Hex rule B2o/S2m34: 1997 write-up with viewable patterns

[pcallahan]

While going over this old write-up of mine, I realize that it is greatly improved if you can see the patterns in action, which has been possible in LifeViewer for some time.

I don't think anyone else has transcribed it for LifeViewer, so here goes. (I'm a little surprised to see the date here, because I had only been in Aberystwyth a little over a month at that point and was writing CGOL infinite-growth searches for Nick Gotts. I must have been very busy.)

Subject: Experiments with a somewhat "Life-like" hexagonal CA (long)

By Paul Callahan (3 Dec 97)

Introduction

Occasionally the question comes up of whether there is a 2-state CA that runs on a hexagonal grid and works something like Conway's Game of Life. I've never seen any detailed discussion of a particular rule, though I had heard rumors that there is one with gliders. I decided to look for such a rule, partly to satisfy my own curiosity and partly to test the generality of some software tools I've been working on. (After working on this for a while, I found out that some rules had been found by Robert Andreen with some similar properties).

After settling on one rule, which I'll describe below, I was able to duplicate many of the properties associated with Life, including the ones needed to embed digital logic. A number of growth-rate results and oscillator-construction methods should also transfer to this rule. In fact, I was surprised at the variety of interesting patterns and reactions, some of which actually lead to simpler constructions than in Life. The rule is tamer than Life in that small patterns usually stabilize very fast.

I'm fairly certain that there is no totalistic rule on a hex grid that supports the variety of interesting reactions found in Life, because if there were it would be better known. In any case, I checked briefly and found nothing promising. In particular B3/S23 (birth on three live neighbors, survival on 2 or 3, like Life) settles down very fast and doesn't seem likely to have spaceships. Birth on 2 neighbors is useful, because an axis-parallel front of 3 neighbors cannot occur in a hex grid. However, always having birth on 2 neighbors seems likely to cause uncontrolled growth.

The Rule

With that in mind, I started playing around with rules based on different-shaped neighborhoods, but insisting on symmetry. Taking symmetries into account, there are three ways each of getting 2, 3, and 4 neighbors, for a total of 13 different neighborhoods. I tried a lot of variations, but the one I settled on only distinguishes between different ways of getting two neighbors. For these, one can use the organic chemists' naming convention for benzene-ring disubstitutions: ortho, meta, and para. Rich Schroeppel and Achim Flammenkamp suggested this, and it may be a useful mnemonic for some people. It has the disadvantage that there is no standard extension to subsets of 3 or 4 neighbors. Fortunately, we don't need this for the rule used here. So let's call these neighborhoods 2o, 2m, and 2p as follows:

Code: Select all

       O O            O .            O .
 2o:  . x .     2m:  . x O     2p:  . x .
       . .            . .            . O

Extending the birth/survival convention used above, the rule is B2o/S2m34.

That is, we always have survival but not birth on 3 or 4 neighbors. On two neighbors we have birth but not survival if the neighbors are adjacent (2o), and survival but not birth if there is one empty neighbor cell in between (2m). If they are opposite each other (2p), there is no birth or survival (which is also the case for 0, 1, 5, or 6 neighbors).

I settled on these rules in order to get a glider. It uses the only really small mechanism I'm aware of. I consider it crucial that gliders show up spontaneously from random starting states.

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2bob$3bo$o2bo$3bo$3bo!

The spark on the left appears every other generation. This is a period-4 glider with two distinct shapes, much like the glider in Life.

(Robert Andreen, and probably others as well, have looked at rules that support this glider.)

Variations

You can vary the rule for 2p. If you have birth but not survival, the same glider works (other factors made this variation less interesting to me). If you have birth and survival on 2p and death on 4 (this can be relaxed somewhat) you get two distinct sparky gliders that propagate at the same speed but higher periods (6 and 8 ).

The following discussion, however, will be limited to B2o/S2m34. This rule is interesting because it doesn't seem to grow explosively but it does have a number of natural high-period oscillators. It also has a fairly common (i.e. observable in nature) infinite-growth pattern.

Initial finds


Patterns usually settle down much faster than in Life. The kinds of things that remain are either little and isolated, or clumped into honey-comb patterns. The latter are very stable, and seem to act as anchors for oscillators (mostly low-period blinkers on the fringes, but interesting high-period sparkers as well). There are a number of reactions--similar to the "kickback" in Life--in which two gliders collide to produce a new glider on a different path. Unlike Life, one glider can also destroy another without being affected.

The first thing I really liked about this rule was the following period-48 spaceship (it flips after 24 moves and moves down one position).

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
obo4b$b2o4b$b3obob$2b2o2bo$3bo3b!

It's fairly easy to observe by trying random starting states, though not nearly as common as the glider. It's very sparky, and I subsequently found a number of glider reflections as well as a glider duplication reaction. I'll explain these in a later section.

Infinite growth is an easier question to answer here than in Life, because there is a stalk that leaves a trail of double honey-combs behind it. It shows up often enough to notice--maybe as often as the replicator in HighLife--but not so much as to dominate the dynamics of random patterns. This one is propagating down.

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
ob2o7b$b2obo6b$2ob2o6b$2b2obo5b$2bob2o5b$3b2obo4b$3bob2o4b$3b3obo3b$6b2o3b$
6bobo2b$6b3o2b$8b3o!

This structure (and others like it) are extremely stable. I can think of some intuitive reasons for why this is so, but I haven't considered the issue carefully. The practical implication is that unlike analogous Life patterns, there is no clear way to "ignite" the trail left behind by this pattern. So far, I have not found any kind of damage that propagates far along these stalks.

Note: gliders cannot propagate down (270 degrees) but are limited to 0, 60, 120, 180, 240 degrees. Clearly, this affects the kinds of constructions we can build out of spaceship/glider interactions.

Some natural oscillators and a period-15 spaceship


There were a lot of oscillators that appeared in random testing, so I ran a program to test 10x10 (rhombic) random starting states for new ones. I'll summarize all the periods later. One very useful early find has period 42. This is its smallest state.

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
b2o3b$bobo2b$4o2b$o3bob$b5o$3bo2b$3bo2b!

This can reflect gliders by 180 degrees and 120 degrees. Two can be combined to make a 60-degree reflector. Loops of recirculating glider streams can be built out of these.

High-period oscillators in this rule have a different feel from Life oscillators. One of the weirdest ones is period 66 and seems to extend and retract a tentacle (which is big and temporarily looks like a permanent, stable part of the oscillator).

Compact form:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2b2ob$3obob$obobob$b2ob2ob$bob2obobob$5b2ob3ob$7b2ob!

40 steps later with "tentacle" extended to the left.

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
4bob$2bo2bob2ob$3o3b2obob$2bo4bobob$2bo4bob2ob$3b2ob4obobob$5b2obob2ob3ob$
12b2ob$13bob!

There are many other natural oscillators, which I'll list later on. One especially nice one is this period-42:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2b2ob$b3ob$3bob$4ob$2ob!

Unlike the period-42 oscillator shown earlier, this one rotates 60 degrees every 7 steps. So far, its spark does not seem as versatile as that of the earlier one, but there is probably a use for it.

Finally, there is another spaceship that comes up in random searching, though it is much less common than the period-48. It's symmetric and moves in glider directions one position every 15 steps:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
bobo5b$2b3o4b$3obo4b$2bobo4b$3b2o4b$9b$4b2o3b$4bobo2b$3b3obob$6b3o$6bobo!

The one shown above moves to the left.

Larger constructions: A Glider Gun

Three of the period-42 oscillators can be combined to make a glider gun. I found this by first pairing up oscillators and testing the results. This didn't yield a glider gun, but it yielded a pair with a larger spark, to which I added the third:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
12b2ob$12bob$12bo4b2ob$2bo9bo5b2obob$b2obo10bo2bob2ob$2ob3o13b2ob2ob$
2bobobo13bobob$2b2ob4o3b2o6bob2ob$2bob2obo5b3o4b3obob$13bobo6bob$12b4ob$
12bo3bob$13b5ob$15bob$15bob!

It seems likely that almost all of the complex constructions that work in Life can be realized under these rules, the main difference being the relatively rarity of sustained explosions (certainly nothing on the order of the R-pentomino of Life, let alone the less common Methusalehs such as rabbits or the acorn). In fact, a very common result of a glider colliding with anything is to kill the glider without damaging the object it hits. For example, this very common 3-cell period-2 oscillator

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
bo$2ob!

acts as an eater much of the time. A honey-comb wall (as produced by the stalk) can also eat gliders.

