Thread for basic non-CGOL questions

[pzq_alex]

I have seen some certain rules discussed within a particular rulespace, the rules have RRO Herschels (or Herschel descendants or whatever) that
interact to create strange extremely slow or high period spaceships like in this rule

Code: Select all

x = 91, y = 113, rule = B2in3-ekny4kt5ery6cin7c8/S2aek3-ace4eiyz5eik6-ck7e8
6bo$5bo$5bo2bo$3bo3b2o$4bo2bo$4bo2bo$4b3o15$3o$obo$2b2o$obo$obo$2o29$
22b2o$21bobo$21bo$21bo2bo8b2o$22b2obo6b2o10b2o$25bob2o17b2o$25b3o14bo
4bo$26bo14bo2bo2bo$45b2o14$87b2o$89bo$86b2ob2o$85bob3o$82bo4b2o$86bo
26$54b2o$54bobo$48b2o4bob2o$47bobo5b2o$48b2o5bo!

most of these kind of spaceships are all in different (but related) rules. where did these kind of rules originate from? is there a rule which led to the discovery of this rulespace?

These are rules where c/2 frontends (B-heptominoes and relatives) act as failed replicators. There are lots of such rules, and it's hard to say for sure which one was first. Thrre are also multiple rulespaces that can be found by hand rulegolfing and search programs exist that sweep through the entire rulespace looking for spaceships.

[confocaloid]

There are oscillators "made out of themselves", where one phase is made of several non-interacting copies of another phase. Which sequences of the number of copies are possible? Examples for sequences (1,2) and (1,2,2), found from this replicator using searchRule-matchPatt.py:

Code: Select all

#C period 59, 2 copies at T = 21
x = 6, y = 1, rule = B3-cr4cejnw6ak/S2-ce34jn5cei6akn
6o!

Code: Select all

#C period 73, 2 copies at T = 21, 2 copies at T = 61
x = 6, y = 1, rule = B3-r4cjw5a6cn7c/S2-c3-c5ai6aen
6o!

Code: Select all

#C period 83, 2 copies at T = 21, 2 copies at T = 52
x = 6, y = 1, rule = B3-r4cw5aqy6c/S02-c3-n4k5ai6a
6o!

Code: Select all

#C period 90, 2 copies at T = 21, 2 copies at T = 72
x = 6, y = 1, rule = B3-ck4cjkw5acjy6a/S2-ce3-c4jnqy5ai6akn
6o!

In the last example, two copies of phase 0 appear at different times in two different ways.

[w33z8kqrqk8zzzw33]

There are oscillators "made out of themselves", where one phase is made of several non-interacting copies of another phase. Which sequences of the number of copies are possible? Examples for sequences (1,2) and (1,2,2), found from this replicator using searchRule-matchPatt.py:

Code: Select all

#C period 59, 2 copies at T = 21
x = 6, y = 1, rule = B3-cr4cejnw6ak/S2-ce34jn5cei6akn
6o!

Code: Select all

#C period 73, 2 copies at T = 21, 2 copies at T = 61
x = 6, y = 1, rule = B3-r4cjw5a6cn7c/S2-c3-c5ai6aen
6o!

Code: Select all

#C period 83, 2 copies at T = 21, 2 copies at T = 52
x = 6, y = 1, rule = B3-r4cw5aqy6c/S02-c3-n4k5ai6a
6o!

Code: Select all

#C period 90, 2 copies at T = 21, 2 copies at T = 72
x = 6, y = 1, rule = B3-ck4cjkw5acjy6a/S2-ce3-c4jnqy5ai6akn
6o!

In the last example, two copies of phase 0 appear at different times in two different ways.

Here's another (1,2,2)

Code: Select all

#C P1064, 1 @ 0, 2 @ 53, 2 @ 98
x = 27, y = 5, rule = B2n3-cy4t5acqy6c7/S2-i3-cknr4ct5cn6ce
o5bo13bo5bo$b2ob2o15b2ob2o2$2b3o17b3o$3bo19bo!

[pzq_alex]

What is the skew torus syntax again? / how do I run this infinitely long wick on a skew torus?

