Thread For Your Unrecognised CA

[iddi01]

live3 is a rule i created a long time ago to get the rotating block, but recently, i found out that it's yet another GOBWR rule:

Code: Select all

x = 215, y = 87, rule = live3
3$42.BA$28.2A12.C$28.2A24$208.A$204.2A2.A$204.2A2.A$208.A2$37.A$37.A$
5.2A30.A43.2A$5.2A30.A43.2A15$37.A$37.A43.2A$37.A43.2A$37.A2$5.2A$5.
2A21$26.A$26.BC!

It have many useful spaceship-based reactions though.

Also, a b2a-type constructional rule:

Code: Select all

x = 335, y = 105, rule = B2an3r4i/S1c2n
7$58b2o$58b2o41$27b2o196b2o100bo$226bo101bo$329bo14$6bo$6bo23$59bo$60b
o76bo$61bo75bo2$112b2o26b2o$86bo4bobo$84bobo4bo18bo$61bo48bo$60bo$59bo
!

"WormLoops" in a 3-state rotate4reflect rule.

SoManyShips-ized variant of this rule:

Wow, 13 likes: this almost beats the record!

Like i suggested before, you really need a thread for your 3-state rules. There are numerous threads for rules far less interesting than yours.

I'm feeling uncomfortable whenever i see there's no thread for rules this good. If there's still no thread for those after a few days, i might even create one myself!

Code: Select all

x = 109, y = 58, rule = B1LifeLife
75.5A$74.A4.A$79.A$78.A8.4A$86.A3.A$90.A$89.A3$87.A$88.A$88.A$86.3A3$
.2A$2A$.A85.4A$86.A3.A$90.A$89.A8$87.B3.B$88.4A$87.A3.A$91.A$86.BA2.A
B2$86.BA2.AB$91.A$87.A3.A$88.4A$87.B3.B11$97.B2.B$98.A$98.A.A$86.2A5.
B5.A.A6.B$86.2A4.B6.A2.AB2.2A$88.B3.B.B4.A.A4.2A$93.B4.A.A4.B2.B$98.A
$98.B!
@RULE B1LifeLife
@TABLE
n_states:3
neighborhood:Moore
symmetries:permute
var a = {0,1,2}
var a1 = a
var a2 = a
var a3 = a
var a4 = a
var a5 = a
var a6 = a
var a7 = a
var a8 = a
var b = {0,2}
var b1 = b
var b2 = b
var b3 = b
var b4 = b
var b5 = b
0,1,0,0,0,0,0,0,0,2
b,2,2,2,0,0,0,0,0,2
a,1,1,1,b,b1,b2,b3,b4,1
1,1,1,b,b1,b2,b3,b4,b5,1
2,2,2,0,0,0,0,0,0,2
a,a1,a2,a3,a4,a5,a6,a7,a8,0
@COLORS
1 255 255 255
2 128 128 128

That's a variant of Brew (not putting a link because you posted in the Brew thread just today)

[confocaloid]

If there's still no thread for those after a few days, i might even create one myself!

Not everything needs a separate thread. I think it's fine to post new rules here, in "Thread For Your Unrecognised CA". There are many CA rules, including many interesting CA rules.

It would be an especially bad idea to create a collection thread for someone else's rules. If they do need/want a collection thread for all their CA rules, they will create it.

[iddi01]

It would be an especially bad idea to create a thread for someone else's rules. If they do need/want a thread they will create it.

It has been done several times already.

[confocaloid]

It has been done several times already. (i remember a third one but i coudn't find it)

Those are threads about specific CA rules, rather than collection threads for all rules by a certain person.

Basically, when someone needs/wants a collection thread for their rules, let them have control over the first post in that thread.
If they do not need/do not want a collection thread for their CA, that means the thread is not needed.

[b-engine]

Code: Select all

x = 109, y = 58, rule = B1LifeLife
75.5A$74.A4.A$79.A$78.A8.4A$86.A3.A$90.A$89.A3$87.A$88.A$88.A$86.3A3$
.2A$2A$.A85.4A$86.A3.A$90.A$89.A8$87.B3.B$88.4A$87.A3.A$91.A$86.BA2.A
B2$86.BA2.AB$91.A$87.A3.A$88.4A$87.B3.B11$97.B2.B$98.A$98.A.A$86.2A5.
B5.A.A6.B$86.2A4.B6.A2.AB2.2A$88.B3.B.B4.A.A4.2A$93.B4.A.A4.B2.B$98.A
$98.B!

