...
> Looking at the benefits and deficits of various models makes
> sense, but if we are to anoint one as THE model (which our profession has
> somewhat done with the relational model), we should be very careful we
don't
> go and apply the theory of relativity to very small objects (an
analogy) --
> we should use it where it is useful. From what I can see, the relational
> model is not close enough to a generalized TOE (theory of everything) for
> data that we should stop reviewing other options.

Couldn't agree more - we should be. The value of O-O is still far from concluded. What's a little different about relational is that it's derived very directly from logic and set theory, and also is the implicit base in most other theories of data - you've said as much yourself regarding multivalues. However, the "extensions" can corrupt the base, which I believe is that case with most other theories, and furthermore they often add nothing.

> I was in the middle of another post that relates to language when I saw
your
> quote here and it fits nicely with that so I could just jump into that
topic
> here, but instead I'll just agree with the comment and note that it seems
> likely that we need not just mathematics and logic, but also language in
our
> discussions of data. Cheers!

True, but I've not seen anything formal, which a computer system needs to be. Chomsky, Wittgenstein and countless others have theories of language; have they been applied to data? So far, all I've heard is "we should consider language", but that's not much good unless we get more concrete. As I said, we can't ever stop thinking and evaluating... but we need something fairly concrete.

• erk