RE: NPR, Godel, Semantic Web

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

The semantic web doesn't trip on godel or 
incompleteness.   It trips on authority. 

No one sensible has said that Prolog-style 
knowledge bases don't work.  They do.  They 
are just expensive to build (to make complete 
in the sense that all terms have definitions, 
and that all definitions that need to be closed 
are), to maintain (hard to keep any system 
of definitions bounded), and conflicts in 
authoritative definitions (you say tomato i say tomato) 
have to be resolved.  Then you get to the 
really expensive part:  the logic layer.  
I expect money to be made there.

The problem of the semantic web isn't 
mathematical; it is operational.  "And 
now what will you pay for the ginzu knife? 
But wait, there's more!!"   We will need a 
standard that says something about the QOS 
of the metadata and its handlers.  That is 
why we talked golems last year, just trying 
to get an understanding of the qualities 
we need for the software that supports this.

Really folks.  Rolling back ten years 
and grabbing yetAnotherNicheFrom the
history of compsci isn't a big NEXT 
STEP (hear the W3C trumpets, see 
the garlanded hero, the whisperer, 
the dogs, the children, the triumph). 
It is just another engineering task. 
Metadata systems have their uses.  They 
just cost, require domain expertise, 
and *a common system* for interoperation. 
Actually, that last bit is ALL I got 
from the Scientific American article: 
this is another standards effort to 
declare another common system to be 
tied to the web:  adding knowledge 
bases to hypermedia.  So, 1985?

The crockness of it isn't the doability; 
it is the need to do it now.  Good 
topic for research, good topic for 
discussion; perhaps not the initiative 
by which all other tasks before the 
W3C et al should be measured or circumscribed.

Len 
http://www.mp3.com/LenBullard

Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h


-----Original Message-----
From: Simon St.Laurent [mailto:simonstl@simonstl.com]
Sent: Monday, May 07, 2001 9:49 AM
To: Robert C. Lyons; xml-dev@lists.xml.org
Subject: RE: NPR, Godel, Semantic Web


At 10:31 AM 5/7/01 -0400, Robert C. Lyons wrote:
>For a simple explanation of Godel's Theorem, see
>
>   http://www.nadn.navy.mil/Users/math/meh/godel.html.

Thanks!

>Here's a site that describes a couple of "common but fallacious
>conclusions" that people make from the theorem:
>
>   http://www.santafe.edu/~shalizi/notebooks/godels-theorem.html

Since a lot of what the Semantic Web proposes to do is precisely "deduction 
from axioms", I suspect these claims don't fall into the "common but 
fallacious conclusions" area.

If anyone knows where they do fall, I'd love to hear it.