From: brucec@phoebus.labs.tek.com (Bruce Cohen)
Subject: Re: APPS: Game Design
Date: 15 Mar 92 23:21:26 GMT
Organization: Software Technology Research Laboratory, Tektronix Inc.




In article <1992Mar15.200525.12243@cpsc.ucalgary.ca> sharp@cpsc.ucalgary.ca 
(Maurice Sharp) writes:

> ... stuff about probability spaces deleted ...
>
> In other words, the addition of participants only adds complexity
> inasmuch as each participant has their own explanation of what is
> going on. However, these views are not critical to directing the
> action. It can be done by some observer using their own explanation of
> what is going on. The participants will orient to the new situation of
> action. Or if they do not, there will be detectable inconsistancies
> that can be used to change the action to something that works.

I think we're talking about two domains of action here, that we might
call micro and macro.  By macro domain I mean the regime of action and
response between actors (that is, both human participants and non-human
agents, what role-playing gamers sometimes call Non-Player Characters
or NPC's).  Dialog between participants is in the macro domain, for
instance.  The micro domain is the regime of sensory interaction and
feedback.  For instance, my hitting a hand on a wall in cyberspace is
probably an event at the micro level, since there has to be feedback to
tell me of the collision (perhaps a "thunk", and vibration against my
hand from a tactile-feedback glove).  But because there doesn't need to
be a response from another actor, the event probably is not significant
at the macro level, at least by itself.

Now this distinction is important because the regimes have different
time scales:

    micro:  visual feedback - 50-100 milliseconds
            auditory    "   - 100 milliseconds
            tactile     "   - ??? (I'd guess < 200 milliseconds)
            kinesthetic "   - < 100 milliseconds

    macro:  conversational response - 5-10 seconds
            attentional response    - 1-2 seconds ("uh-huh")

It seems to me that the problem of inconsistencies can be handled at the
macro level, as Maurice suggests, by having the participants note them
and modify their actions, after all, that's what people do in
conversations and other improvisational situations.  What concerns me is
the micro level, because I don't think that strategy will work there.
First, the sub-cognitive processes people use in that regime don't
handle large changes in goal very well.  The changes usually have to get
bucked up to full cognitive processes to examine the need for goal
modification, and these higher-level processes are a lot slower.
Second, even if the lower-level processes can handle the changes, I
suspect that changing the action will involve several rounds of
negotiation among the participants before everyone settles down to the
new behavior; even at micro-level speeds, this will take time which will
intrude on the progress and the versimilitude of the simulation.

--
"The end cause ... is too often handed off as an afterthought to harried
interface designers who follow programmers around with virtual brooms
and pails." - Brenda Laurel in "Computers as Theatre"
------------------------------------------------------------------------
Speaker-to-managers, aka
Bruce Cohen, Software Technology Research Lab   email: brucec@strl.labs.tek.com
Tektronix Laboratories, Tektronix, Inc.         phone: (503)627-5241
M/S 50-662, P.O. Box 500, Beaverton, OR  97077