Re: performatives and trust

From: "pat hayes" <phayes@ai.uwf.edu>

> >Well I, for one, agree with you, Lynn.  I have argued in other trains and
> >with the graph [2] that people (and automated agents) will experience of
the
> >Semantic Web\Cloud quite incoherently, because regardless of their
> >processing capabilities they will be always have encountered an
incomplete
> >and inconsistent set of statements.
>
> Just for the record, logic has no problems whatsoever with incomplete
> sets of statements.

... and how does logic do with inconsistent and contradictory statements ?

> The vision of the semantic web coming out of W3C doesnt sound at all
> like a 'cloud'.

To me the distinction between a cloud and a web is that a web contains
vancable (defrefrerncable) links.  A cloud for the most part does not.  Web
pages form a web because the hyperlinks for the most part work and all refer
to the same resource.  But the URI used for RDF will not form these same
kind of links; there could be thousands of RDF nodes contained on hudreds of
servers and local applications all identified with the same URI scribing
quite different descriptions.  All of these nodes are *not* accessable to
any given agent at any given moment (unlike web pages).  No, the Semantic
Web will not be at all like the WWW ... it will be a cloud .. not a web.

>And I think that where $$ are involved, the
> semantic web is likely to be extremely sensitive to inconsistencies.
>I don't see banking agents making million-dollar
> trades with their electronic heads lost in a cloud of contradictions.
> I think it going to be more like a kind of automated lawyer's
> convention, with agents checking everyone else's proofs down to the
> last jot and tittle, and not parting with a microcent until all the
> inconsistencies are completely ironed out, or at any rate until some
> broker agent has agreed, in secured and checked DAML (or
> XML-legal-rules, or whatever) to take the risk.

Well Im sure there will be some of that ... achieved, If i may repeat
myself, by isolating the memory of those agents who must function at a high
level of security and trust.  But I think that the majority of packets that
flow through the Semantic Cloud and into people's living rooms, will be of a
different nature entirely.  Just like the inventors of the printing press
could not have anticipated the content of today's books, I doubt that you or
I can anticipate today the content of the messages in the Semantic Cloud a
decade from now.

[2] http://robustai.net/mentography/CoherentExperience.gif

Seth Russell

Received on Thursday, 7 June 2001 00:39:26 UTC