That's a variant of Brew (not putting a link because you posted in the Brew thread just today)

Not quite, as state 2 cells in B1LifeLife are born when they have 1 state 1 neighbor, while state 2 cells in Brew are born when state 1 cells die.
Also a natural puffer:

Code: Select all

x = 94, y = 78, rule = live3
B2C.B.B.ABAB3.B2CA.BC.3A2.CB.2A5.BA4.BA3.B2.2A.A.A2BA.B.ABCB3C.3A3.2B
.C2B.2C.C.A2.A$2.3C2.AC6.ACA.B2.CA2.B.A2.ACBA5.C.C.2B2C.B2.A.A2BA2.C.
BA2.C.B.2C.CA2.B.A2.B.CB3.AB.CA$.C2.A3.A.C.A2B.A.2A.ACACB3A.A3.B6.3AC
.BCAB5.2C4.CB.BA4.C.C.C4.BC3.2CA.A3.B$2.2AC3.BAB.C.B.2CB.A2.3A.A.CAC2B
CAB4.C3.B.CBC3.BCB.C2.B4.AB2.AB2CB4.B.B.A3.CB2.2B.B$3B2.2C.ACA.2A.AC.
B5.C3.B.B4.A4.2A.B2.A6.2C.B4.B3.C4.CABCA2.A2.C4.2B2.A$.A2.BA2.B4.CAC.
B4.A.BC.A.B.B.AB.C2.2B.A2.CA2.AB2.B2.A2CA8.B6.ACB2A.AB2.A.2B4.A$2.BA4.
B6.BA.2A3.3ABA3.A.AB5.C.A.2A2.C3.BA2CA.CA2C.AC3.A3.C2.CA2C.B2A.CACB.B
2.B$3.C.AB3.2A.2AB5.B3.C.CA3.A2.B.C3.BA2C2.C.A.ABC.A.CB.ACAB.B4.3AB.B
C2.AB.2ABA.2A2.BCB$.B2A.C13.B2A2.2A3.C3.3B6.B.A2.B2.BCB5.C3.A6.2BA.BC
2.AC3B.2AB.C2.C.B$A.BC.ABACB.C3.AB.A2.ABAC.2C.A2.2C2.B.ABCB2AB2ACBA2C
2.C.B.C2BC2B.BC.C2.C.AB2.BA2.B2.A3.B3C.A$3.C3B.B.C.C.B3.2A.2B2.C2A.AB
3.B3.B.AB2.B.A.A4.B3.B2.CBA2.B2.BC2.A.AB2.AB.3A2B.C3.B.C$A2.CA.2A.C2.
BCA2.BA2.C3.BCB6.B2.C.2ABC2BA3.C3.2A.2B.ACBCAC.C.C.AB7.BAB4.3BC3.A$2A
C4.2B5.A.B2.CBC4.B.2C.3A6.CB.C.A2.CAB2.A.2A4.AB.C2B2A3.B.A2.3A2BACB.B
.A3.A$A2.C3.B3.C10.3C2BC5.A3.BC5.B4.2C2ABCAB.CA2.BCB.AC4.AC5.B2.C2B.2B
C$3.B.C7.B5.BA3.C.3C2.C4.B.B.BCBA.B2.AB.B2.B2.A.2AB4.A2.B.A.AB.B.AC.C
2.C2.B2.AC.B$.4C6.A4.BA.B.CA.B3.2BA9.BC.C5.B.A.BCB.A.ABA3.CA.ACB2AC.C
.2A.ACA.B2.2A2B.2A$2.2CA.C2.BC2BAC.2C3.AC2.A2.B2.CB2.A4.A.B5.AB9.B4.C
B.CA2.B.2A.A.CABA.BC.A.BC2.A$.2A.C5.C3AC.C.B.C2.A3.BC.CA.BAC2.A.2A.C.
