W3C home > Mailing lists > Public > www-ws-arch@w3.org > August 2002

Re: Choreography and the Semantic Web

From: Geoff Arnold <Geoff.Arnold@Sun.COM>
Date: Sat, 10 Aug 2002 22:57:03 -0400
To: www-ws-arch@w3.org
Message-id: <0374F022-ACD6-11D6-BECC-000393991304@sun.com>


On Saturday, August 10, 2002, at 10:40  PM, Champion, Mike wrote:
> I think this demonstrates my point.  The proposals we're 
> discussing seem to
> be written by people trying to do with with today's technology, 
> hard-coding
> syntax onto semantics. They can point to a large body of 
> experience that
> does similar things, albeit with less interoperability or 
> scalability than
> the Web technologies have exhibited.  It would seem that those who 
> say that
> it should be done with RDF, DAML-S, OWL, or whatever are looking to the
> future.  Maybe it's a gap in my education, but I don't know of existing
> large-scale commercial systems that use agents that handle the partial
> understanding problem, or employ RDF-based technologies in the way 
> being
> described here.

No, it's not a gap in education. The problem here is how to 
kick-start the
network effect (i.e. how to avoid being the only person with a 
photocopier).
The answer (in the agent community) seems to be, essentially, 
imagine how
an intelligent agent might communicate (we're good at doing that, 
because
we can use ourselves as prototypes ;-) and then use that style of 
communication
between our relatively dumb (typically FSM) agents. That way, when the
first slightly-less-dumb agents emerge they'll find someone to talk to
in their own language (even if the conversations are slightly stilted).

As a bonus, it turns out that putting a little effort into the 
communication
models can lead to a significant payoff in terms of expressiveness, 
even for dumb
agents.

> Is there some modus vivendi possible here? ... along the lines of a WSA
> framework that is rich to describe the *principles* of coordination,
> conversations, reliability in a useful way that is abstract enough 
> to be
> implemented with either a stack of special purpose schemas and 
> layers on top
> of SOAP, or with specific ontologies expressed in a general 
> purpose semantic
> language?

Absolutely. A corollary to this is that we must eschew appeals to 
"the argument
from incredulity" (the one that goes "I can't imagine why, or how, 
you'd ever
need X"). And as I mentioned elsewhere, no Procrusteam beds, please!


Geoff Arnold
Sun Microsystems Laboratories
Received on Saturday, 10 August 2002 22:57:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:04 GMT