# Re: Pizza Example

Date: Fri, 23 Apr 2004 08:41:38 -0500

Message-ID: <c6b6f5$nkd$1_at_news.netins.net>

"Tony" <andrewst_at_onetel.net.uk> wrote in message
news:c0e3f26e.0404230129.7e2d756e_at_posting.google.com...

*> "Anthony W. Youngman" <wol_at_thewolery.demon.co.uk> wrote in message
*

news:<kaXeN8ESmwhAFw3Z_at_thewolery.demon.co.uk>...

> > In message <c0e3f26e.0404191515.429e6820_at_posting.google.com>, Tony

*> > <andrewst_at_onetel.net.uk> writes
**> > >> Yes, you CAN try to analyse and store that information, but you
***CANNOT
**

*> > >> do it in Science - entropy is a one-way-street. If you treat a
*

database

*> > >> as an exercise in Pure Maths, I might agree with you, but as soon as
*

you

*> > >> drag the real world into it (and if you don't, what's the *point* of
*

a

*> > >> database), you have to deal with entropy.
**> > >
**> > >Since nothing is lost by the decomposition process (by definition),
**> > >entropy is about as relevant as Schrodinger's Cat and all your other
**> > >pop-science book obsessions.
**> >
*

> > Let's decompose the real world into its fundamental particles. Ooops ...

*> > we've suddenly lost cause-and-effect !
**> >
**> > There is a limit to lossless decomposition - quantum mechanics says so.
**> > Okay, a GUT or TOE might revise our opinion of this, but that's beyond
**> > current knowledge.
**>
**> Quantum mechanics concerns properties of matter, not data. It is one
**> big, fat irrelevance to data modelling. In any case, normalisation of
**> data never "splits the atom" (i.e. attribute values). This really is
**> becoming a futile discussion!
*

If I understand the misunderstanding, it might be something like this ... Wol is giving analogies to try to show that when doing "database theory" you don't want to just sit inside a mathematical model and "do the theory" because you are modeling something beyond that theory. If you are not constantly aware of the actual problem statement and working to show that your model really does model the problem and isn't just a process of playing with mathematics, then it is not useful. Because the theory is just a metaphor, it could very well be missing some key aspects needed to do the best job of solving the problem.

While I'm sure there are folks who don't care what the theory is used for or what it is useful for, there are others who don't want to do theory in the abstract without ensuring it is useful. The (awful) example I've used is that if folks are working on theories related to seriel killers and they don't actually lead to finding those killers, then I'm not too interested in THAT theory.

and on that note -- cheers! --dawn Received on Fri Apr 23 2004 - 15:41:38 CEST