B5.2C2.A.2B.A3.A2CB3.B3.2B2.2B2.C$B3.2A4.A2B4.C.2B.2A.A2.C.B.A.A3B11.
B.BA3.2C3.B.C5.BAC.2CA.B.ACA.C.B2.2A3.A$C.2B.AB3A.B.CB.A2.C.C.B2C.B2.
2A.B2CBA.AC.3C.BA2.C.C3.A2B3.ACB2.A.A2C.C2.A.C.A2.AB.CA.B.C$2.CB2.C6.
BA3.CB.BC.BC2.B2.A.BAC.B3.2A2.C2.B.CAC.B.B.CB2.BA3.B9.2CA3.2CAB.B2.B.
A$.A2BC3.BC.A.A3.2BC.AB3.AC5.CB.B.B6.2C.B3.2B5.C5.C2.A.B.CAB2C.A.CB3.
BCA.2A2.C$2.ACB.A4.C.A2.C4.BA2.B7.2B2.BCBA3.BC2.ACAC.AB.BC.B2.BC2.A.B
C.CB.BAC7.2B3.2ABC$2.B.B2.C.B3.C5.2C2.CA.CA3.C3.C.C2ABA4.BA.B.BC.BCA6.
C.A2.2BC.A3.3BA2.A.C2.C2BCB2C$A2.BA2.C.A.CBC.C.B2.AC2BC.B3.AB4.B7.B2.
C3.A4.C.2CB3.C.A.A3.A3.B3.BACBC2.C3.BA$2B.C2.C2.C.C3.A2C7.B.A2C.B.2C3.
CB.BC3.2C4.A3.B.3C.B.B2.A.A.2AB.CB2.ACB.B.2ABCA.AB$.BABC.BA.C.C.BCA3.
2B3.A2BC.A2.C2.2A2.A.ABC.B.2C2.A4CA.CBAC.2B2.2B.A.C.B.C2.A.A.B3.2A.B.
B$.B3.2BC5.B.B.AC4.ABC.B3.B.C2A3.AB.C.CB2.A2.A3.C3.B2.B.2C.2A3.A3.CAC
4.AC4.BA2.C$ACBA2.A2.B3.AB.B2.C.A2.B3.2A2.C.A.A2B.CB.B7.B.2C2.2CA2.B.
A.B.B.2B.B.B.B4.CB3A2C.AB.C$2.BA.AC6.B.A.BA.A3.BA.C7.AC2.B3.2CBAC3.BA
.B2.2A.A2C.C.CB2.C2AB2C.CA.2A2.AC.C2.CBCA$C.C2.C4.B.CBABA2B2.B3.A.BC.
ACAC.BC2.B.CBCBAB2.A.C4.C2.BA.2C2.BC2.A4.A.2A2BA3C.BCA.BCBA$CAC.A.B2A
4.CBC2.AC.A.2CB3.C3.C2.CA.B.C.ABA.CAC.B3.CA2.C2.B2.A.ABA4.BA2B.C4B2.A
.B.B2C$A3.BAC.CB.A.B2.BCA.2B2.CB3.CB2.B.C7.ACBC4.A.AB2.2CA.A5.A.B.AC4.
BC3.2B.B.2C$5.2BC2.C2B2.2A.BC2.B7.AB.C2BC.C.3B.C.C2.A2.B.B6.2BA.AC.BA
.2A3.B.CA.C4.A.B2.2A$B4.AB.C.2C.B2.3C2.B.C2.BA.BA3BC.A3.BA4.CA.C.B.2C
2.2A2CA2BA.2AB2.2B2.A.CBA3.A3C2BC$B.CABC2.3A3.B3C.AC.C.AC3.A.B.CA2.2A
2.CB2.CACB.A.B3A.C2.B3.B3.BA.C.C.A2.CAB.2B.CB.2C$2.A3.B2.C2A4.BA3.BCB
.AC.BCB.2C.BCBC.CB.B5.A.BAB.4C3.B3.3C.B3.C2AC3.2C.CB2C.B2.CB$.B2.C4.A
2C.A.A.CACB2A.BA.A2B3.CAC.2ACACB2.CA3.AC2.BA.B.CBCB2.2CA.C2.B.C.B2A.A