Code: Select all

x = 43, y = 168, rule = FeynmanDiagramGen
2C$2C$CD$.2C$.2C$.CD$2.2C$2.2C$2.CD$3.2C$3.2C$3.CD$4.2C$4.2C$4.CD$5.2C
$5.2C$5.2C$5.2C$5.2C$5.2C$5.2C$5.2C$5.CD$6.2C$6.2C$6.CD$7.2C$7.2C$7.C
D$8.2C$8.2C$8.CD$9.2C$9.2C$9.CD$10.2C$10.2C$10.CD$11.2C$11.2C$11.2C$11.
2C$11.2C$11.2C$11.2C$11.2C$11.CD$12.2C$12.2C$12.CD$13.2C$13.2C$13.CD$
14.2C$14.2C$14.CD$15.2C$15.2C$15.CD$16.2C$16.2C$16.CD$17.2C$17.2C$17.
2C$17.2C$17.2C$17.2C$17.2C$17.2C$17.CD$18.2C$18.2C$18.CD$19.2C$19.2C$
19.CD$20.2C$20.2C$20.CD$21.2C$21.2C$21.CD$22.2C$22.2C$22.CD$23.2C$23.
2C$23.2C$23.2C$23.2C$23.2C$23.2C$23.2C$23.CD$24.2C$24.2C$24.CD$25.2C$
25.2C$25.CD$26.2C$26.2C$26.CD$27.2C$27.2C$27.CD$28.2C$28.2C$28.CD$29.
2C$29.2C$29.2C$29.2C$29.2C$29.2C$29.2C$29.2C$29.CD$30.2C$30.2C$30.CD$
31.2C$31.2C$31.CD$32.2C$32.2C$32.CD$33.2C$33.2C$33.CD$34.2C$34.2C$34.
CD$35.2C$35.2C$35.2C$35.2C$35.2C$35.2C$35.2C$35.2C$35.CD$36.2C$36.2C$
36.CD$37.2C$37.2C$37.CD$38.2C$38.2C$38.CD$39.2C$39.2C$39.CD$40.2C$40.
2C$40.CD$41.2C$41.2C$41.2C$41.2C$41.2C$41.2C$41.2C$41.2C$41.CD!
@RULE FeynmanDiagramGen
1 = determined off
2 = determined on
3 = active off
4 = active on
@TABLE
n_states: 5
neighborhood: vonNeumann
symmetries: none
var a.1={0,1,2,3,4}
var a.2=a.1
var a.3=a.1
var a.4=a.1
var a.5=a.1
var b.1={1,3}
var c.1={2,4}
a.1,a.2,b.1,a.3,a.4,1
a.1,a.2,c.1,a.3,a.4,2
a.1,a.2,a.3,3,c.1,4
a.1,a.2,a.3,4,b.1,4
a.1,a.2,a.3,3,b.1,3
a.1,a.2,a.3,4,c.1,3
a.1,a.2,a.3,a.4,a.5,0
@COLORS
0 0 0 0
1 60 60 60
2 255 60 60
3 127 127 127
4 255 127 127

[DroneBetter]

Being that Margolus/block neighbourhood rules can be reversible by ensuring that the state transition diagrams are bijective (which can be done isotropically, as in Critters), and a 16-state MAP rule (with D4_x symmetry) can encode such a rule (by making each cell four cells and each iteration two iterations), are there any known bounds on the minimum number of states and maximum order of symmetry necessary for a rule that is enacted homogenously at a cell granularity to be reversible (ie. proofs by example or disproofs of specific cases)? (Excepting those that are trivial due to being uninteresting, say I would like one that allows patterns to exceed their bounding octagons, I daresay even requiring Turing-completeness, if that isn't disallowed by any parity rules.)

[muzik]

Is it possible for both of these oscillators to coexist in one rule by alternating two non-isotropic rules? From what I can see, S2a is the main issue here, but only four out of every eight possible rotations/reflections of 2a are present at any given generation. The oscillators need not still work after being reflected.

Code: Select all

x = 6, y = 6, rule = B3a/S2ae
2b2o$3bo$o3b2o$2o3bo$2bo$2b2o!

Code: Select all

x = 6, y = 6, rule = B3a/S3kq
3bo$2b2o$bo2b2o$2o2bo$2b2o$2bo!

[confocaloid]

Is it possible for both of these oscillators to coexist in one rule by alternating two non-isotropic rules? From what I can see, S2a is the main issue here, but only four out of every eight possible rotations/reflections of 2a are present at any given generation. The oscillators need not still work after being reflected.

Code: Select all

x = 6, y = 6, rule = B3a/S2ae
2b2o$3bo$o3b2o$2o3bo$2bo$2b2o!

Code: Select all

x = 6, y = 6, rule = B3a/S3kq
3bo$2b2o$bo2b2o$2o2bo$2b2o$2bo!

