## Puffers and Rakes

Jason Summers and I independently discovered a small pattern
that acts like a period-38 c/2 "chaos puffer" -- it repeats after
38 generations with a chaotic mess behind it. It is shown below in
two forms; the larger one is more robust to perturbations, but
nevertheless is eventually killed off by the growing chaos behind
it. The third pattern below is a larger almost-puffer found by
Jason, that also eventually gets killed off by its own debris.

We don't know how to make a tame p38 puffer from these patterns.
Fortunately, by perturbing the smoke very early, Jason was able to
make spaceships of other periods, such as
8, 10, and 20:

He also found another chaos puffer perturbation, with period
24.

The p20 spaceship is easy to turn into a simple puffer that
puffs 2 or 4 tubs:

It proved to be difficult to turn this into a rake (puffer that
emits only gliders), but after considerable noodling around, Jason
came up with this p40 example:

The rake about sends its gliders sideways, to the right. It
would also be useful to send gliders forwards and backwards. Period
40 is not large enough to build a forward rake. Using a 3c/6
spaceship, the rake can be thinned out to p120, allowing a forward
rake to be built. Jason thinks there's probably a better way to do
this; e.g. using p4 c/2 spaceships to turn the 2c/5 spaceships.

Period 80 would also work, but we don't know how to build a p80
sideways rake. Here's a way to build a chaotic "p80" puffer based
on de-phasing a p6 spaceship:

Now that we have rakes, it should also be possible to use them to
construct breeders. However, since these
rakes are unable to produce closely-spaced glider salvos directly,
one would either need to combine forward and back rakes to make
such salvos, or to use a gun synthesis
that allows the gliders to come in one-at-a-time.

B35/S236 -- Cellular Automata -- D.
Eppstein -- UCI Inf. & Comp. Sci.