.CA.B2.B3.C$CA2.B.B.C.BA2.AB.BAB2.A5.B3.A.C.B2.CBA2.2C3.C.C2.CB.AC2.A
.A.ACB.C2B.A2C.C.BCB2.2AB3.C2.A$CBCB2AB.A.2B2.A2.AB2.CB.BCBCB.2A.A4.B
2.ABC2.CB2C3A.C.C.B.2CA.A2BA4.A3.B.C10.A.B.B$5.B5.BC.CB3.2C.B2.A.C.B.
CBC5.A2.CA5.ACA2C2.B2.B2.A.C3.B2.B3.2A.2B3A.B4.A2.B$B2CB2.B2CA.C.C.C3.
2B.2AC2.BA2.2CB3.A.B.A6.B3.2C.A5.B3.A2.CB.2A4.B.C2.B.2C.B4.CA$CB2.2AB
4.B.A2.C.A.2CA2.2ACB2.B2.A2.2A4.2C2AC8.BA2.CB2.C.A.C3.A2C.A.A2.2ACAC.
C2B3C$4.C.A.B.A.C4.B.B.A2.C.A.A3.B.A.BC2.BC.C.A4.A.BABC3.B.CAC2A.B.B2.
2AC2.2A.B4.B.B.A2.2B$.C2.B5.ABA4.2C2A.C2.4C.C.CA2B2CB2.AC2.A4.BC3.CBA
2.A2.B3.AB.B.B.2CB.BA2.BCB2.2A2C.2C$2B.C.CB2C.B.AB.2ABC.C.2A2C2.ACB2.
C2A.B.B9.CBA3B3.2A5.2C.CA.A2BA.CB3.CB2.B$CB2.C4.CA.C.3BCB4.ACACB2.B2.
A.BAC.C2.2C.C6.CA2.ACA2B.BC2.AB3.C.C.2C2.2C2.C2.C5.A$.A.CAC2.C.AB.4C3.
BC4.3AB3.2C.BAC.A.C3.B.BC.A.2B2.CA3.2B2.B3.C3.C7.B.B.2BC3.AB$C2A.2BC2.
C5.A8.ABCA.B.C.CA2.A.2B.A.B.A2.BCB.B.AC.A2CBC.A2CB.ACA2CB.C2.AB.BCAB2.
CB3.B$3BC2.AB3.ACBA2.B2.C3.A3.ABCA.C.B4.A2CAB3.CB2.C2A.2A2.C.A.BAC.C3A
4.C.BA4.CA3CB2.BC$C.B.CBCA4.A6.B2A.CBC2B.A.2AB.B2.B3.C.A.B.BCBC2.C.C.
B2.AB.A7.BACA2.B.A.AB.2B3.C2AC$2.A.B.B2C2.CA2.CBC.CA.BA.BA.BC.B4.B.BC
2.CA4.A2BAC.AC.BAC2.BA2.B.C2.AB2CAB.B2A.BA2.C.B2.B$.C.C.C.B.BA3.2C.BC
.A.C.C2A.A.ABC.A.AC.C2.A.A.BC3.C.B.AC.A.B4.A3BA.A.B3.BA3.2B.CA.C3.C$.
C.C2.ABAB.2B.2BA.2C3.CBC.AC.C2B.CBC2B.B2.B.ABA2C.B4.A6.C.2B2A.AC.BCBA
2B3.2BAC2.2A.BA$3.C12.AB.A.C.B.B2.CA.AB.B2.2CA2.2A.A.BAC.2A4.C5.C3.3C
A.CAC.A.C.B7.A.A2.B$.A.3ABA2.3B2.B.A2.A.BAB.B.A7.CA2.BC.AB.C2.A2.AC2B
2.A.CBA2CB3.A.CA.C.A.B.A.2B3.CAC2.BC$B2.3B.B2.AB2.3CA.CB2.C.A.A5.B2.A
.A.AB2.B3.A.BA.ABC.A.A4.C2.AB5.C2.2BA.2A2BC.CACA.AC$B.A.BA2.BA2.2A2BA
2CB3.2A2.BC.CBA5.2B3.B4.A2C.A2.B.C.3C3.A7.ACB2.C2.B4.C3.BA.B$B3.A2.2B
