One very recent example is the p16 puffer engine found by Tim Coe. The following mail digest contains a number of new patterns not archived elsewhere. These letters illustrate the speed at which such a discovery can be applied to new designs, given the tools and experience that have developed over the years. Images appearing on this page are links to patterns that can be downloaded and viewed using a Life simulator. If your browser supports Java, then an animator will be invoked as well.
I was looking through the output of one of my flipper searches and I saw a ship that looked liked it might be robust with respect to some amputation. I ended up with the following period 16 c/2 puffer engine that by itself leaves no exhaust. It is quite robust to perturbation and I have included some amusing puffers that I have managed to construct.
Have a lively day,
-Tim Coe coe@vitsemi.com
A period 16 c/2 puffer engine that by itself leaves no exhaust.
A period 16 c/2 block puffer. Omitting the top spaceship results in a period 16 blinker puffer.
A period 16 c/2 tub puffer.
A period 128 puffer that emits a backwards going spaceship and assorted other junk. The lower left spaceship is there to perturb the exhaust in such a manner as to allow the generated spaceships to escape.
The new engine seemed ideally suited to an automated collision enumeration (being so small and clean). The enumeration (interactions between engines and spaceships) turned up a smaller block puffer:
The enumeration also found a blinker puffer (I haven't checked if it's identical to the one mentioned by Tim Coe). I don't trust my choice of parameters, and I'm going to try another enumeration, but for now I have 14 results total. Many are "dirty" puffers. Some are simply perturbations of the spark that do not actually result in a puffer. I put all the results together into the following pattern. They don't destroy each other but they do interact, so it will be necessary to cut and paste to examine each other separately:
I noticed that two of the interactions with an MWSS result in identical exhaust. However, the placements themselves are slightly different. I haven't done any "human" filtering of this file. It's just the raw results.
Another promising idea may be to react mirror images of the new engine against each other.
--Paul
Here are two more clean puffers based on the new engine. These were found by enumerating reactions of spaceships and dirty puffers composed of the engine and another spaceship. I viewed all non-self-destructive combinations in Xlife, and picked these out by eye.
I haven't found a rake (i.e. glider puffer) yet.
--Paul
I knew my collision enumerator would come in handy again one of these days (I haven't found much with it in months before today).
Anyway, here is a period 16 rake using Tim Coe's engine. Gliders propagate backward with respect to rake propagation.
I haven't found a forward rake, but the enumeration is still going, and I'll post if one shows up.
--Paul
I went over all the different puffers that came up in the enumeration and selected those that seemed well-behaved. Besides p16 puffers, there are p32 puffers for blocks and tubs, and a p48 blinker puffer. Here's the full set:
Besides puffers, there were lots of spaceships. Two had what seemed to be interesting sparks. The spark on this one is really big and chaotic:
The next one just seemed to have kind of a pretty spark. I've caught it sparking in the following picture:
--Paul
Two of the dirty puffers I got with my original enumeration produced forward gliders with period 32. I tried adding another spaceship and managed to stabilize the exhaust down to a block, but not eliminate it entirely. Finally, I added another HWSS to get rid of the block, though I was originally hoping I wouldn't have to. Anyway, here is the resulting p32 forward rake:
I don't know if there is a better solution than this.
--Paul
>I don't know if there is a better solution than this.
Neato! The only improvement I see is that the trailing HWSS could just as well be a MWSS and still delete the blocks cleanly.
NDE
Paul> Besides puffers, there were lots of spaceships.
Any o or o or oo or o o o o o o o sparks? Or do we only wantsingle-engine rakes? E.g., if you're *desperate*,
#C Synchronous P32 breeder based on Tim Coe's P16 engine: #C Breeder makes new puffer every 32 generations; #C Puffers make new blinker every 32 generations.
-- Mark D. Niemiec
Sorry for delays in responding to mail,
To Bob Wainwright:
To Dave Buckingham:
P256gun is extremely elegant (and not obvious; even after seeing the original,
I spent two days trying to re-construct part of it from memory, and was not
able to stumble across the correct placement of even one of the blocks!)
To Tim Coe: Congratulations on the new spaceship!
This is so simple (and the large part is similar to other
almost-spaceships which so often arise naturally), that it is amazing
that nobody found this one before.
Some minor additions to the new P16 spaceship:
A simple 2-part glider synthesis (eventually needing 9 gliders):
This requires four gliders from the right side, making it suitable
for construction by puffers. If the MWSS is made first and the
LWSS is inserted instead, it requires one glider from each direction,
making it more suitable for gun-based construction.
The spaceship can be widened by one (although I can't think of
how this would be useful, since it doesn't provide any extra
sparks, and is more difficult to construct and get close to
with catalyzing spaceships):
The spaceship can be used as an improved HWSS for sparking other
puffers, since it has a P4 side-spark, but only a P16 tail-spark,
making it much less intrusive.
This simple blinker puffer works catalyzed by the P16 (and works in
3 out of the 4 possible phases), but eats a HWSS employed for the
same purpose. There are probably lots of other almost-ran puffers
which have been overlooked because they attack an HWSS, but could
work successfully with the P16.
Here are some more P32 mechanisms. The first is a clean
P32 spaceship with a plume over 100 pixels long:
-- Mark D. Niemiec
I am in awe at how quickly so many neat things
have been created using this new P16 engine.