With rules changed to be closer to each other, the first rule has S2a and second rule has S3q:

Code: Select all

x = 6, y = 6, rule = B3a/S2ae3k
2b2o$3bo$o3b2o$2o3bo$2bo$2b2o!

Code: Select all

x = 6, y = 6, rule = B3a/S2e3kq
3bo$2b2o$bo2b2o$2o2bo$2b2o$2bo!

Each of 8 subconditions of S2a/D2a and each of 8 subconditions of S3q/D3q occurs once in period. I think it will work if each of those subconditions is alternated between S and D:

Code: Select all

x = 18, y = 35, rule = B/S012345678History
4.3F8.F$4.F.F7.2F$4.F.F8.F$4.F.F8.F$4.3F7.3F4$2.4D6.4D$2.D2CD6.D2CD$
4DC3D2.3DC4D$DC3D2CD2.D2C3DCD$D2C3DCD2.DC3D2CD$3DC4D2.4DC3D$2.D2CD6.D
2CD$2.4D6.4D2$18F2$2.4D6.4D$2.2DCD6.DC2D$3D2C3D2.3D2C3D$2DC2D2CD2.D2C
2DC2D$D2C2DC2D2.2DC2D2CD$3D2C3D2.3D2C3D$2.DC2D6.2DCD$2.4D6.4D4$4.3F8.
F$4.F.F7.2F$4.F.F8.F$4.F.F8.F$4.3F7.3F!

[confocaloid]

How cheap (in terms of glider synthesis cost) can be a RCT, if you can choose any isotropic rule that is both fertile and mortal, and then you can choose (to be used in a synthesis) any glider that works in the chosen rule? What if Golly ruletables are allowed as well?

[pzq_alex]

How cheap (in terms of glider synthesis cost) can be a RCT, if you can choose any isotropic rule that is both fertile and mortal, and then you can choose (to be used in a synthesis) any glider that works in the chosen rule? What if Golly ruletables are allowed as well?

Three.

[confocaloid]

How cheap (in terms of glider synthesis cost) can be a RCT, if you can choose any isotropic rule that is both fertile and mortal, and then you can choose (to be used in a synthesis) any glider that works in the chosen rule? What if Golly ruletables are allowed as well?

Three.

As far as I understand, a three-glider RCT would be something like this?

(1) Two initially present gliders collide to make the stationary part of the RCT and exactly one escaping glider.
(2) Later, third initially present glider collides with the escaping glider from the first collision to make an approaching spaceship (with speed equal to 1/3 of the speed of the glider) and exactly one added glider that would bounce between the approaching spaceship and the stationary part.

Probably such a rule exists (and would be possible to find) in the rulespace of isotropic Golly ruletables, since there are so many rules and so many degrees of freedom. OTOH it does not seem very likely that there is an INT rule where this would work.

Two gliders would not suffice since there are only finitely many 2G collisions.

[pipsqueek]

how was the rule for this langtons ant simulator found?

Code: Select all

x = 4, y = 2, rule = B3-cjkq4knqy5cnr7/S02-cn3-ak4eiq5kq6e7e
b3o$o2bo!

or what about irregular replicators?

Code: Select all

x = 2, y = 4, rule = B2e3ijn6n/S1c2cek3n
o$bo$bo$2o!

or this SMOS?

Code: Select all

x = 2, y = 3, rule = B2k3-ejny4knqtw5cjny6en7c/S01e2aen3ijnry4ikqtyz5-cekr6ckn7c
2o2$bo!

how could I find cool/weird things in non-totalistic rules?

[silversmith]

how was the rule for this langtons ant simulator found?

Code: Select all

x = 4, y = 2, rule = B3-cjkq4knqy5cnr7/S02-cn3-ak4eiq5kq6e7e
b3o$o2bo!

or what about irregular replicators?

Code: Select all

x = 2, y = 4, rule = B2e3ijn6n/S1c2cek3n
o$bo$bo$2o!

The way I found those two patterns was by starting with an idea of exactly what behavior I wanted(eg. the Langton's Ant), and finding surprises along the way(eg. the irregular replicator).

I knew a Langton's Ant pattern would need a way to store the state of each "metacell" to know which way to turn. I chose the simplest possible option, where a dot represents a "right" turn and an empty space would represent a "left" turn. I also chose to use a R-pentomino as the base pattern, for no particular reason other than a guess that B3ai would be easier to work with.

First, I rulegolfed by hand until I found a rule where the R-pentomino moved forward and rotated. Then modified the rule to make it leave a dot where it started. This was the "left" turn.

