Re: Pizza Example

From: Eric Kaun <>
Date: Fri, 23 Apr 2004 18:13:51 GMT
Message-ID: <z1dic.11395$>

"Anthony W. Youngman" <> wrote in message
> I've actually managed to work out a formal statement of what I always
> suspected.
> "Normalisation by non-loss decomposition" is impossible. The law of
> entropy says so.

Does entropy apply to information?

> So, to rephrase your last sentence, "By extracting the data from the
> real world and storing it in ANY database you are changing it in some
> sense, and losing data IS INEVITABLE."

This is true.

> The more you change it, the more you lose.

This is not. Or at least, you also gain. And I'd distinguish potential information ("raw data"?) from information structures. I don't prefer to have my system encode every bit of verbiage spouted by a potential user in requirements-gathering.

> And the more you decompose
> it, the closer you get to the point where "cause and effect" ceases to
> have meaning :-) (which means it gets more and more likely, that an
> attempt to get back to where you started will leave you somewhere else
> entirely :-)

That assumes you need to get back to where you started, which isn't a goal I necessarily have in mind. If I do, it's called a text document. My goal is not to reassemble a paper purchase order from data in a database - I usually need to present different views of the data, and do something interesting other than regurgitating it to the user verbatim.

Received on Fri Apr 23 2004 - 20:13:51 CEST

Original text of this message