Re: [Slate] American Religious fervor, by the numbers

Koen Holtman (Koen.Holtman@cern.ch)
Mon, 5 Jul 1999 11:13:58 +0200 (MET DST)


On Fri, 2 Jul 1999, Gregory Alan Bolcer wrote:

> That's pretty funny. Of course, as you know, the Web
> scales where hypertext systems couldn't because of
> of inexactness, informality, and inconsistency. If
> it weren't all those agents and bots would be on
> unemployment.
>
> Just because something's inexact, doesn't mean you can't
> derive value from it. Although I agree, for a mathematics
> aficionado, exactness is a good property to observe.

Oh, I agree that inexactness has value. My complaint was not about the
use of language which allowed for inexactness. It was about your
restatement of Godel's results being *grossly inexact*. Your original
statement:

# Godel's theorem. Any system (or universe) can be proven either
# consistent or complete, but not both.

leaves out the fairly important 'sufficiently complex' between 'any'
and 'system'.

There are in fact many systems which can be proven both consistent and
complete. And the use of such systems plays an important role in web
content negotiation systems like RFC2295 and RFC2533. In Prolog, too.
There is a whole class of interesting stuff out there.

> Greg

Koen.