AB.AC.C6.A.C6.BCA.B2.2ABCB2.2CA2.B6.BCA.B3.A3.CBA3.B2.AC.2CB.A2CB.A$.
2B3.2CAB.A3.A2.CB2.B.CB2.BC.BA2.3AB3.CB.3C.C.A.C.A3.CBC2.C3.CBAB.BC3B
.C.A3.BA.BA2.C$A2.BCB.C2.A3.A.AC2.B3.2C3.A.C4B.CAB.A6.B2C.A2.2C2.B.2C
AB.BC2.C5.C5.BC.A.C2B2.2C$CB2.2C.AC.C2.A.B.A5.2B3.2B.2A.2A.B.2B.CACB3A
C3.2A.2B2.BAC.A3.2ABC.A.C.B3.B.A2.CB3.C$A6.AC.A4.2A4.B5.B.AB6.2AC6.C.
B.2C.B2.A5.C.C4.B.2C2.BA2C2.BC5.BAC2.A$C.A2.A.A2.A.C2.BCA.2BC4.AB.ABC
A.B.B.B2A2C2.C.3B.2A.B.ABC2.B.ACA2.B4.BC.ABA.C.CABC2.CB3.B$2ABAC3.CB2.
2A2.ABA.A2.B2.B3.4C3.BCA2.C.2A2.C.B2.ACBC.2AC.B.C2BC.CA2.C4.CAB4.B6.2C
$3.B.B.2B3.C.2A.CBCB.C3.CA2.A.C.C.B2.C.A3.B.A2B2.CBABC.A2C.C.C3.CAB.B
.A4.AB2AB2A.B.C$CA.CA3.2CA2B3.BA2.2AC2.CBC.CBC2.C2.A.A.C.B.C.A3.A6.A2.
B2C2A6.C.2CABC3.C.C5.C$A.BA.BC6.A2.A.B2.AC3.B2.B.A.CB.ACA2.2A2.A.AC4.
AB.B2C.CB.A.CA.A4.C.C.2BA2.A.2ABA.C.ACA$.A.C.BC.BC2.AB3.BA5.A.B.B2.A3.
A2.AC.A3.AC2B.B3.B4.CA.C.CB2.A.A.BA2.B3AC.2B2.C.A$2.A.AB2.C2B.B3A4.A.
AB2A.2C3.BA.C2.C2.CB2C2.B.A3.B.C.C.C.A3.C.2B3.C2.2C2.C3.B3.2B2A2BC$3.
B2.ACA3.C.BC2.CA4.2CAC.C2.2C.A2.C.ABABA.C3.C.AC.AB.A2.B2.C.A.A.ACABCB
A.A2.B.C.A2.A2.A2C$C.2A2C.A2.B.A.2C.B.C2.2A2.ACABC2.A.B3.C.B.A4.C.CA.
A.C5.B.AB2C2.C2.A.BA.CAB3.4C.2AC2AC$2.CA4.CACB.A.C3.B.ACA.CBA2.A2.C2.
AB.CB2CBA2.C.BA2B2.3A2.AC2A2.A.C.A.ABAC.ABC.BA2C.AC.AB.B$.C2.BCACAC2.
C3.B.C2.B.A3.B3CA.B2.C.A.A4.BA4.A.2AC7.2C6.CBCAB4.A.AC2.AC2.B2.C$2.A.
BA2.AC2.C2.2A.B3.C2.A.BC.CB2A.C2AC2.BA.CA.3C.2A.2C3.2A2.BA3.C.C3.B.A3.
A2B.C.AB2C.A$CBA.A.C2.CB.B2.C2AC2.B6.C2.B2.CA.2C3.3C2.2CBC.A2.B3.BCAB
C.2BC2.B2.B.B.BC3.2A.3B3AC.C$.B.2B7.A2B2.B.B.CBA7.B2.AB2.A.C2.B2.B2C3.
BC2.B2.2B.A3.2C.C2.C.2B2.A.A.2A3.B2.C$.B.2C.A2.AC2.CB.A.ABA2.B.BCA5.B
C.B.2CB2A2.B5.2BAB.C6.A.C2.CBA.A.2C2A.B.A3.B3.BA2CB!