The right turn was found by using EPE to find a rule where the R-pentomino would turn right if there was a dot in front of it, with the same spacing. After letting the search go for a while, it eventually found the result above.

The irregular replicator was something I found by chance while rulegolfing a different project.

My general suggestion is to start from scratch (I think there are enough B3/S23 variants). Have an idea of exactly what core features a rule should have, like a ship, oscillator, weird symmetry, or whatever else you want. Then keep adding transitions for a long while, exploring each rule as you go. Finally, after you have a rule with many different pattern you like, and you can't change the rule without breaking them, then it's complete. Afterwards you share your results, and start running some searches.

You will probably have something unlike what anyone else has made, and has a decent chance of being interesting enough for you or someone else to enjoy for a while.

or this SMOS?

Code: Select all

x = 2, y = 3, rule = B2k3-ejny4knqtw5cjny6en7c/S01e2aen3ijnry4ikqtyz5-cekr6ckn7c
2o2$bo!

EnumPattEvo I believe.

[yujh]

I’m pretty sure the last one is found as a part of a search for sssss by the sssss script

[pzq_alex]

How cheap (in terms of glider synthesis cost) can be a RCT, if you can choose any isotropic rule that is both fertile and mortal, and then you can choose (to be used in a synthesis) any glider that works in the chosen rule? What if Golly ruletables are allowed as well?

Three.

As far as I understand, a three-glider RCT would be something like this?

(1) Two initially present gliders collide to make the stationary part of the RCT and exactly one escaping glider.
(2) Later, third initially present glider collides with the escaping glider from the first collision to make an approaching spaceship (with speed equal to 1/3 of the speed of the glider) and exactly one added glider that would bounce between the approaching spaceship and the stationary part.

Probably such a rule exists (and would be possible to find) in the rulespace of isotropic Golly ruletables, since there are so many rules and so many degrees of freedom. OTOH it does not seem very likely that there is an INT rule where this would work.

Two gliders would not suffice since there are only finitely many 2G collisions.

You do realize that you could encode extra information in the gliders themselves, right?

[BokaBB]

Hi there. Is there any way you could possibly make a rule with more than 256 states? Which kinds of edits would it require to Golly, LifeViewer and such? Is it arbitrarily chosen or not?
Good morning/afternoon/evening/night and sweet dreams!
Enjoy your life and all the best! May THE LORD help you!
Have a suuuuperior day (and this all also to your own)!

BokaBB

[muzik]

Can an odd-period oscillator exist in B345/S5?

[BokaBB]

Hey ya. Is there a way for a rule without small orthogonal spaceships to have Orthogonids and for one without diagonal ones to have Demonoids? Thanks in advance!
Good morning/afternoon/evening/night and sweet dreams!
Enjoy your life and all the best! May THE LORD help you!
Have a suuuuperior day (and this all also to your own)!

BokaBB

[rutabaga]

Code: Select all

x = 6, y = 6, rule = B345j/S23e67
obo$o2bo$o2b2o$b2o2bo$2bo$3b3o!

Has anyone seen this p10 flipper oscillator in this rule? I'm assuming so, but wanted to make sure, and also to point it out because it's really cool.

[Period1GliderGun]

Code: Select all

x = 6, y = 6, rule = B345j/S23e67
obo$o2bo$o2b2o$b2o2bo$2bo$3b3o!

Has anyone seen this p10 flipper oscillator in this rule? I'm assuming so, but wanted to make sure, and also to point it out because it's really cool.

https://catagolue.hatsya.com/object/xp1 ... 11/b34s267

[squareroot12621]

Is there a rule such that two spaceships, A and B, can be moved and/or phase-shifted to create 0, 1, and 2 of the spaceship C? (Bonus points if it's within 5 transitions of CGoL.)
A near-example:

Code: Select all

x = 27, y = 27, rule = 12ColorMarkedLife
.pB13.O2.O6.D$pB.pB3.E2.E4.O9.D.D$pB.pB7.E3.O3.O5.3D$pB.pB3.E3.E3.4O6.
D.D$.pB5.4E13.D.D2$4.17pB3.2N$24.N.N$.pB2.D2.D3.2B4.N2.N3.2N$2pB6.D4.
B2.N7.N.N$.pB2.D3.D3.B3.N3.N3.2N$.pB3.4D7.4N$3pB9.B12.2R$10.R13.R$9.R
14.R$9.3R12.R$25.2R$4.17pB2$2pB4.E2.E5.O2.O$2.pB7.E3.O$.pB4.E3.E3.O3.
O$pB6.4E3.4O$3pB$6.R11.R$5.R13.R$5.3R9.3R!