Representation of Patterns

So far, pictures have sufficed to give an idea of this rule, but I hope at least a few people will try to run these patterns on their own CA programs. The format for electronic distribution is unfortunately less standardized than Life or other 2-state CAs on a cartesian grid. The obvious way to emulate a hexagonal grid is to use the Moore neighborhood omitting two opposite diagonal corners, which is what I did. But we have to agree on which corners. I chose the northeast and southwest corners for removal. So the embedding is given by the following
transformation:

Code: Select all

ABC                   A B C
DEF          ->      D E F
GHI                 G H I

The cell E has cells A, B, D, F, H, and I, but not C or G, as its neighbors. Following this convention, all the patterns I will post here can be run on a sufficiently general Moore-neighborhood CA simulator. In fact, the patterns above can all be converted back to a rectangular grid just by deleting the white space. The actual spacing in the cartesian grid is given by the dots. I will adhere to this convention for all pictures. For larger patterns, I will use run-length encoding as described in David Bell's recent posting. These patterns encode cells on a cartesian grid with the neighborhood defined above. At the risk of annoying people who are against posting code, but at least guaranteeing that anyone with a copy of this article can decode the patterns, here is a short perl program to convert these patterns to a list of cell coordinates:

[Of historical interest only. I stopped writing in Perl many moons ago.]

Code: Select all

#!/usr/local/bin/perl
$x=0; $y=0;

findpat:
while(<>) {
last findpat if (/x *= *[0-9]+ *, *y *= *[0-9]+/);
}

main:
while(<>) {

while ( /([0-9]*)([bo\$])(.*)/ ) {

if ($1 eq "") { $count = 1; }
else { $count = $1; }

if ($2 eq "b") { $x+=$count; }
elsif ($2 eq "\$") { $x=0; $y+=$count; }
else {
for ($i=0; $i<$count; $i++) {
print $x++, " ", $y, "\n";
}
}

$_ = $3;
}

last main if (/\!/);

}

Again, we have to agree on conventions. The cells are listed by horizontal (x), then vertical (y) coordinate. Vertical increases from top to bottom. For example, the cells A, C, and I above would be listed as "0 0", "2 0", and "0 2", respectively.

Glider Reflections

There are many ways of reflecting gliders both from stationary oscillators and moving spaceships. I won't try to list the 180-degree reflections exhaustively, since there are so many. I will include some of them in the larger patterns. So far, I have found one way of turning a glider 60 degrees and four ways of turning one 120 degrees. I haven't tried all the high-period oscillators, so there may be others as well. It's especially interesting that the period-48 spaceship can reflect gliders, because gliders move fast enough that they can circulate around a flotilla of moving objects in this way. Here are the three non-180-degree spaceship
reflections:

Glider is reflected 60 degrees, traveling up:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
14bob$14bob$14bo2bob$4o10bob$o2bo11bob$b3ob$b2ob!

Glider is reflected 120 degrees, traveling to the right:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2bob$ob$bob$2bob$3b2ob$b$b$b2ob$2bob$3bob$3bobob$4bobob$2b3obob$b4ob2ob$4bobob$
4bob!

Glider is reflected 120 degrees, traveling up:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
8bob$6bob$7bob$8bob$9b2ob$b$b$b$2ob$bob$2bob$2bob$2bob$b5ob$4bob$4bob!

In the above three examples, the spaceship is traveling down. Thus, the two distinct 120-degree reflections are not interchangeable.

Here are two more 120 degree reflections with stationary oscillators. In both cases, the reflected glider travels to the right.

Period-20:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2bob$ob$bob$2bob$3b2ob$b$b$b$b$4b3ob$6b2ob$6bobob$7b2ob$7bobob$8b2ob$8bob$9bob$
9bob$7b4ob$9bobob$10bob!

Period-42:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2bob$ob$bob$2bob$3b2ob$b$b$b$4bob$5b2ob$4b4ob$6b3ob$6bobob$7b4ob$7bo3bob$7b5ob$
10bob$11bob!

Finally, these can be used to send gliders in a loop. First, here's one based on period-42 reflectors:

Code: Select all

x = 89, y = 89, rule = B2o/S2m34H
29bo$30b2o$5bo22b3ob3o$4bobo23bobo2bo$5b4o21bob4o7bobo$4b2o3bo20b3obob
o5bob2obo$6b4o23b2o8b2ob2o$5b2obobo22bo9bobobo$7b5o32b3obo$10bo30bo3bo
b2o$38b4o3b3o$47bo$2b2o5bo19bo$2bob2o4bo19bo$b2obo5bo17bo2bo$2ob3o4bo
21bo$2bobobo3bo21bo$2b2ob2o$2bob2obo$5bo8$12b4o$16bo41bo11bo$61bo9bobo
$14bo46bo7b5o$61bo8bo3bo$60b2o9b4o$70b2obobo$72b3obo$75b2o$75bo2$72bo$
73bo$4b2o68bo$2b3ob3o66b2o$4bobobo72bo$3b2ob4o70b2obo$5b2obo71bob2o$5b
ob2o70b4ob2o$7bo72bobobo$12b2o66b3ob3o$14bo68b2o$15bo$16bo2$13bo$12b2o
$12bob3o$13bobob2o$14b4o9b2o$14bo3bo8bo$15b5o7bo46bo$15bobo9bo$18bo11b
o41bo$73b4o8$83bo$81bob2obo$82b2ob2o$56bo21bo3bobobo$56bo21bo4b3ob2o$
57bo2bo17bo5bob2o$58bo19bo4b2obo$59bo19bo5b2o$41bo$41b3o3b4o$40b2obo3b
o30bo$40bob3o32b5o$41bobobo9bo22bobob2o$41b2ob2o8b2o23b4o$41bob2obo5bo
bob3o20bo3b2o$43bobo7b4obo21b4o$53bo2bobo23bobo$54b3ob3o22bo$57b2o$59b
o!

Here's one that was considerably harder to set up, but shows that a glider can be sent in a loop around a moving spaceship flotilla:

Code: Select all

x = 270, y = 51, rule = B2o/S2m34H
17bo$17bobobo$17bo3b2o$20b2o$20bo$3bobo14b2o$3b4o13bob2o$4bo2bo12bo2bo
$5bobo15bo$5b4o$5bo2$o$obobo13bo$o3b2o12b3o$3b2o13bo2bo$3bo13b3ob2o$3b
2o15b2o$3bob2o14b2o$3bo2bo$6bo74bo$81bo$81bo2bo$81bo$82bo5$31b3o7b3o$
30b2o2bo5b3o2b2o219bo$31bo11bo222bobo$31b3o10bob2o220bo$31bobo10bo220b
2o$34bo232bobo$268b2o$268b2o$268bo5$12bo3$16b4o$15b2o$15b5o2$20bo$20bo
bo!

I found only two reflections by the period-15 spaceship, and both are 180-degrees. These are not as useful as they would be if the glider was traveling in the same direction as the spaceship. Maybe someone else can think of a use:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
bob$ob4ob$b2ob3o4bob$2bo3bo2bob$2bo3b2o2bob$11bob$3bo3b2o3b2ob$4bo3bob$4b2ob3ob$
4bob4ob$6bob!

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2bob$2b3ob$3ob2ob$2bo2b2ob$3bob2o5bob$10bob$4bob2o3bob$4bo2b2o3bob$3b3ob2o4b2ob$
6b3ob$7bob!

Glider duplication

While doing collision enumerations to find reflections, I stumbled across a very surprising reaction between a glider and a spaceship:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
ob$3bob$3bob$3bob$2b2ob$b$b$b2ob$b2ob$bobob$2o2bob$b6ob$3bobob$4bob!

What happens here is that the glider interferes with the spaceship without destroying either of them. The result is a large spark that is very similar to an escaping glider, but has an extra cell that soon causes it to vanish. I've seen nothing similar in Life, where it's very hard to get any effect out of a glider without destroying it in the process. In fact, there are several ways of placing a second spaceship so that the spark becomes an escaping glider, resulting in a glider duplication that leaves one glider undisturbed. With this, we can convert the moving glider loop into a rake, albeit a large one with a slow growth constant:

Code: Select all

x = 270, y = 51, rule = B2o/S2m34H
17bo$17bobobo$17bo3b2o$20b2o$20bo$3bobo14b2o$3b4o13bob2o$4bo2bo12bo2bo
$5bobo15bo$5b4o$5bo$50bo3bo$o50b2o2bo$obobo13bo33bobobo$o3b2o12b3o30b
3o2b2o$3b2o13bo2bo32b3o$3bo13b3ob2o$3b2o15b2o$3bob2o14b2o$3bo2bo$6bo
74bo$81bo$81bo2bo$81bo$60bo21bo$60bo5bo$62bo2bo$63b3o$63b2o$31b3o7b3o
19b3o$30b2o2bo5b3o2b2o16bo2bo199bo$31bo11bo21bob2o197bobo$31b3o10bob2o
20bo199bo$31bobo10bo220b2o$34bo232bobo$268b2o$268b2o$268bo5$12bo3$16b
4o$15b2o$15b5o2$20bo$20bobo!

