Re: Does an agent have "live" on a world???

[glen e. p. ropella]

Hey Laszlo!  Good to hear from you.

Laszlo Gulyas writes:
 > > Whether or not (3) or (4) or (3) and (4) are sufficient for (2) is
 > > another question.  And I'm pretty sure the answer is "no."  I think
 > > the further requirement of (6) would be sufficient to ensure (2),
 > > though.  It's questionable as to whether or not the truth of (5)
 > > is sufficient for (2), because a method in an agent can be a null
 > > op.  Does that null op actually "affect the sim"?  It certainly 
 > > does at some level.... but, maybe not at a high enough level for
 > > anyone to care.
 > 
 > Well, I may be too theoretical but an internally motivated null op do
 > really effect the sim? (I mean it is possible that (6) is not sufficient
 > since that poor agent might send messages only to other agent's null op 
 > method.)

That's true.  I was applying the qualifier that *any* message
that gets sent will at least result in the cost of searching for
the method in the object's class, unless the object is a null
object.  But, even if the object is 0x0, there's still some 
possibility that the course of the sim could change if a message
is sent to it.

So, in essence, there is *some* level at which the sim is affected
by null ops.  But, that level could be negligible for modeling 
purposes, especially if the state of the model isn't changed by
the operation.

The reason I think that internally motivated messages to other
objects is sufficient to ensure the sending object's existence
in event space is because the sending of a message, in Swarm,
is an event, even if it's just a null message to another object
or a message to a null object.  When we implement event logging
in Swarm to keep track of what goes on in a Swarm model, we 
will need to log both the messages sent via the schedule and 
those sent directly between objects, in my opinion.  But, there
does lie they possibility of divying up these events into more
than just those two types.  We could identify "set" methods as
more interesting than "get" methods, for instance.  And at each
gradation of interesting-ness designated for types of events,
we introduce another level of "existence" or relevance to the 
model.

glen
-- 
{glen e. p. ropella <address@hidden> |  Send lawyers, guns, and money!  }
{Hive Drone, SFI Swarm Project         |            Hail Eris!            }
{http://www.trail.com/~gepr/home.html  |               =><=               }


                  ==================================
   Swarm-Modelling is for discussion of Simulation and Modelling techniques
   esp. using Swarm.  For list administration needs (esp. [un]subscribing),
   please send a message to <address@hidden> with "help" in the
   body of the message.
                  ==================================