Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!apple!xanadu!hibbert
From: hibbert@xanadu.com (Chris Hibbert)
Newsgroups: comp.object
Subject: Re: Re: Stash Collection (Economic models)
Summary: hire an agent
Keywords: agoric systems
Message-ID: <1990Nov30.184708.18687@xanadu.com>
Date: 30 Nov 90 18:47:08 GMT
References: <12080@life.ai.mit.edu> <-283749995@hpcupt1.cup.hp.com>
Organization: /etc/organization
Lines: 20

thomasw@hpcupt1.cup.hp.com (Thomas Wang) asks:
>> Is the market economy model applicable to computer programs? [...]
>> Human beings do this with relative ease.  But [...] programs would
>> have difficulty dealing with this sort of decision making.  

The simple answer is: read the articles by Miller & Drexler in the
collection by Huberman mentioned earlier in this thread.  The quicker
answer is that Carl Manning was right: most small agents will probably
form coalitions to have their resources negotiated as a bundle freeing
the individual objects from the overhead.  Other small objects that
[or whose owners] don't want to live or die with an unrelated cluster
will hire an agent that specializes in negotiating for resources.
Part of the contract would be that the agent would find a way to save
the object's state when the agent decides to stop spending the
object's money on computrons.

Miller and Drexler can be reached as miller@xanadu.com and
drexler@xanadu.com, respectively

Chris