Glider duplications also exist for some stationary oscillators. It's likely that they are quite common and would be an important part of the toolkit for anyone building designing large patterns in this rule.

Period-42:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
12bobob$4bo8b2obob$11b3ob2ob$8bo3bobobob$5b4o3b4obob$14bob3ob$17bob$2ob3ob$
ob6ob$bobob2ob$3ob2obob$3b2ob$4bob!

Period-20:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2bo7bob$5bo2b4ob$5bo3b2ob3ob$5bo5b2obob$bo2b2o7b2ob$obo12b2ob$3bo13b3ob$
bobo14bob$2bobo13b3ob$3b2o13bob$3bobob$4b2ob$6bob$6bob$7bobob$6b4ob$9bob$10bob!

Here's a period-210 gun built by inserting a p42 glider duplication into a period-210 glider relay consisting of two 180-degree reflectors:

Code: Select all

x = 69, y = 26, rule = B2o/S2m34H
49bo$50bo$48b5o$49bo3bo$50b4o$50bobobo$51bob3o$50b2obo$52bo10b2o$2bo
50b2o6b3ob3o$b2o45b4o2bo2bo4bobobo$2b2o44bo9bo4b3obo$3ob2o50bo2bo3bo2b
3o$bob4o53bobo2b2o$2bobobo53bob4o$2b2ob3o58bo$2bob2o$56bo$54bobo$53b4o
$54bobobo$55b6o$54b2o3bo$56b4o$56bobo$58bo!

The period-20 duplication can be used to make a compact high-period gun by sending a period-42 stream through it. The duplication only occurs once every 420 steps, and in other cases, the glider either passes through or gets destroyed without hurting the period-20 oscillators. Shown below is a period-840 gun that illustrates two other reactions as well, namely the 3-cell glider eater, used here to get rid of gliders that pass through, and a period-8 oscillator that thins period-4n streams down to period-8n streams:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
4bob$3bobob$4b4ob$5bob$5bob$6bob$4b3ob$5bobob$5b3o5b2ob$2o5b2o4bob$bo6bo14bob$
b2o18b3ob$2bo20bob$23b2ob$22b2obob$2bo18b2ob$2bo17b2obob$2bo2bo14bob3ob$
2bo18b2ob$3bo8b2o8bob$14bob$15bob$16bob$10b2o2bob$9b3ob$8b2obo11bo3b2ob$
8bob3o11b4obob$9bobo11b3obobob$9b2ob2o3b3o6b2ob2ob$9bob2o6b3o6b2obob$
11bo7bobo7bob$20b5ob$19b2o3bob$21b5ob$21bobob$24bob!

The glider to the left is the one that escapes. It is just below the period-8 filter. The one headed 60 degrees above horizontal passes through the duplication device, triggering the duplication.

Spaceship factories

Spaceships are a lot more interesting when they can be produced on demand. Just as in Life, there are ways to synthesize these from gliders. Here's a three-glider synthesis of the period-48 spaceship:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
18bob$19bob$19bobob$2o10b4o4bobob$2bo9bo8bob$3bob$4bo11bob$2bob!

There is also a two-glider collision that can be sparked in various ways to produce this spaceship. This results in a more compact spaceship factory:

Code: Select all

x = 36, y = 93, rule = B2o/S2m34H
10bo$11bo$11b4o$11bobo$14bo$14bo$15b3o$8bo6bobo$2bo5bo7b3o$3o6bo15bo$
2bo21b3o$b3o21bob3o$4b2o12bo7bobo$6b2o9bo7b3ob2o$5b2ob2obo14bob2o$7b2o
b2o14b3obo$7bobo2b2o14bo$9b3o$11bo2$30bo$17bo13b2o$5bo9bo13b3ob2o$4b5o
7bo6b4o2bobobo$5bobobo7bo5bo6b2ob3o$6b4o8b2o10bob2o$6bo3bo16bo$6b5o13b
2o$9bobo12bobo$10bo14b4o$25bo3bo$25b5o$28bo$29bo6$15b4o$19bo2$17bo22$b
o$obo3bo$b4ob2o$2bo2b2ob3o$2bo4b2obo$9b3o13b2o$11b2o11b2ob3o$22b3obobo
$23bo2b2ob2o$17bo5b4ob2o$8b2o7bo6b5o$8bo8bo2bo5bo$17bo$18bo$25b2o$26bo
bo$17b3o5bo3b2obo$16b2o13b2o$11bo3b2o3bo3bo4b3ob2o$11bo2b2ob3o4b2o2b2o
bobo$10b5ob2o8bo3b2ob2o$13bo11bob2obob2obo$14bo8b3obo5bo$24bobobo$25b
4o$25bo3bo$26b5o$28bo$29bo!

I haven't looked for a synthesis of the period-15 spaceship. Since it is symmetric, the most likely approach would be to try to synthesize half of it and fit the two halves together.

One use of these spaceship factories would be a "standing breeder" like the one built in Life by Dean Hickerson. That is, it should be possible to generate a set of spaceship streams through which a glider stream can be sent on a zig-zag path. If this path intersects a stream of glider duplication flotillas, then each such reaction point would be a source of linear growth, resulting in quadratic growth overall. The standing breeder could also be used to realize other growth rates, since it converts any f(n)-growth stream of gliders into an O(n f(n)) breeder. I haven't built such a pattern--so far, anyway. There is a simpler way to get quadratic growth, which I'll explain in the next section.

Quadratic Growth: The Stalk Breeder

The stalk is common enough that it seemed likely that there was a simple way to synthesize it. As with the period-48 spaceship, there are several. Here's one using 3 gliders:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
ob$3bob$3bob$3bob$2b2ob$b$b$b$b$b$5bob$4bobob$5bobob$7bob$8bo4bob$13bob$
13bo2bob$13bob$14bob!

Since the stalk is a linear growth pattern, a reasonable way to get quadratic growth would be to reproduce this collision again and again using a moving pattern. I've already presented a rake that should work for this purpose. The problem is not quite this simple, because one must make sure that the rake does not collide with the growing stalks, something that would almost certainly happen with any three-rake construction to reproduce the above synthesis. A different stalk synthesis turns out to be more useful for building a breeder:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
10bob$10bob$2bo8bo2bob$3b2o7bob$13bob$b$b$2ob$2bob$3bob$4bob$2bob!

This combines two gliders with the 3-cell period-2 oscillator that we've seen already as an eater. The synthesis is dirtier than the first, but eventually results in a stalk. The glider duplication reaction can be modified to produce the 3-cell eater instead of a glider, giving us a puffer for these objects. The glider entering from the right is not coming in a direction that we can produce with the rake given earlier, so we need to add two extra spaceships as reflectors.

That's all there is to this pattern, though it takes up a lot of space:

Code: Select all

x = 413, y = 328, rule = B2o/S2m34H
396b2o$397bo$398bo$398bobo$399bobo$397b3obo$396b4ob2o$399bobo$15bo383b
o2$15b2o$17bo$17bo2bo$16bobobo$16b3obo$16bobobo$16bobo$19bo6$391bo$
392b2o$394bo$393bobo$394bo31$2o$o$o$bobo$b5o$3bo$4bo22$273bo2$274b2o$
274bo$274bobo$274bobo$275b3o$276bobo$276bo23$131bo$131bobo$133bo$132b
2o$132bo42$214bo$214bo$206bo8bo2bo$207b2o7bo$217bo3$204b2o$206bo$207bo
$208bo$206bo4$373bo2$373b2o$375bo$374bobo$375bobo$375b3o$375bobo$378bo
11$372b3o$373bob2o12bo$372b2o2bo$373b4o12b2o$376bo14bo$377bo12bobo$
391bobo$391b3o$391bobo$230b4o160bo$231bo2bo$232b3o$219b2o13b2o$221bo$
222bo$220bobo$219b5o$222bo$222bo4$365b2o$366bo$367bo19b2o$367bo20bo$
367bo20bo$366b5o16bobo$369bo17bobo$369bo18bob3o$262bo125b2ob4o$390bobo
$240bo21b2o129bo$264bo$241b2o20bobo112bo$241bo22bobo110bobo$239bo2bo
21b3o112bo$240bobobo19bobo109b2o2bo$241bob3o21bo109b3o$242bobobo131bob
2o11bo$245bobo118bo11bo13b4o$245bo120bobo24bob2o$367bo25b4o$269bo97bo
2b2o25bo$269bo99b3o$266bo2bo98b2obo$267bo104bo$266b4o$268bobo$253b4o
12bobo2bo$253bo2bo18bo$254b3o19bo2bo97b2o$254b2o23bo98bo$278b4o97bo$
278bobo98bo$278bobo98bo$217bobo158b5o$218b2o161bo$216bob3o160bo$203b2o
11bo2b2o58bo$204b2o14bo$204bobo72b2o$204bo2b2o72bo$203b6o71bobo$205bob
o73bobo$207bo73b3o$281bobo$284bo$408b2o$410bo$411bo$409bobo$408b5o$
411bo$411bo2$266bo$267b2o$269bo$269bo$247b3o19bo$248bob2o15b3o$228b2o
17b2o2bo15bobo$230bo17b4o18bo$231bo19bo$229bobo20bo$228b5o$231bo$231bo
$248b2o$249bo$248bobo$247bo2bo2bo$240bo7bo2b4o$240bobo6bo2bo2bo$239bo
2bo10bo10b2o$239bo14bo10bo$239b4o5bo16bobo$240bo2bo7bo11bo2bo2bo$240bo
10bo11b4o2bo$251bo11bo2bo2bo$250b2o14bo$266bo7$264b3o$265bob2o$264b2o
2bo$265b4o$268bo$269bo9$250bo$251b2o$251b3o$250bo$251bo2bo$251b6o$254b
o$255bo!

