xml-dev - Re: RDF, the "semantic web",and the nadir of AI (was RE: Realistic propo

[uche.ogbuji@fourthought.com]

> BTW, if by semantic web they mean 

Egads!  I hope you don't mean the slightest implication that I speak
the W3C's words.  Look, I don't understand half what TBL says, and except for
occasional lapses, I avoid the term "semantic web" because I'm not quite
clear of what it, er, means.

It looks as if I picked an opportune time to finally catch up with
XML-DEV.

> "RDF is the key to a proven design pattern, in which we build portal and
> intranet-type Web applications by marshalling numerous XML snippets. It
> helps us build multi-dimensional structures of object relationships,
which
> are usually cumbersome and unmanageable using traditional database
designs."
> - Uche Ogbuji 

It will be instructive to the rest of the discussion to note that I
restrict my claims to "portals and intranet-type applications".  Note that
these are closed systems.  I shall defend my claim in that space.  If you
wish to migrate the discussion to the open Web, prepare to drop me off
because I don't claim to pack that long a water-stick.

> per the XML.COM article, then this inference engine approach 
> could be compared to OLAP systems.   I am not sure what the 
> inference engine approach buys us that walking a recordset 
> and comparing values to determine actions doesn't.

OLAP, Decision Support, Executive Info Systems: all code-words for taking
the (admittedly old and hoary) ideas of AI and trying to make it work
for the ludicrously structured world of relational databases.  Note
that I say "AI".  I'm not afraid to name names.  I, too am a reformed
"machine intelligence" bigot, and I know that there is not much of the
"I" to be had from "AI".  But like it or not, AI in its maturity really
stands for knowledge-representation and frame-based systems.  Hardly
intelligence, as many people here have pointed out, but quite effective
stuff when it doesn't over-promise.

> Full 
> multi-dimensional (cubes) and query systems are already supported 
> on internet dbs such as SQLServer2000.

Whatever.  Relational databases couldn't find another dimension if they
were made of super-strings.  Uh oh.  Here comes Ken North with a
Shillelagh: better practice my drunken style.

Seriously.  The market triumph of RDBMS is nothing more than a grafting of
astounding hubris on the really quite modest work of Codd and co.  You
can't impose a priori structure on even a subset of the real world, and,
you certainly can't impose a priori structure on business.  Period.

As another poster here mentioned, the semi-structured folk have been
howling this for years.  I'm arguing that they are right.  I'm also
arguing that RDF (warts and all) is probably the most market-viable means
of restoring balance now.  I do hope something better comes along, but in
the meantime, I as a consultant have been solving real business problems
with RDF.

> Enlightenment?  Multiple means exist.  Why use one or the 
> other particularly if one is still considered a failed 
> technology, eg, AI, Prolog, etc.  Yes, that is something 
> to compare to markup because one re-emerged and the other 
> hasn't quite.  Why? 

AI is a failed technology like Gerald Manley Hopkins was a failed
poet.  Nuff aid.

Hopefully someone will come up with a recognized name that drops both the
"A" and the "I", but until then, "AI" is an acceptable name for a
technology that has made and continues to make important contributions to
modern data processing.

And oh by the way, all "Object Oriented Development" ever did was kidnap
frame-based representation and shove it into a straightjacket.  With all
the money that ended up flushed down the OO sink, we probably _could_ have
had computer systems of something that might be called "intelligence".  I
don't claim any insight as to whether this would have provided any
benefit to businesses.


-- 
Uche Ogbuji                               Principal Consultant
uche.ogbuji@fourthought.com               +1 303 583 9900 x 101
Fourthought, Inc.                         http://Fourthought.com 
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python