[Supersuthiastic]

I was recently having lots of fun with the rule B4568/S123457 and It has a pattern which lives for 1941229 gens before stabilizing into a p9 oscillator. Here is the pattern below:

Code: Select all

x = 11, y = 11, rule = B4568/S123457
11o$o9bo$o9bo$o9bo$o9bo$o4b2o3bo$o3b2o4bo$o4bo4bo$o9bo$o9bo$11o!

Can you find longer ones?
I can find it, and it lasts for something like 40 million gens before stabilizing:

Code: Select all

x = 11, y = 12, rule = B4568/S123457
11o$o9bo$o9bo$o9bo$o9bo$o9bo$o4b2o3bo$o3b2o4bo$o4bo4bo$o9bo$o9bo$11o!

My 3rd post!

[b-engine]

Can you find longer ones?

(I can modify quotes in order to shorten it.)

Welcome to the forums!

The lifespan of a soup seems to grow with the size of soup.

This behaves like replicator:

Code: Select all

x = 143, y = 4, rule = B4568/S123457
143o$o62bo78bo$o62bo78bo$143o!

This isn't much explored as spaceships cannot exist without B0, B2 and B3 and with B1. Plus, the present day is the era of Isotropic non-totalistic rules. Most Life-like rules are already far explored, which makes new Life-like rules rarely exists in this forums, sadly.

[confocaloid]

Life-like rules are already outdated, sadly.

What do you mean "outdated"? I fail to see how Conway's Life (B3/S23) or Pedestrian Life (B38/S23) or Seeds (B2/S) or 2x2 (B36/S125) or any other explored Life-like CA could possibly be "outdated". Newer discoveries make things more interesting, not less.

[tommyaweosme]

What do you mean "outdated"? I fail to see how Conway's Life (B3/S23) or Pedestrian Life (B38/S23) or Seeds (B2/S) or 2x2 (B36/S125) or any other explored Life-like CA could possibly be "outdated". Newer discoveries make things more interesting, not less.

there is only like 2^16 possibilities for life-like rules, meanwhile isotropic has like 2^102

[confocaloid]

there is only like 2^16 possibilities for life-like rules, meanwhile isotropic has like 2^102

Having fewer possibilities may be viewed as an advantage. Definitions of Life-like cellular automata are simpler (==> easier to understand and investigate the rulespace; no need to learn Hensel notation to be able to do rulegolfing). Still, there are enough degrees of freedom to have curiosities like Day & Night or 2x2 (for example).

Yes, Hensel notation allows to define more rules. But very many of resulting rulestrings look and feel arbitrary to someone who meets a rule for the first time. How on earth would someone arrive at a rulestring such as B2ci3ai4c8/S02ae3eijkq4iz5ar6i7e? Where all these letters and digits came from? How to navigate in this rulespace? That is possible, but harder. And definitely quite hard for newcomers.

[...]
Can you find longer ones?
I can find it, and it lasts for something like 40 million gens before stabilizing:

Code: Select all

x = 11, y = 12, rule = B4568/S123457
11o$o9bo$o9bo$o9bo$o9bo$o9bo$o4b2o3bo$o3b2o4bo$o4bo4bo$o9bo$o9bo$11o!

[...]

My guess is that it will not be very hard to find seeds with higher lifespans, but most of them will probably settle into low-period oscilllators.
Here is a p260 oscillator that occurred twice over three of my attempts. Are there higher periods in the same bounding box?

Code: Select all

x = 11, y = 11, rule = B4568/S123457
11o$ob9o$11o$11o$11o$11o$11o$11o$11o$9obo$11o!

[H. H. P. M. P. Cole]

Having fewer possibilities may be viewed as an advantage. Definitions of Life-like cellular automata are simpler (==> easier to understand and investigate the rulespace; no need to learn Hensel notation to be able to do rulegolfing). Still, there are enough degrees of freedom to have curiosities like Day & Night or 2x2 (for example).

I second that comment. Speaking of which, which rulespaces are small enough to have at least a few interesting rules? The smallest I found is the 2^6-large rulespace of totalistic R2 1D rules, and the only 'interesting' rule has totalistic code 24 (if you want to run this in LifeViewer use the string R2,C2,S1,3,B2,4,N@003C00). Here is the catagolue census for this rule. Are there any more with sizes less than 2^18?

I define 'interesting' very simply: a rule is interesting if any sufficiently large random pattern, can, in principle, expand their bounding dimensions and is non-explosive (for some definition of non-explosive). Others may have different definitions.

