Re: OO's best feature survey results

From: Shayne Wissler (thales000_at_nospam.yahoo.com)
Date: 11/03/03 

Date: Mon, 03 Nov 2003 19:37:05 GMT

JXStern wrote:

> On Mon, 03 Nov 2003 17:05:26 GMT, "Shayne Wissler"
> <thalesNOSPAM000@yahoo.com> wrote:
>>
>>When you say RDBMS are on a "rock solid theoretical foundation", do you
>>mean that they are designed based on a theory that was inductively formed
>>and validated in a way similar to how physics was built up over the
>>centuries, with experiments and all? Or do you mean that this "rock solid"
>>foundation is someone's arbitrary, unproven hypotheses and theorems that
>>some mathematician dreamed up, and then deduced all of the RDBMS theory
>>from?
>>
>>If it's the second, then again I say: So what?
>
> It's sort of in-between, isn't it?
>
> A consistent, formal mechanism was created from well-established
> mathematical components. It does what it does. It is perhaps
> inductive intuition that it is an appropriate basis for managing data
> in the real world -- I do not believe this is the kind of thing that
> can be proven.

I don't see how this is answering my question. Again: So what?

> Even science is based on skepticism and best-theories
> rather than proof.

I don't really want to get into a debate about scientific methodology since
those just spiral into nothingness in this forum, so I'll just say that I
disagree with your philosophical viewpoint on it. And I will point out that
in science, a strong relationship must be established between theory and
experimental evidence in order to be able to declare it to be true in any
sense. I hope we can agree on that at least in some woozy terms.

So if you want to compare relational theory to scientific theory, you're
going to have to come up with not merely a self-consistent formal system,
but an argument that ties it to evidence. Formal != correct. There have
been lots of totally bogus formal systems invented. Just because some
crackpot writes a fantasy down in formal terms does not lend it any more
credibility. Just because it "works" does not mean it's right.

> So, the "rock solid" business is true enough,

I don't see how, unless you mean that the implementations run without
crashing.

> it's just not clear what
> weight it carries in the real world.

The issue is that some people seem to want to intimidate people into
thinking that relational systems are correct, without actually arguing
their case. It's a common ruse with today's pseudo-intellectuals: just
declare that your system is mathematical and formal etc., and state some
esoteric theorems, and maybe no one will notice that even though the system
is internally consistent, it hasn't in any sense been tied to a proof that
it's the right way to look at things. I don't know if that's what Alfredo
was attempting to do, but I've never seen a defender of relational do
anything other than this.

The next predictable thing that comes is usually not even a smidgeon of
evidence, but rather a character attack. This is all I have ever seen
regarding relational theory: either intimidation with fancy-sounding but
empty jargon, or ad hominems on those who point out that they haven't
actually proved anything. It's exactly like dark-ages religion: intimidate
and overwhelm the sheep, and burn the heretics.

Underneath it all, for all the theoretical work they claim to have done, an
RDBMS is merely a system among many that happens to work. It has not in any
way been shown to be more correct than a different type of database; it
merely has more dogmatic, pseudo-intellectual rationalizers on its side.

Shayne Wissler

Relevant Pages

  • Re: OT: The film Expelled: No Intelligence Allowed
    ... In saying that there is no god, ... theorems stating inherent limitations of all but the most trivial ... American mathematician and philosopher. ...
    (rec.arts.mystery)
  • Re: Writing a mathematical work more efficiently
    ... I would just write down a few Theorems, Lemmas, etc. that I have ... I then realized that this method of writing a mathematical ... I once asked a mathematician friend of mine what he does when he ... and prove them in a paper then begin writing them in LaTeX or does he ...
    (sci.math)
  • Re: A suggestion for infinity
    ... an infinite amount of possible theorems? ... series of symbols which can be interpreted by a competent mathematician ... If a mathematician did look over the monkeys' output and picked out a ...
    (talk.origins)
  • Re: Godels Theorem and Variations
    ... you see a more complete and familiar statement: The truth ... If this were a wff with a free variable, ... both the 1st and 2nd Theorems using simple Theory of Computation ...  Godel’s Theorem, proved by the German mathematician Kurt Godel in 1931, ...
    (sci.math)
  • Re: When a mathematician makes a mistake...
    ... Theorems are not evidence for axioms, ... Unsolicited bulk E-mail subject to legal action. ...
    (sci.math)