There is probably a much smaller quadratic growth pattern. A more compact puffer would be a big step in this direction.

Other guns

The period-20 oscillator actually yields a compact period-20 gun. Two combined produce a spark that is almost a glider: all that is needed to fix it is the 3-cell eater:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
8b2ob$8bob$b$5b6ob$5bob$2bo5bob$b5ob$2bob2o8bob$3b2o8b3o4bob$2b2o9b2ob2o2bob$
3o12b2ob4ob$2bo14bo2bobob$b3o17bob$4bob!

If we use a period-25 spark instead of the 3-cell eater, we can increase the period to 100. Combining this with the period-8 filter, we get a small period-200 gun:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
5bob$6b3ob$7bob$7b3ob$3bo2b2ob$2b2ob2ob$obob2ob$b3ob2ob$4b2obo11bob$6b4o7b3ob$
8bo10bob$9bo9b2ob$18b2obob$17b2ob$16b2ob2ob$5bo8b3ob2ob$6bo9b3obob$6bobo7bo2bob$
7bobob$8bo5b2ob$7bo5b3ob$13b3ob$8bo5bobob$8b2o5b2ob$8bobo4bobob$15b3ob$18bob$
18bob$19bobob$18b4ob$21bob$22bob!

There is a period-43 oscillation that shows up occasionally, though it is not nearly as common as period-42 or period-20. Technically, it's part of a period-86 oscillator, because the patterns that realize this period have some period-2 cells. These can also be used to build a glider gun, which is interesting since it produces an odd-period (in fact prime-period) glider stream as output:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
11b2ob$2o11bobob$2bobo8b3ob$2b3o8bobob$2bobo9b4ob$3b3o8bo2bob$2b2o2bo7b5ob$
4b3o8b2ob$4b2obo8bob$4bob2ob$2bo2bobob$b3o2b2ob$o5bob$b$b$b$b$b$b$b$b$18bo2bob$
19b3ob$20bob3ob$20bobobobob$20b7ob$21bo2bob$25bob!

Finally, we can build high-period guns using glider/glider interactions. What's interesting here is that we can generate the stream "A and not B" without affecting the stream "B" because (unlike Life) there are lot of ways for a glider to delete another without being destroyed. To illustrate this, here's a gun I built before I found the glider duplication reaction:

Code: Select all

x = 86, y = 49, rule = B2o/S2m34H
o$o10bo$bo9bobo$9b5o$10bo3bo$11b4o$11bobobo$11b5o6bo$13bo7b2obo$20b2ob
2o$18bo3bobo$19bo2b2obo$18b3obob2o$20b2ob2o$24bo2$7bo$7bo18bo$8bo2bo
15b2o$9bo17b2o$10bo16bob2o$26b4obo$29bobo$29bob3o$29b3o$31bo$15bo9b2o$
15b2obo6bo$13b3ob4o4bo$15bobobo5bo20bo31b2o$14b2ob3o8bo16bobo31b2obo$
16b2obo26bobo30bob2o$16b2o2b2o26bo31b2ob2o$20b2o27bo31bobo$22bo58bob2o
$81b3obo$23bo14b2o43bo$23b3o12bo$22b2obo13bob2o$22bob5o10b3obo$23bobo
2bo4bo5bobobo$23b2obo2bo4b2o4b2ob2o$23bob3o4b3obo5b2obo$26bo4b2obobo5b
obo$33b4o$33bo3bo$33b5o$36bobo$36bo!

It works by sending a stream through a glider relay, which leaves a hole in the stream wherever it interacts with the glider in the relay. A second gun inverts this stream, resulting in a high period stream. The gun as such is made obsolete by direct glider duplication, but there are undoubtedly many uses for the one-sided glider deletion reaction. Among other things, it could be useful for building the sort of pseudorandom gun that has been realized in Life by Dean Hickerson and others.

Miscellaneous oscillators

Many different oscillators turned up in a random survey of 10x10 (rhombic) starting states on an otherwise empty infinite grid. Most of these consisted of an oscillating part stabilized by an anchor organized into honey-combs. Like the period-43 mentioned in the preceding section, these often have a low-period fringe that multiplies the total oscillation by 2, 3, or 5, without affecting the interesting part, which I call the principal oscillation.

I'm going to list oscillators here by principal oscillation. Sometimes it may be possible to eliminate the discrepancy between overall period and principal oscillation, but I haven't tried to. It is also likely that many of the stabilizing anchors below can be reduced in size. This is not an exhaustive list of known oscillators, and I've left out common low-period patterns that are easy to generate. Oscillators that have appeared already in this article will not be repeated.

Period 7:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
3bob$2b2ob$b2ob2obob$3bob3ob$ob2obo2b2ob$b2obo3bob$2ob4o2bob$2b2o3bobob$
2bobob2o2bob$2b3ob2ob3ob$5b2ob2ob$7b2ob!

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
ob$ob$b2ob$2b3ob$b2o2b2obob$bo4b3ob$2bo5bob$9bob$8b2ob$8bob$9bobob$8b4ob$11bob$
12bob!

Period 9:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2b2ob$3obob$bob2ob$b3obob$4b2ob$4bob$3b3ob$3bo2bob$2b3obob$2bobo2bob$3b2ob2ob$
5b2ob$6bob!

Period 11:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2bob$b2ob$2b2obob$3obobob$2bobo2b2ob$2b6ob3ob$2bo3b2o2bob$8b3ob$10bob$11bob!

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2obob$o2b2ob$2b2obob$3ob2ob$bob2obob$2b2ob2ob$4b2ob!

Period 14:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
b3ob$2obob$o3b2ob$2ob2ob$2b2ob2ob$2bob2ob$4bob!

Period 17:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
2bob$2bob$6ob$2bob2ob$3b2ob$b$3b2o2bob$2b2ob2ob$2bo2bobob$3b2ob2ob$5b2ob2ob$
7b2ob$7bob!

Period 18:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
bobob$2b3ob$b2obob$2ob2ob3obob$ob2obobob4ob$b2ob5ob2ob$3b2obob3obob$3bobo4bob!

Period 19:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
bo3bob$2o2b3obob$ob2obob3ob$b2ob5obob$bob2o2bob2ob$3bobob3ob2ob$7b2ob3ob$
8b3o2bob$10bob!

Period 22:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
bob$2bob$5ob$2bobob$2b5ob$3bobob2ob$6b4ob$6b2ob$b$10bob$11b2ob!

Period 26:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
bob$bobob$5ob$bobob2ob$bo2b2obob$3b2o2bob$5b5ob$7bobob$5b6ob$2b2obobobobob$
4bob2obob$8b3ob$8bo2bob$11bob!

Period 29:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
3bob$3b2ob$2b2ob2ob$3o4b3ob$bobo3bobob$2b4o2b3ob$4o2b3ob$2o2b4ob$b3obobob$
2bo2bob$b$6bob!

Period 46:

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
3bob$3b3ob$b3obob$obob5ob$b5o2bob$2bobobo2bob$b6ob$4bob$4bob!

General Comments and Conclusion

After working with this rule for several weeks, I was surprised at the variety of interesting objects and reactions that existed in it. It's a fairly simple rule, though admittedly more complicated than Conway's Life. However, like Life, it is far simpler than the phenomena it supports. I think it's interesting, among other things, for demonstrating that there is a 2-state hexagonal CA with at least some Life-like properties. I had been skeptical of this, since I had never seen a discussion of one.

