[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: Re: Silly indemnity FUD

• To: "'roger.day@g...'" <roger.day@g...>
• Subject: RE: Re: Silly indemnity FUD
• From: "Bullard, Claude L (Len)" <clbullar@i...>
• Date: Thu, 14 Aug 2003 09:08:55 -0500
• Cc: xml-dev@l...

One way to use a registry is to vette the code it 
references.  Discovery isn't necessary but that 
is one alternative.  It is tough to automate away 
the need to inspect and test, and unless one 
is willing to use boilerplate language at a 
very large scale, contracts always require 
human eyes.  Natural language is tricky.  If 
one gets into automated contracts, then RDF 
is a good bet.

I took up the problems of Golems (say agents 
with both vested authority and power) 
in a paper published 
by the MIT press two years ago.  That journal, 
unfortunately, isn't available.  The discovery 
challenge for authorized agents is the same 
as for the humans; negotiation of authorized 
terms where the owner of the agent must determine 
the negotiatble articles.  Hairy stuff.

len


From: roger.day@g... [mailto:roger.day@g...]

If, instead of matching bits on the wire, the, uh, fraternising object
models used a discovery mechanism which is good enough to encapsulate
"policy" as well as the aforementioned bits, then this discovery mechanism
might answer some of your questions. In other words, we invest the
discovery mechanism with some degree of in abstentia authority to discover
objects it trusts to use and it trusts to be used by. This "usage" is in
the legal and the mechanistic senses.

• Prev by Date: RE: Symbol Grounding and Running Code: Is XML Really Extensible?
• Next by Date: Re: XACML adding new functions
• Previous by thread: RE: Re: Silly indemnity FUD
• Next by thread: Adding standalone="yes" to XML document
• Index(es):
  • Date
  • Thread