Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!ogicse!milton!brucec%phoebus.labs.tek.com@RELAY.CS.NET
From: brucec%phoebus.labs.tek.com@RELAY.CS.NET (Bruce Cohen;;50-662;LP=A;)
Newsgroups: sci.virtual-worlds
Subject: Re: Toward a Typology/Topology of Virtual Worlds
Message-ID: <12225@milton.u.washington.edu>
Date: 3 Dec 90 18:25:50 GMT
References: <12157@milton.u.washington.edu>
Sender: hlab@milton.u.washington.edu
Organization: Tektronix Inc.
Lines: 45
Approved: hitl@hardy.u.washington.edu



In article <12157@milton.u.washington.edu> PLai@cup.portal.com writes:
>...  Thus the cyberspace
> generated by the ai is a computational mapping of how objects are related
> in space, and associated textually.  A few general principles stand out:
> 
> 1. That the objects have only a spatial meaning in terms of other
>    objects - i.e. a thumb is a inch away from the palm
> 2. Objects are spatially related by transforms, not through coordinate 
>    references.
> 3. Objects are chunked into groups at varying levels, each group able
>    to have spatial transforms to other groups.

Umm ... I'd be careful about making coordinate transformations your primary
relationship.  That's the way it's done in mechanical CAD and related
fields, and I can tell you from bitter personal experience that it is very
difficult to get data structures designed to handle coordinates to deal
with things like forces applied to a moving linkage, or constraints
propagating back through a a set of connections.  A good example of this
problem is animating a human or animal body: you want to describe the body
based on the skeleton and its joints, and control the way it interacts with
the body so that its feet don't go through the floor when it walks.  It's
almost impossible to get realistic motion without addressing these issues.

The software used for this sort of animation is quite different from that
used to simply move a hierarchicly-related group of spatial transforms
around.

> 4. The layout and design of the cyberspace is done for benefit of the ai,
>    for the human to see the ai's spatial reasoning requires another layer
>    of logic.

Yes, that's clear, but I suspect that you will want to include that extra
level, if only for testing.  The case is similar to that of expert systems,
where the designer often builds in an explanation facility even of the user
will not be using it, just so that the designer can convince herself that
the ai understands the test environment in the expected way.

--
------------------------------------------------------------------------
Speaker-to-managers, aka
Bruce Cohen, Computer Research Lab        email: brucec@tekchips.labs.tek.com
Tektronix Laboratories, Tektronix, Inc.                phone: (503)627-5241
M/S 50-662, P.O. Box 500, Beaverton, OR  97077


Brought to you by Super Global Mega Corp .com