There are some properties missing from this rule that make Life interesting. Unlike Life, patterns die down very rapidly unless they stabilize into honey-comb structures. The result of a collision between gliders and stable objects is similarly tame, with the most common results being an eater reaction or mutual annihilation. On the other hand, this can actually be beneficial for designing patterns to realize some logical function or growth rate.

Unlike Life, this rule does not seem to support finite still-life patterns. The honeycomb structure is stable and can present an arbitrarily long stable boundary, but corners always result in cell-producing neighborhoods that usually result in low-period oscillations. The stability of the honey-comb lattice is also very unlike Life, in which a small change to a dense lattice almost always propagates very fast, destroying it. It may be worth doing some systematic tests on random mutations of a lattice background.

[Naszvadi]

To the forum maintainers: where is the 99999999 like button?
To the authors: this publication has been floating around for ages, and it is pretty cool to share it this way. My 2cents are to disclose the original url, which is http://radicaleye.com/lifepage/hexrule.txt

[pcallahan]

To the authors: this publication has been floating around for ages, and it is pretty cool to share it this way.

Glad you like it. It took most of my Sunday morning, using a combination of scripting and hand-editing. I didn't bother to get the right bounding boxes for the RLE converted from images, and that doesn't seem to matter.

The big enabler is LifeViewer. I did not even realize it could handle hex rules until this morning.

[Sarah]

A p60 with a p3 and p20 part, sorry if this is already known.

Code: Select all

x = 12, y = 12, rule = B2o/S2m34H
bo$3o$bob3o$2b2obo$2bob2o$2b3ob2o$4bo3b2o$8bobo$9b2o$9bobo$10b2o$9b3o
!

[pcallahan]

A p60 with a p3 and p20 part, sorry if this is already known.

Look at the section "other guns" in the original post, and you'll see the same p20 part used to make a gun. I don't remember a lot about the specific oscillators I found, but they were pretty easy to generate from small random seeds that I ran and tested for periodicity. I may have missed some. I also didn't try any more sophisticated searches, so these are all natural objects.

[blah]

Code: Select all

x = 10, y = 10, rule = B2o/S2m34H
ob2o7b$b2obo6b$2ob2o6b$2b2obo5b$2bob2o5b$3b2obo4b$3bob2o4b$3b3obo3b$6b2o3b$
6bobo2b$6b3o2b$8b3o!

This structure (and others like it) are extremely stable. I can think of some intuitive reasons for why this is so, but I haven't considered the issue carefully. The practical implication is that unlike analogous Life patterns, there is no clear way to "ignite" the trail left behind by this pattern. So far, I have not found any kind of damage that propagates far along these stalks.

Here's a proof it can't be destroyed from the side, if nobody's proven it already.
Take this pattern, and assume it repeats vertically forever (* represents a live cell, . represents a dead one, and ? is unknown. Assume it's stabilised on the other side):

Code: Select all

* * . ?
 . * . ?
* * a ?
 . * b ?
* * . ?
 . * . ?

(Assume a and b are dead.)

In order for any *s to die from outside interference, they must be overcrowded.

If it is impossible for a to be born, overcrowding is impossible.

a already has 3 cells. Birth requires 2 cells. No changes to the ?s or even b will cause a to be born. Therefore, the pattern is indestructible.

In fact, it logically follows that since a can't be born, neither can b. So even the empty space around the pattern is indestructible.

I don't think it's indestructible from the ends though.

[GUYTU6J]

Brilliant rule! It surely warrants a LifeWiki OCA article, serving as a mature example of isotropic non-totalistic hexagonal rule. On catagolue it is named "Hexrule b2o"; is that name agreed by everyone?

And also, it may become the first Turing-complete such rule with an explicit Rule 110 unit cell construction.

BTW how far is the rule from being omniperiodic? Catagolue has recorded a large amount of oscillators at even periods, and the original write-up describes several low prime periods.

[pcallahan]

BTW how far is the rule from being omniperiodic? Catagolue has recorded a large amount of oscillators at even periods, and the original write-up describes several low prime periods.

I notice I was careful in the writeup to list "principal oscillation". I am not sure there is any way to eliminate the small-period cells around the fringes. A large-prime oscillator might be constructible but it would have to reset any finite stabilizers. It might be worth seeing if glider collisions can do this cheaply.

Added: I found this just by doodling. The "period-29" from the writeup is really period-87 because of the p3 cells on the stabilizer. However, a single glider collision can change the phase, so it should be possible to apply this method to remove multiples of 3 from the period of large engineered oscillators.

Code: Select all

x = 43, y = 30, rule = B2o/S2m34H
#C [[ GPS 2 ]]
2$10bobo$11b2o$10b2ob2obo$9b2o4b2o$8b2obo3bobo$7bo2b4o2b2o$8bob2o2b3o
bo$9b2ob4o$10bo2bo2bo$11bobo6$17bobo$18b2o$17b2ob2obo7bo$16b2o4b2o8bo
$15b2obo3bobo7bobo$14bo2b4o2b2o8bobo$15bob2o2b3obo8bo$16b2ob4o$17bo2b
o2bo$18bobo!

[praosylen]

I notice I was careful in the writeup to list "principal oscillation". I am not sure there is any way to eliminate the small-period cells around the fringes. A large-prime oscillator might be constructible but it would have to reset any finite stabilizers. It might be worth seeing if glider collisions can do this cheaply.

Added: I found this just by doodling. The "period-29" from the writeup is really period-87 because of the p3 cells on the stabilizer. However, a single glider collision can change the phase, so it should be possible to apply this method to remove multiples of 3 from the period of large engineered oscillators.

Code: Select all

rle

Another strategy that can work in some cases is symmetrizing, or more generally using additional copies of the original rotor in place of stabilizers at other periods. For example, here's a 26-cell oscillator that's solely p29:

Code: Select all

x = 8, y = 7, rule = B2o/S2m34H
2b3o$4obo$bo2b2o$2b4o$2b2o2bo$2bob4o$3b3o!

[pcallahan]

Another strategy that can work in some cases is symmetrizing, or more generally using additional copies of the original rotor in place of stabilizers at other periods.

Oh, cool! I can use this approach to fix up my p43 gun so it is no longer p86.

Code: Select all

x = 35, y = 39, rule = B2o/S2m34H
o$bobo$b3o10b2o$2b2o11b2o$4bo10bobo$4b2o7b6o$b5o8bo2b2o$bobo2bo7b4o$4b
3o9bob2o$5bobo8b3o$4b5o6b2obo$5bobo9b4o$6b3o7b2o2bo$6bo2bobo4b6o$7b5o
5bobo$7b2o9b2o$8bo10b2o$9b2o$9b3o$9bobo$12bo6$20b2o$21bo$21bo$22b3o$23b
ob2o$23bobobo2bo$23b9o$24bo2bobobo$28b2obo$30b3o$33bo$33bo$33b2o!

While the p2 cells never really bothered me much, it's nice to have a solution acceptable to purists.

Added: Just some doodling in Lifeviewer, but I see there's a simple gun period tripler. I don't think I found this before.

Code: Select all

x = 86, y = 79, rule = B2o/S2m34H
$59bo$59bo$59bob2o$59b3o$62b2o$51b2o6b2ob2o$49bob4o3b6o$50b2obo7bo2bo
$50bo10b4o$51b5o7bob2o$52bobo8b3o$52b3o7b2obo$52bobo9b4o$51b5o8bo2bo$
56bo8b6o$53bob2o8b2ob2o$52b4obo7b2o$54b2o11b3o$66b2obo$69bo$69bo$60bo
$59bobo$60bo$61b2o$63bo2$66bo$64bobo2bo$65b5o$69b3o$70bob2o4bo$70bobo
bobobo$69b11o$70bobobobobo$70bo4b2obo$77b3o$79b5o$79bo2bobo$82bo3$12b
o48bo$13b2o44bo$11b3obo3bo40bo$11b2o2b5o2bo38bo$12b2obobobob2o39b2o$12b
o2b5o2b2o$15bo3bob3o$20b2o$22bo$12bo$6b4ob2o15bo20bo$9bob2o14bobo20bo
$8b2o2b3o13bobo17bo2bo$8bob3obo2bo12bo21bo$10b3ob4o13bo20bo$13b2obob2o
14bo$15b4ob3o12bo$15bo2bob3obo8b3o$18b3o2b2o10bo$20b2obo11bo$20b2ob4o
7b3o2bo$20bo15b6o$36bo2bo$37b4o$37bobo$38b3o$39bobo$38b4o$39bo2bo$37b
6o$39bo2b3o$43bo$43bo$43b3o$43bo$44bo!

Added: It doesn't work with even periods. That might be why I missed it in 1997.

[confocaloid]

Unfortunately Golly doesn't (yet? version 4.2) support the isotropic hexagonal rulestring notation. I believe the following weighted-neighbourhood rulestring is an equivalent definition of the rule, so patterns can work unchanged:

Code: Select all

#C Source: https://conwaylife.com/forums/viewtopic.php?p=113322#p113322
#C A glider loop based on period-42 reflectors
x = 89, y = 89, rule = R1,C0,S5,7,10-11,13-15,17,19-23,25-30,34-35,37-46,49-54,56-58,60,B3,6,12,24,33,48,NW010200200004001008H
29bo$30b2o$5bo22b3ob3o$4bobo23bobo2bo$5b4o21bob4o7bobo$4b2o3bo20b3obob
o5bob2obo$6b4o23b2o8b2ob2o$5b2obobo22bo9bobobo$7b5o32b3obo$10bo30bo3bo
b2o$38b4o3b3o$47bo$2b2o5bo19bo$2bob2o4bo19bo$b2obo5bo17bo2bo$2ob3o4bo
21bo$2bobobo3bo21bo$2b2ob2o$2bob2obo$5bo8$12b4o$16bo41bo11bo$61bo9bobo
$14bo46bo7b5o$61bo8bo3bo$60b2o9b4o$70b2obobo$72b3obo$75b2o$75bo2$72bo$
73bo$4b2o68bo$2b3ob3o66b2o$4bobobo72bo$3b2ob4o70b2obo$5b2obo71bob2o$5b
ob2o70b4ob2o$7bo72bobobo$12b2o66b3ob3o$14bo68b2o$15bo$16bo2$13bo$12b2o
$12bob3o$13bobob2o$14b4o9b2o$14bo3bo8bo$15b5o7bo46bo$15bobo9bo$18bo11b
o41bo$73b4o8$83bo$81bob2obo$82b2ob2o$56bo21bo3bobobo$56bo21bo4b3ob2o$
57bo2bo17bo5bob2o$58bo19bo4b2obo$59bo19bo5b2o$41bo$41b3o3b4o$40b2obo3b
o30bo$40bob3o32b5o$41bobobo9bo22bobob2o$41b2ob2o8b2o23b4o$41bob2obo5bo
bob3o20bo3b2o$43bobo7b4obo21b4o$53bo2bobo23bobo$54b3ob3o22bo$57b2o$59b
o!

[wildmyron]

Unfortunately Golly doesn't (yet? version 4.2) support the isotropic hexagonal rulestring notation. I believe the following weighted-neighbourhood rulestring is an equivalent definition of the rule, so patterns can work unchanged:

Code: Select all

<snip rle>

While the isotropic hexagonal rule notation isn't supported by Golly, the non-totalistic hexagonal rules are supported with hexagonal MAP rules. This post provides a script to convert the isotropic hexagonal notation to a MAP rule, along with an example pattern in this very rule.

As a demonstration, here's Callahan's B2o/S2m34H glider gun:

Code: Select all

x = 25, y = 13, rule = MAPFAMgVwhXAP4AP4B+gH4A6A
17b2o$2bo15b2obo$b2obo13bob2o$2ob3o13b2ob2o$2bobobo13bobo$2b2ob4o3b2o
6bob2o$2bob2obo5b3o4b3obo$13bobo6bo$12b4o$12bo3bo$13b5o$15bo$15bo!

[Naszvadi]

While going over this old write-up of mine, I realize that it is greatly improved if you can see the patterns in action, which has been possible in LifeViewer for some time.
...

6 unit cells for rule 110:

Code: Select all

x = 1491, y = 267, rule = B2o/S2m34H:T1536,0
69bobo253bobo253bobo253bobo253bobo253bobo$68bob2obo250bob2obo250bob2ob
o250bob2obo250bob2obo250bob2obo$69b2ob2o251b2ob2o251b2ob2o251b2ob2o
251b2ob2o251b2ob2o$70bobobo251bobobo251bobobo251bobobo251bobobo251bobo
bo$70bob4o250bob4o250bob4o250bob4o250bob4o250bob4o$70b3o3bo8b2o239b3o
3bo8b2o239b3o3bo8b2o239b3o3bo8b2o239b3o3bo8b2o239b3o3bo8b2o$10bo61bo
13bobobo175bo61bo13bobobo175bo61bo13bobobo175bo61bo13bobobo175bo61bo
13bobobo175bo61bo13bobobo$11b2o72b2o2b2o176b2o72b2o2b2o176b2o72b2o2b2o
176b2o72b2o2b2o176b2o72b2o2b2o176b2o72b2o2b2o$87b3ob2o250b3ob2o250b3ob
2o250b3ob2o250b3ob2o250b3ob2o$86b2obobo250b2obobo250b2obobo250b2obobo
250b2obobo250b2obobo$88b2ob3o250b2ob3o250b2ob3o250b2ob3o250b2ob3o250b
2ob3o$88bob2o252bob2o252bob2o252bob2o252bob2o252bob2o5$13bobo253bobo
253bobo253bobo253bobo253bobo$13b3obo251b3obo251b3obo251b3obo251b3obo
251b3obo$14bob2o52bobo197bob2o52bobo197bob2o52bobo197bob2o52bobo197bob
2o52bobo197bob2o52bobo$12bo2bobo51bob2o195bo2bobo51bob2o195bo2bobo51bo
b2o195bo2bobo51bob2o195bo2bobo51bob2o195bo2bobo51bob2o$3bobo5b6obo51b
2ob2o184bobo5b6obo51b2ob2o184bobo5b6obo51b2ob2o184bobo5b6obo51b2ob2o
184bobo5b6obo51b2ob2o184bobo5b6obo51b2ob2o$4b2obo5bobob3o51bobo2bo183b
2obo5bobob3o51bobo2bo183b2obo5bobob3o51bobo2bo183b2obo5bobob3o51bobo2b
o183b2obo5bobob3o51bobo2bo183b2obo5bobob3o51bobo2bo$2b3ob4o6b2o53bob2o
b2o180b3ob4o6b2o53bob2ob2o180b3ob4o6b2o53bob2ob2o180b3ob4o6b2o53bob2ob
2o180b3ob4o6b2o53bob2ob2o180b3ob4o6b2o53bob2ob2o$4bobobo8b2o53b2ob2o
183bobobo8b2o53b2ob2o183bobobo8b2o53b2ob2o183bobobo8b2o53b2ob2o183bobo
bo8b2o53b2ob2o183bobobo8b2o53b2ob2o$4bob3o8bo56b2obo182bob3o8bo56b2obo
182bob3o8bo56b2obo182bob3o8bo56b2obo182bob3o8bo56b2obo182bob3o8bo56b2o
bo$5b2obo65bobo184b2obo65bobo184b2obo65bobo184b2obo65bobo184b2obo65bob
o184b2obo65bobo$7b2o254b2o254b2o254b2o254b2o254b2o$8b2o254b2o254b2o
254b2o254b2o254b2o3$89bo255bo255bo255bo255bo255bo$86b4obo250b4obo250b
4obo250b4obo250b4obo250b4obo$88bob2o252bob2o252bob2o252bob2o252bob2o
252bob2o$89b2ob2o251b2ob2o251b2ob2o251b2ob2o251b2ob2o251b2ob2o$90bobo
253bobo253bobo253bobo253bobo253bobo$90bob3o251bob3o251bob3o251bob3o
251bob3o251bob3o$90b3o253b3o253b3o253b3o253b3o253b3o$92bo255bo255bo
255bo255bo255bo26$68bob2obo250bob2obo250bob2obo250bob2obo250bob2obo
250bob2obo$69b2ob4o249b2ob4o249b2ob4o249b2ob4o249b2ob4o249b2ob4o$70bob
obo251bobobo251bobobo251bobobo251bobobo251bobobo$69b2ob3o250b2ob3o250b
2ob3o250b2ob3o250b2ob3o250b2ob3o$71b2obo12bo239b2obo12bo239b2obo12bo
239b2obo12bo239b2obo12bo239b2obo12bo$73bo12b3o240bo12b3o240bo12b3o240b
o12b3o240bo12b3o240bo12b3o$86b2ob2o251b2ob2o251b2ob2o251b2ob2o251b2ob
2o251b2ob2o$87b3obo251b3obo251b3obo251b3obo251b3obo251b3obo$87bobobo
251bobobo251bobobo251bobobo251bobobo251bobobo$86b4ob3o248b4ob3o248b4ob
3o248b4ob3o248b4ob3o248b4ob3o$88bob2o252bob2o252bob2o252bob2o252bob2o
252bob2o$91bo255bo255bo255bo255bo255bo5$138b2obo252b2obo252b2obo252b2o
bo252b2obo252b2obo$137b2ob2o251b2ob2o251b2ob2o251b2ob2o251b2ob2o251b2o
b2o$69bob2o62b3obobo183bob2o62b3obobo183bob2o62b3obobo183bob2o62b3obob
o183bob2o62b3obobo183bob2o62b3obobo$70b2ob3o60b5ob2o182b2ob3o60b5ob2o
182b2ob3o60b5ob2o182b2ob3o60b5ob2o182b2ob3o60b5ob2o182b2ob3o60b5ob2o$
71bobobo52b2o11b2o184bobobo52b2o11b2o184bobobo52b2o11b2o184bobobo52b2o
11b2o184bobobo52b2o11b2o184bobobo52b2o11b2o$71bob4o49b3ob3o8bo185bob4o
49b3ob3o8bo185bob4o49b3ob3o8bo185bob4o49b3ob3o8bo185bob4o49b3ob3o8bo
185bob4o49b3ob3o8bo$71b3ob2o51bobobo194b3ob2o51bobobo194b3ob2o51bobobo
194b3ob2o51bobobo194b3ob2o51bobobo194b3ob2o51bobobo$74b2o52bob4o196b2o
52bob4o196b2o52bob4o196b2o52bob4o196b2o52bob4o196b2o52bob4o$74b2o52b3o
bo197b2o52b3obo197b2o52b3obo197b2o52b3obo197b2o52b3obo197b2o52b3obo$
76bo53bo2bo198bo53bo2bo198bo53bo2bo198bo53bo2bo198bo53bo2bo198bo53bo2b
o$133b2o254b2o254b2o254b2o254b2o254b2o$133bo255bo255bo255bo255bo255bo$
87bo255bo255bo255bo255bo255bo$88b2o254b2o254b2o254b2o254b2o254b2o$88b
2o254b2o254b2o254b2o254b2o254b2o$88bob2o252bob2o252bob2o252bob2o252bob
2o252bob2o$88b3obo251b3obo251b3obo251b3obo251b3obo251b3obo$90bobo253bo
bo253bobo253bobo253bobo253bobo$89b2ob3o250b2ob3o250b2ob3o250b2ob3o250b
2ob3o250b2ob3o$91b2o254b2o254b2o254b2o254b2o254b2o$87bo5bo5b2o242bo5bo
5b2o242bo5bo5b2o242bo5bo5b2o242bo5bo5b2o242bo5bo5b2o$87bobo7b3obo241bo
bo7b3obo241bobo7b3obo241bobo7b3obo241bobo7b3obo241bobo7b3obo$85b5o9bob
o239b5o9bobo239b5o9bobo239b5o9bobo239b5o9bobo239b5o9bobo$38bo47bo3b2o
7bob2o191bo47bo3b2o7bob2o191bo47bo3b2o7bob2o191bo47bo3b2o7bob2o191bo
47bo3b2o7bob2o191bo47bo3b2o7bob2o$39b2o45b5o8b3obo191b2o45b5o8b3obo
191b2o45b5o8b3obo191b2o45b5o8b3obo191b2o45b5o8b3obo191b2o45b5o8b3obo$
89bobo10b4o239bobo10b4o239bobo10b4o239bobo10b4o239bobo10b4o239bobo10b
4o$89b4o11bo240b4o11bo240b4o11bo240b4o11bo240b4o11bo240b4o11bo$66bo24b
o13bo216bo24bo13bo216bo24bo13bo216bo24bo13bo216bo24bo13bo216bo24bo13bo
$64bob3o251bob3o251bob3o251bob3o251bob3o251bob3o$65b2obo44bo2bo76bo
127b2obo44bo2bo76bo127b2obo44bo2bo76bo127b2obo44bo2bo76bo127b2obo44bo
2bo76bo127b2obo44bo2bo76bo$66bobo12bo30b3obo76bobo126bobo12bo30b3obo
76bobo126bobo12bo30b3obo76bobo126bobo12bo30b3obo76bobo126bobo12bo30b3o
bo76bobo126bobo12bo30b3obo76bobo$65b2ob3o10bo29b2obo2bo73b5o125b2ob3o
10bo29b2obo2bo73b5o125b2ob3o10bo29b2obo2bo73b5o125b2ob3o10bo29b2obo2bo
73b5o125b2ob3o10bo29b2obo2bo73b5o125b2ob3o10bo29b2obo2bo73b5o$67b2obo
10bo2bo28b4o75bo3b2o125b2obo10bo2bo28b4o75bo3b2o125b2obo10bo2bo28b4o
75bo3b2o125b2obo10bo2bo28b4o75bo3b2o125b2obo10bo2bo28b4o75bo3b2o125b2o
bo10bo2bo28b4o75bo3b2o$67bob3o9bo31bo3b2o73b5o126bob3o9bo31bo3b2o73b5o
126bob3o9bo31bo3b2o73b5o126bob3o9bo31bo3b2o73b5o126bob3o9bo31bo3b2o73b
5o126bob3o9bo31bo3b2o73b5o$71bo10bo30b5o77bobo129bo10bo30b5o77bobo129b
o10bo30b5o77bobo129bo10bo30b5o77bobo129bo10bo30b5o77bobo129bo10bo30b5o
77bobo$71bo44bobo76bobo129bo44bobo76bobo129bo44bobo76bobo129bo44bobo
76bobo129bo44bobo76bobo129bo44bobo76bobo$79b3o35bo77b2obo136b3o35bo77b
2obo136b3o35bo77b2obo136b3o35bo77b2obo136b3o35bo77b2obo136b3o35bo77b2o
bo$82b2o112b3o139b2o112b3o139b2o112b3o139b2o112b3o139b2o112b3o139b2o
112b3o$72bo7b2o11bo234bo7b2o11bo234bo7b2o11bo234bo7b2o11bo234bo7b2o11b
o234bo7b2o11bo$71b2o21b2o231b2o21b2o231b2o21b2o231b2o21b2o231b2o21b2o
231b2o21b2o$70b2ob5o18bo229b2ob5o18bo229b2ob5o18bo229b2ob5o18bo229b2ob
5o18bo229b2ob5o18bo$72bobob3o16bobo230bobob3o16bobo230bobob3o16bobo
230bobob3o16bobo230bobob3o16bobo230bobob3o16bobo$72b2ob2o19bo231b2ob2o
19bo231b2ob2o19bo231b2ob2o19bo231b2ob2o19bo231b2ob2o19bo$72bob2o7bobo
242bob2o7bobo242bob2o7bobo242bob2o7bobo242bob2o7bobo242bob2o7bobo$84b
2o115b3obo134b2o115b3obo134b2o115b3obo134b2o115b3obo134b2o115b3obo134b
2o115b3obo$84b2o116bob2o134b2o116bob2o134b2o116bob2o134b2o116bob2o134b
2o116bob2o134b2o116bob2o$84bobo114b4ob2o132bobo114b4ob2o132bobo114b4ob
2o132bobo114b4ob2o132bobo114b4ob2o132bobo114b4ob2o$83b4o114bo2bobo132b
4o114bo2bobo132b4o114bo2bobo132b4o114bo2bobo132b4o114bo2bobo132b4o114b
o2bobo$83bo3bo113b2obob2o131bo3bo113b2obob2o131bo3bo113b2obob2o131bo3b
o113b2obob2o131bo3bo113b2obob2o131bo3bo113b2obob2o$84b5o114bob2obo131b
5o114bob2obo131b5o114bob2obo131b5o114bob2obo131b5o114bob2obo131b5o114b
ob2obo$86bo255bo255bo255bo255bo255bo$87bo255bo255bo255bo255bo255bo9$
93b3o253b3o253b3o253b3o253b3o253b3o$96b2o254b2o254b2o254b2o254b2o254b
2o$94b2o254b2o254b2o254b2o254b2o254b2o7$bo255bo255bo255bo255bo255bo$2b
o255bo255bo255bo255bo255bo$5o251b5o251b5o251b5o251b5o251b5o$bo3bo251bo
3bo251bo3bo251bo3bo251bo3bo251bo3bo$b5o251b5o251b5o251b5o251b5o251b5o$
4bob2o252bob2o252bob2o252bob2o252bob2o252bob2o$4b3o253b3o253b3o253b3o
253b3o253b3o$3bo255bo255bo255bo255bo255bo$5bo255bo255bo255bo255bo255bo
3$93bo255bo255bo255bo255bo255bo$94b2o254b2o254b2o254b2o254b2o254b2o$
96bo255bo255bo255bo255bo255bo$5bo89bobo163bo89bobo163bo89bobo163bo89bo
bo163bo89bobo163bo89bobo$6b4o86bo165b4o86bo165b4o86bo165b4o86bo165b4o
86bo165b4o86bo$4b3o2bo250b3o2bo250b3o2bo250b3o2bo250b3o2bo250b3o2bo$5b
ob3o251bob3o251bob3o251bob3o251bob3o251bob3o$6bobobo251bobobo251bobobo
251bobobo251bobobo251bobobo$6b2ob3o250b2ob3o250b2ob3o250b2ob3o250b2ob
3o250b2ob3o$6bob2o252bob2o252bob2o252bob2o252bob2o252bob2o$8bobo253bob
o253bobo253bobo253bobo253bobo12$93b3o253b3o253b3o253b3o253b3o$96b2o
254b2o254b2o254b2o254b2o$94b2o254b2o254b2o254b2o254b2o17$28bo176b2o77b
o176b2o77bo176b2o77bo176b2o77bo176b2o77bo176b2o$28bo64bo111bob3o74bo
64bo111bob3o74bo64bo111bob3o74bo64bo111bob3o74bo64bo111bob3o74bo64bo
111bob3o$29bo64b2o106bo3bobo76bo64b2o106bo3bobo76bo64b2o106bo3bobo76bo
64b2o106bo3bobo76bo64b2o106bo3bobo76bo64b2o106bo3bobo$96bo107bob2obo
142bo107bob2obo142bo107bob2obo142bo107bob2obo142bo107bob2obo142bo107bo
b2obo$95bobo107b2ob3o140bobo107b2ob3o140bobo107b2ob3o140bobo107b2ob3o
140bobo107b2ob3o140bobo107b2ob3o$96bo110b2o143bo110b2o143bo110b2o143bo
110b2o143bo110b2o143bo110b2o3$200bo255bo255bo255bo255bo255bo$200bo255b
o255bo255bo255bo255bo$200b2o254b2o254b2o254b2o254b2o254b2o$199b2obo
252b2obo252b2obo252b2obo252b2obo252b2obo$201b5o251b5o251b5o251b5o251b
5o251b5o$200b2o3bo250b2o3bo250b2o3bo250b2o3bo250b2o3bo250b2o3bo$202b5o
251b5o251b5o251b5o251b5o251b5o$145bob2o53bobo196bob2o53bobo196bob2o53b
obo196bob2o53bobo196bob2o53bobo196bob2o53bobo$28bo117b2ob3o52bo79bo
117b2ob3o52bo79bo117b2ob3o52bo79bo117b2ob3o52bo79bo117b2ob3o52bo79bo
117b2ob3o52bo$28bo117bobobo133bo117bobobo133bo117bobobo133bo117bobobo
133bo117bobobo133bo117bobobo$29bo117b3obo133bo117b3obo133bo117b3obo
133bo117b3obo133bo117b3obo133bo117b3obo$148bob3o251bob3o251bob3o251bob
3o251bob3o251bob3o$147bo3bo251bo3bo251bo3bo251bo3bo251bo3bo251bo3bo$
149bo255bo255bo255bo255bo255bo2$93b3o47bo205b3o47bo205b3o47bo205b3o47b
o205b3o47bo205b3o47bo$96b2o45b2o207b2o45b2o207b2o45b2o207b2o45b2o207b
2o45b2o207b2o45b2o$94b2o45b3obo204b2o45b3obo204b2o45b3obo204b2o45b3obo
204b2o45b3obo204b2o45b3obo$142bobob2o250bobob2o250bobob2o250bobob2o
250bobob2o250bobob2o$143b4o252b4o252b4o252b4o252b4o252b4o$142b2o3bo
250b2o3bo250b2o3bo250b2o3bo250b2o3bo250b2o3bo$144b5o251b5o251b5o251b5o
251b5o251b5o$144bobo253bobo253bobo253bobo253bobo253bobo$146bo255bo255b
o255bo255bo255bo12$78bob2o11bo240bob2o11bo240bob2o11bo240bob2o11bo240b
ob2o11bo240bob2o11bo$79b2ob3o9b2o239b2ob3o9b2o239b2ob3o9b2o239b2ob3o9b
2o239b2ob3o9b2o239b2ob3o9b2o$80bobob2o10bo239bobob2o10bo239bobob2o10bo
239bobob2o10bo239bobob2o10bo239bobob2o10bo$80bob3o10bobo238bob3o10bobo
238bob3o10bobo238bob3o10bobo238bob3o10bobo238bob3o10bobo$80b3obo11bo
239b3obo11bo239b3obo11bo239b3obo11bo239b3obo11bo239b3obo11bo$85bo255bo
255bo255bo255bo255bo$84b3o253b3o253b3o253b3o253b3o253b3o$86bo255bo255b
o255bo255bo255bo2$85b2o254b2o254b2o254b2o254b2o254b2o$86b2o254b2o254b
2o254b2o254b2o254b2o$84b3obobo249b3obobo249b3obobo249b3obobo249b3obobo
249b3obobo$85bob6o248bob6o248bob6o248bob6o248bob6o248bob6o$86bobo2bo
250bobo2bo250bobo2bo250bobo2bo250bobo2bo250bobo2bo$86b2obo252b2obo252b
2obo252b2obo252b2obo252b2obo$86bob3o251bob3o251bob3o251bob3o251bob3o
251bob3o$88bobo5bob3o243bobo5bob3o243bobo5bob3o243bobo5bob3o243bobo5bo
b3o243bobo5bob3o$96b3obo251b3obo251b3obo251b3obo251b3obo251b3obo$96bob
o2bo250bobo2bo250bobo2bo250bobo2bo250bobo2bo250bobo2bo$97b4o252b4o252b
4o252b4o252b4o252b4o$96b2o3bo250b2o3bo250b2o3bo250b2o3bo250b2o3bo250b
2o3bo$98b4o252b4o252b4o252b4o252b4o252b4o$98bobo253bobo253bobo253bobo
253bobo253bobo$101bo255bo255bo255bo255bo255bo!

[Naszvadi]

UC

Period is 2184, if I was correct.

[iddi01]

Brilliant rule! It surely warrants a LifeWiki OCA article, serving as a mature example of isotropic non-totalistic hexagonal rule.

I created a stub article at OCA:B2o/S2m34H.

To editors: feel free to add stuff to the page!

[confocaloid]

I made a four-state variant of this CA. Surviving cells retain their state; a newborn cell either "inherits" the state or takes on the remaining live state. There are several hybrid gliders and other hybrid objects.

Code: Select all

#C The pattern shows a p42 gun and a p43 gun, shooting hybrid gliders.
#C The ruletable is appended at the end.
x = 55, y = 47, rule = HexRGB_B2oS2m34_v0
19.2ABC$19.A2$23.A7$35.B$21.2C13.B$22.2C10.6B$22.C.3C9.2B.3B$20.4C.C.
2C11.B.3B2.B$16.C2.2C.C.2C2.C11.B.B.B3.B$16.2C.C.4C15.9B$18.3C.C17.B
3.B.B.B$21.2C18.B2.3B.B$22.2C23.3B.2B$49.6B$52.B$53.B$38.4A$38.2A.2A$
40.3A.A$41.A.2A$41.A.A.A.A$8.BC31.9A$8.C34.A.A.A.A$8.C37.2A.A$8.C37.A
.3A$11.C36.2A.2A$15.B.B31.4A$16.2B$.A2.2A10.4B$3A.3A10.2B.3B$.A.2A2.A
3.2C5.3B.B$2.A.A.2A2.5C5.B.B$2.2A.2A4.2C.C5.B.2B$2.A.2A5.2C.C6.2B.B$
4.A.A4.C.C.C$12.4C$12.C3.C$13.4C$13.C.C$15.C!

@RULE HexRGB_B2oS2m34_v0

https://conwaylife.com/forums/viewtopic.php?p=182451#p182451

@COLORS

0   48  48  48
1  240  48  48
2   48 240  48
3   48  48 240

@TABLE
n_states:4
neighborhood:hexagonal
symmetries:rotate6reflect

var x={0,1,2,3}
var x1=x
var x2=x
var x3=x
var x4=x
var x5=x
var x6=x
var a={1,2,3}
var a1=a
var a2=a
var a3=a
var a4=a

# birth rules (B2)
0, a1,a1, 0, 0, 0, 0, a1
0,  1, 2, 0, 0, 0, 0,  3
0,  1, 3, 0, 0, 0, 0,  2
0,  2, 3, 0, 0, 0, 0,  1
# survival rules (S2)
a, a1, 0,a2, 0, 0, 0,  a
# survival rules (S3)
a, a1,a2,a3, 0, 0, 0,  a
a, a1,a2, 0,a3, 0, 0,  a
a, a1, 0,a2, 0,a3, 0,  a
# survival rules (S4)
a,  0, 0,a1,a2,a3,a4,  a
a,  0,a1, 0,a2,a3,a4,  a
a,  0,a1,a2, 0,a3,a4,  a
# otherwise, the middle cell becomes dead
x, x1,x2,x3,x4,x5,x6,  0