[confocaloid]

Is there a rule such that two spaceships, A and B, can be moved and/or phase-shifted to create 0, 1, and 2 of the spaceship C? (Bonus points if it's within 5 transitions of CGoL.)
A near-example:

Code: Select all

x = 27, y = 27, rule = 12ColorMarkedLife
.pB13.O2.O6.D$pB.pB3.E2.E4.O9.D.D$pB.pB7.E3.O3.O5.3D$pB.pB3.E3.E3.4O6.
D.D$.pB5.4E13.D.D2$4.17pB3.2N$24.N.N$.pB2.D2.D3.2B4.N2.N3.2N$2pB6.D4.
B2.N7.N.N$.pB2.D3.D3.B3.N3.N3.2N$.pB3.4D7.4N$3pB9.B12.2R$10.R13.R$9.R
14.R$9.3R12.R$25.2R$4.17pB2$2pB4.E2.E5.O2.O$2.pB7.E3.O$.pB4.E3.E3.O3.
O$pB6.4E3.4O$3pB$6.R11.R$5.R13.R$5.3R9.3R!

I tried to find a rule where this becomes an actual example:

Code: Select all

#C LWSS + LWSS => 2G
#C [[ ZOOM 8 THEME LifeHistory ]]
x = 19, y = 4, rule = B36k/S234q
o2bo11bo2bo$4bo9bo$o3bo9bo3bo$b4o9b4o!

Code: Select all

#C Advance one LWSS, and you get 1G
#C [[ ZOOM 8 THEME LifeHistory ]]
x = 18, y = 5, rule = B36k/S234q
o2bo$4bo9b2o$o3bo8b2ob2o$b4o9b4o$15b2o!

Code: Select all

#C Advance again, and you get 0G
#C [[ ZOOM 8 THEME LifeHistory ]]
x = 18, y = 5, rule = B36k/S234q
o2bo$4bo8b4o$o3bo8bo3bo$b4o8bo$14bo2bo!

(2TFL, toggled B6kS4q)

[yujh]

What is maximum period that an oscillator with a bounding box no larger than 4x4 can have? The best i know is 64.

[pipsqueek]

What is maximum period that an oscillator with a bounding box no larger than 4x4 can have? The best i know is 64.

Logically, it should be 2^(4*4) = 2^16 = 65536. but that is only the best you could do (in a non-isotropic rule. in an isotropic rule, there would be symmetry restrictions) there may not be a p65536 in a 4x4 bounding box (which is very likely to be true)

that is assuming it stays within that bounding box. you could probably use LLS to figure that out.

and out of curiosity, can I have the RLE for that oscillator?

[muzik]

and out of curiosity, can I have the RLE for that oscillator?

The oscillator is included in the OmniOsci project as a minimal nrv example: https://catagolue.hatsya.com/object/xp6 ... n4iy5ei6ak

[toroidalet]

I had a long post I was just about to submit calculating an upper bound of 28,088 for the period of a 4*4 oscillator or spaceship in an isotropic rule, but it was lost in an error and I really don't want to repeat all of that work. The result was that there were 62,120 asymmetric patterns in a 4*4 box, comprised of 5 unique asymmetric 2*3 patterns, 28 unique asymmetric 2*4 patterns, 22 unique asymmetric 3*3 patterns, 591 unique asymmetric 3*4 patterns and 6,381 unique asymmetric 4*4 patterns. 5 of those could not be the phase of any oscillator:

Code: Select all

x = 67, y = 5, rule = B3/S23
27bo21b2o13bo$3bo$o2bo12bo13bo$19bo10bo18bo2bo10bo2bo$19bo!

So the maximum possible period is 4*(5+28+22+591+6381-5)=28,088. It seems unlikely that an oscillator could exist at that period, but it wouldn't surprise me if there was one with a period over 10 thousand.

Anyway, LLS found this p68 very quickly, so we could probably find a much higher period:

Code: Select all

x = 4, y = 4, rule = B2cn3-ceir4-eqryz5-ckqy6ace7/S02cen3-cer4-ceqt5cei6-k7e8
b3o$bobo$o2bo$2b2o!

There is some ambiguity with the question about whether the pattern stays within a specific 4*4 box, or if the box can move around, but even with the stricter requirement, we could get some very high periods (probably 200). One way to find this would be to have a depth-first search where it starts from some pattern and sees what 4*4 patterns it can reach from there, repeating until it finds a pattern it has already visited.