Great stuff Paul, Bill, and Mark.
Live long and prosper,
-Tim Coe coe@vitsemi.com
A period 32 LWSS backrake. The P32 forward glider rake prevents period multiplying exhaust reactions. The P32 sparker's purpose is to get rid of a pesky beehive that I could not eradicate with spaceships.
This is the period 4 flipper ship that I cut up to create the P16 puffer engine. This illustrates how two engines can share a stabilizing spaceship and be either 60 or 180 degrees out of phase.
[...] I ended
up with the following period 16 c/2 puffer engine that
by itself leaves no exhaust.
Nice work.
It is quite robust
I'll say
to perturbation and I have included some amusing puffers
that I have managed to construct.
[...]
#D A period 16 c/2 block puffer.
[...]
#D A period 16 c/2 tub puffer.
[...]
A period 24 hive puffer.
#D A period 16 c/2 block puffer.
[...]
#D A period 16 c/2 tub puffer.
[...]
#D A period 24 hive puffer.
A.k.a. p40:
The new p16 engine is real nice! I have impacted gliders into it in various
ways to see what results. Many different objects can be formed without
destroying the engine, including one of the rare still-lifes.
This is the most important result. The following shows a forwards glider
being converted into a LWSS which travels upwards:
The reaction just barely fails to be repeatable every 16 generations, sigh.
But it easily runs every 32 generations.
Here is a sideways LWSS rake based on this and the recent p32 forwards rake:
If a smaller forwards rake can be found, then this LWSS puffer can be
correspondingly improved.
Finally, here are some glider turning reactions that I found.
In these three, forwards gliders are reflected back the way they came.
The outgoing phase differs for each of these reactions.
Here is a glider turning reaction which turns a backwards traveling
glider 90 degrees still moving to the back. The rear spaceship is just
there to delete a loaf.
It is a shame that no trivial backwards to forwards glider turning reaction
exists with this engine. But that can be done using a few standard spaceships
anyway...
BCNU,
-dbell-
Here are more results concerning the new puffer engine...
Here is a relatively small p16 forwards glider rake. This is easily
built from the previous backwards p16 rake and a glider turning reaction:
Here are two reactions in which a LWSS is turned into a glider by the new
p16 engine. One of these creates a forwards glider and the other one creates
a backwards glider. (These might be useful because of my previously shown
glider to LWSS reaction.)
Here is a reaction in which the new p32 big spark turns a block into
a glider:
Finally here is a reaction in which the p32 big spark engine turns a block
into a glider and another block in the same column as the original one:
By using two copies of this reaction in succession, and adding an extra
spaceship to delete the extra glider, you can create a "readout" device,
where a row of blocks can be non-destructively scanned into corresponding
glider outputs. The blocks can be placed every 16 cells along the row.
The following shows the block pattern 1101 being read out:
This is pretty complicated for what it does, but was interesting to find!
Does anyone know of a simpler readout device which can non-destructively
turn a row of objects into gliders?
BCNU,
-dbell-
Here is a backwards p32 rake made from the new p16 engine:
If the rear beehive is deleted, then a different backwards glider is
produced from the other side of the puffer, as in the following rake:
Deleting the behive on demand with forwards gliders also works,
switching the output gliders from one side of the rake to the other:
As a recap, here are the best known forwards and backwards p16 and p32
rakes based on this engine all together in one picture:
BCNU,
-dbell-
Here is a rather large and ugly p32 sideways MWSS rake constructed from
the new p16 engine. It turned out to be amazingly difficult to find a
correct pre-block from any of the the puffer debris of a p16 engine placed
such that it could be hit by a glider to form a MWSS. There must be a
better solution this this!
Here is another clean big p32 spark that can be used to perturb stuff:
BCNU,
-dbell-
Date: Sat, 28 Oct 1995 13:13:16 -0700
From: Mark D. Niemiec
Subject: P16 spaceship, etc.
> Some time ago, you indicated that a list of sparkers and eaters was
> available. Is this still true? If so, I would like a copy. Either list or
> both would be greatly appreciated.
I have recently downloaded the entire Life archives, and am working my way
through 4 years of back mail, fleshing out my collections of oscillators,
sparkers, eaters, etc. When done, I intend to post them, including any
other stuff I have kicking around which has never been sent to the list.
Date: Sat, 28 Oct 95 17:12:19 PDT
From: coe@vitsemi.com (Tim Coe)
Subject: p32 LWSS backrake
Date: Sat, 28 Oct 1995 23:26-0700
From: Bill Gosper
Subject: Period 16? puffer engine
Date: Sun, 29 Oct 1995 01:56-0700
From: Bill Gosper
Subject: Period 16? puffer engine
From: "David I. Bell"
Subject: P32 sideways LWSS rake
Date: Sun, 29 Oct 1995 21:55:18 +1100
From: "David I. Bell"
Subject: More about the new p16 engine, and "readout" devices
Date: Mon, 30 Oct 1995 23:48:52 +1100
From: "David I. Bell"
Subject: P32 backwards rake from Tim Coe's engine
Date: Thu, 09 Nov 1995 21:23:20 +1100
From: "David I. Bell"
Subject: Some more p16 engine results
Date: Sat, 11 Nov 1995 15:32:16 +1100
Back to Paul's Page of Conway's Life Miscellany