[info] [croquet-user] Controlling Avatars with a Java Program

[Eugen Leitl]

----- Forwarded message from Les <hlhowell at pacbell.net> -----

From: Les <hlhowell at pacbell.net>
Date: Sun, 30 Sep 2007 22:37:39 -0700
To: croquet-user at duke.edu, jyoshimi <jeffyoshimi at yahoo.com>
Subject: Re: [croquet-user] Controlling Avatars with a Java Program
X-Mailer: Evolution 2.10.3 (2.10.3-4.fc7) 
Reply-To: croquet-user at duke.edu, Les <hlhowell at pacbell.net>

On Sun, 2007-09-30 at 20:47 -0700, jyoshimi wrote:
> Hello and kudos to those who made croquet.  It's quite nice.   I'm writing to
> ask how hard it would be to control avatars in Croquet with a separately
> running java program (in particular, this one: http://www.simbrain.net/). 
> The idea would be a user could run Simbrain, which is used to build neural
> networks (simulated brain circuits) and have a client window of croquet
> running with an avatar in it, and link the neural network to the avatar to
> control it.  As the avatar walked around the user could observe and
> manipulate its simulated brain.  If this wouldn't be too hard any pointers
> would be welcome.    
> 
> I did a few quick searches on the topic of java and croquet and did not find
> the answer, but apologies if I overlooked something obvious in the docs or
> forums.
> 
Well, Lets think about this a bit.  First of all have you already worked
with simbrain? It apparently can generate a creature of some sort in a
world  you can manipulate.  That would be the first thing to do.  Next,
what are you wanting to get the avatar to do?  The avatar interface is
made up of a viewscreen which is a two dimensional projection of a 3
dimensional world (4 if you count time).  It has ordered things about it
which are the objects in the island you are currently visiting and its
entire environmnet, but your interaction with that environment is via
the view screen, mouse and keyboard.  So on one level, the immediate
thought would be to use the screen for input (one very large neural net
to process what is being seen, maybe), And to have the IM bar scanned by
a second network for interaction with other denizens.  Finally your
neural net would have to generate mouse and or keyboard input to get the
avatar to interact in Second Life.  But to achieve anything approaching
real autonomous effects, the full neuronet would have to run inside
croquet.  I wouldn't want to tackle that by myself, but maybe you are up
to the challenge.  However my experience (which is quite limited)
suggests that with the number of inputs, the random interaction of
others, and the inability to deal directly with the environment would
make this a real test of your patience and programming skills.  Anything
in software can be done, if you can envision the right approach, can do
the coding, and can establish the working boundaries within which you
wish to work.  One problem is that in Croquet, you will not easily have
control over the environment if your project begins to "crawl the space"
of all the croquet worlds.  

	So first learn the simbrain system, then with some experience behind
you, design a creature in the world of simbrain, and see how complex it
can become.  If you succeed in reasonable fashion, then by all means go
for it.  I don't know what to tell you about the incompatibilities of
Java/C++ relative to Squeak or smalltalk, but lets just say they are
different languages used to achieve nearly the same thing. But the devil
as always is in the details.  Someone may give you more or better
information than this.

Regards,
Les H

----- End forwarded message -----
-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE