Inquiries about Conway's game of life
Jesse Thompson
I'm wondering a few things about Conway's Game of life.
Firstly: is there a FAQ on the subject? I saw a post in this group
from 1994 pointing to a faq about CA in general.. the link redirects
to www.alife.org now and a FAQ is no longer apparent there.
Secondly:
Excuse me as I tend to think of Clife in rather high-level terms :)
but I am aware of numerous forms of still life, and numerous forms of
glider eaters.. I wonder if anybody has had success in Clife building
a fence?
By "fence" I mean a structure designed to survive the encounter of
simple spaceships from as many different directions as it can.
For instance, eaters can eat spaceships that hit them from very
predetermined directions. If a spaceship comes from another direction,
or just one cell offset, the eater is either destroyed (ow) or starts
a fire (yikes!)
I'm pondering the idea of a competetive life simulation, where
opponents build machines in specific areas on the board with the
purpose of destroying each other's machines first. Such a competition
should encourage evolutionary forces and greater emergence in a game
like life.
However, such a competition is overly aggressive without some form of
insulator or fence that can absorb or resist sparks. Any fire can
apparently generate sparks, and any savvy marksman can make a million
different clever guns, but there seems little defense against such
tactics. Most constructions in life are monsterously complex and
outlandishly fragile: one rogue spark will destroy almost every
construction I have ever seen regardless of size. It's like Conway's
life emulates nuclear fission: send one neutron in and it splits an
atom (starts a fire) which looses a half a dozen more neutrons and in
short order armageddon ensues.
So, what say ye? Any interest in constructing shapes to absorb
space-dust? Conway firewall anyone? :)
Thanks for listening
- - Jesse
Tim Tyler
: Firstly: is there a FAQ on the subject? I saw a post in this group
: from 1994 pointing to a faq about CA in general.. the link redirects
: to www.alife.org now and a FAQ is no longer apparent there.
http://cafaq.com/lifefaq/ now hosts the material that used to be there.
: Excuse me as I tend to think of Clife in rather high-level terms :)
As you suggest, the GOL is - by nature - highly inflammable. No matter
how you arrange things if they are made of "life stuff" they are much more
likely to burst into flames on impact than perform any useful function
as a barrier.
The most resiliant structures I am aware of are rows of eaters - but
they only defend against a few types of collision - and are thus
themselves likely to start fires - and may well have a net negative
effect overall - by transmitting and fuelling fires.
Basically the GOL is not suitable for this sort of thing. Other automata
allow construction of more impenetrable walls - or allow the existence of
more rapid self-replicating creatures - so if someone destroys the parent
it doesn't matter so much.
--
__________
|im |yler http://timtyler.org/ t...@tt1.org
Ilmari Karonen
>
> By "fence" I mean a structure designed to survive the encounter of
> simple spaceships from as many different directions as it can.
>
> For instance, eaters can eat spaceships that hit them from very
> predetermined directions. If a spaceship comes from another direction,
> or just one cell offset, the eater is either destroyed (ow) or starts
> a fire (yikes!)
While I agree with Tim Tyler that Life is probably not very well suited
for this...
If I had to construct such a fence, I'd build it out of more or less
random debris. I'd make it sparse enough for fires not to propagate but
broad enough to catch most incoming things. I'd be relying on the fact
that the fragility of Life constructs also extends to spaceships, so
that any random collision is likely to destroy them. The collisions
would undoubtedly also erode the "fence", but hopefully they'd produce
enough additional debris to make up for it on average.
Another idea, perhaps even better, would be to follow the old rule: The
best defense is a good offense. That would mean building rake guns or
glider gun arrays to construct moving walls.
--
Ilmari Karonen
(Trying out a new posting address, let's see if this works...)
If replying by e-mail, substitute .net for .invalid in address.
Tim Tyler
: If I had to construct such a fence, I'd build it out of more or less
: random debris. I'd make it sparse enough for fires not to propagate but
: broad enough to catch most incoming things. I'd be relying on the fact
: that the fragility of Life constructs also extends to spaceships, so
: that any random collision is likely to destroy them. The collisions
: would undoubtedly also erode the "fence", but hopefully they'd produce
: enough additional debris to make up for it on average.
This seems like a good idea for a fence.
It may be that replicating GOL creatures might be better off
reproducing rapidly than putting much energy into building walls, though.
: Another idea, perhaps even better, would be to follow the old rule: The
: best defense is a good offense. That would mean building rake guns or
: glider gun arrays to construct moving walls.
...but I'm not sure how well this would work. I reckon that often,
the flotillas would eventually hit something, the incoming continuous
stream of gliders would keep the flames active, and the fires would
expand back along the glider streams until they consumed the guns.
Jesse Thompson
the lexicon it points to, very fascinating. :)
Let me also randomly plug that Life32 is about the most bitchin
life-related thing I've ever seen, followed closely by Alan Hensel's
Java applet, and Paul Calahan's still life generator. I haven't played
with any more sophisticated pattern-searching utilities yet, but I'm
likely to :)
So far as fences: rows of eaters and buffers of debris have been the
best ideas I've had so far as well. After I write this I'm going to
aim a couple of P30 guns at one another through the holes in a couple
of debris feilds, each trying to punch through the other's sheild.
What I did that lead me to begin to think about competing life
constructs was that I was watching a P30 glider gun ratatatting away
one day, and felt like trying to interupt the stream of gliders.
I threw some garbage in the way. In my firt few attempts, the lead
glider kamakazi'd the garbage and the train continued. After a bit a
glider caught fire, and the stream sat there feeding a fire for a
short time, which eventually died out. Then in a hundred generations
or so, the stream cleared away the garbage bit by bit until it snaked
through a narrow passage.
Just as I was beginning to think "jeepers, this stream sure is an
influential event!" I threw another peice of garbage in the way, the
stream fed a fire, and then a rogue spark climbed inexhorably back up
the line and demolished the engine.
Now, in my book, this is Drama! :)
I am heartened also by Ilmari's suggestion that using vast quantities
of space as a rescource could be a key. A dense debris field, say the
natural remnants of a fire, catches fire again pretty easily. Widely
scattered debris survives easier I think. Whenever I destroy complex
things and study the aftermath, the debris never extends more than a
few dozen cells from what origionally burned, on average. sometimes
more and sometimes frightfully more, just like real fire, but this is
something you could bank on the odds of. My game idea wouldn't require
machines that could survive forever, just outlast their opposition
often enough to garner respect.
And, on the other hand, self-replicating machines could prove a nifty
strategy as well. This already sounds like competetive alife :)
----
Here are the rules I am throwing around for a competition possibility.
Lets say each player is allocated a square space on a rather large
life board. Let's also say the board has dead boundries (not
toroidial, not indefinite) sufficiently far away, just to help prevent
the game processor from being overtaxed in sandbagging techniques and
to focus the defensive energies of each team. I think 2 players at a
time would be a fine start, and they should be seperated from one
another by quite a hefty buffer: probably at least 10x as distant as
the size of their allocated home-spaces. I might even place them
diagonally from one another, like a poor man's chinese checkers.
Each player would draw or copy/paste their machine(s) into their
home-area, their constructions would have to fit the space allowed.
There could optionally be a limit on number of live cells employed.
Additionally, each player would designate one live cell on their board
as "the flag". The goal of a player is to turn off his opponent's
"flag" cell which undoubtably resides within a heavily guarded stator.
This regicide approach does hamper the idea of a "nomadic" or "run
away, run away!" space-ship oriented strategy, but it's also very
simple to implement and test for. If you cannot defend your stator you
lose. Your opponent wins by, probably, trying to demolish whatever is
in your home area. If he sets your castle aflame then, yeah, in all
probability your flag will get toggled.
In the simplest rules I've thought of, each player can't see what the
other player is unleashing, and then the two fortresses duke it out on
their own. That's pretty simple to implement and kind of hard to play.
Other variants I've considered include each player getting to call up
to X "timeouts" where the game pauses, and during each timeout players
can't see the interior of their opponents area, and each player is
allowed to continue construction (pasting, drawing, etc) in their area
only, for up to a specified time limit. They cannot move their flag
and it must remain on when their done drawing. (or else they'd lose in
a flourish of stupidity ;)
I think that sums up some rules for what could be a promising
competetive game of life. Lemme know what you think, and thanks again
for your comments :)
- - Jesse
Tim Tyler <t...@tt1.org> wrote in message news:<HJvnn...@bath.ac.uk>...