Yes, Hensel notation allows to define more rules. But very many of resulting rulestrings look and feel arbitrary to someone who meets a rule for the first time.

Hensel notation looks arbitrary, but not as arbitrary as non-isotropic, or MAP notation. However, if one wants a notation that is a little less arbitrary, try HROT. Try reading about Bosco's Rule, one of the first HROT rules to be explored in-depth.

With all due respect, Hensel notation is not arbitrary - that is, if one learns the effect of each transition in the isotropic non-totalistic (INT) rulespace, and practice rulegolfing in it. Sometimes, adding/subtracting conditions can give rules with very different types of objects. Take this p62 statorless oscillator, discovered by confocaloid:

Code: Select all

x = 24, y = 24, rule = B3/S234c
9bo$8b3o$7b2o2bo$9b3o4$10b2o9bo$10b2o9b2o$20bob2o$15b2o3bobo$15b2o3b2o$2b2o3b
2o$bobo3b2o$2obo$b2o9b2o$2bo9b2o4$12b3o$12bo2b2o$13b3o$14bo!

This does not work in Life, despite looking like it works in Life. The INT rulespace is so large that tools like EnumPattEvo2 were created to help navigate through this rulespace easily. (Also, FWKnightship is doing some seminal work on 3-state INT rules.)

Of course, such metrics of arbitrariness are arbitrary themselves, and the metric of arbitrariness is dependent on whoever is creating the metric of arbitrariness.

I would invite anyone who has at least some experience in rulegolfing to join in.

[tommyaweosme]

i has experience in rulegolfing
an example: misshaped honeyfarm life gliderless variation b35n6ei/s235e6ei
cool pi puffer

Code: Select all

x = 0, y = 0, rule = b35n6ei/s235e6ei
3o$2bo$3o!

[confocaloid]

Correction regarding the comparison between B3/S234c and B3/S23: all four transitions (birth 0 -> 1, survival 1 -> 1, absence of birth 0 -> 0 and death 1 -> 0) are possible, in each of the two cellular automata. The difference between two CA is in the survival condition S4c. In B3/S234c, an alive cell with four alive diagonal neighbours and zero alive orthogonal neighbours survives to the next generation. In B3/S23, it dies in that case.

[tommyaweosme]

Correction regarding the comparison between B3/S234c and B3/S23: all four transitions (birth 0 -> 1, survival 1 -> 1, absence of birth 0 -> 0 and death 1 -> 0) are possible, in each of the two cellular automata. The difference between two CA is in the survival condition S4c. In B3/S234c, an alive cell with four alive diagonal neighbours and zero alive orthogonal neighbours survives to the next generation. In B3/S23, it dies in that case.

that one, simple s4c can make so much difference. any addition or subtraction to any isotropic rule makes close to infinite difference. there is infinite uniqueness in isotropic, so much that we will never run out, but the magic of totalistic rules wore out a long time ago, we discovered all there is to discover in that small, useless rulespace. a few good things came out of it, but those good things could also be found through isotropic rules. i believe that isotropic rules should have been where we started in the first place, because it is so big and amazing, the hardworking people in the 1970s could have found lots and lots of more cool things.

[confocaloid]

small, useless rulespace

It's unfortunate that you somehow managed to arrive at that conclusion. I stopped reading your post when I saw that, because it's ridiculous.
Even in plain Conway's Life (B3/S23), every year there are new discoveries. And many Life-like cellular automata remain nearly entirely unexplored, or at very least underexplored. I mean B34/S34, B36/S23, B38/S238, B356/S23, B2/S0, B345/S5, B3678/S34678, and many others.

Hence strictly speaking it's unnecessary to bother with non-totalistic CA or higher ranges; people will never run out of discoveries in Life-like cellular automata.
Of course non-totalistic CA are interesting by themselves. They allow things impossible in Life-like CA. But that doesn't make Life-like CA any less interesting.
Otherwise, someone could respond to your post by rejecting R1 Moore INT rulespace, saying "look, MAP rulesets allow to toggle any out of 2^9 = 512 distinguishable 3x3 conditions selectively; surely they must be more interesting, so why bother with Hensel notation?"

Life-like cellular automata are simple. The universe is simple (it's an Euclidean plane tiled with squares all the way), the neighbourhood is simple (range-1 Moore), the ruleset is simple (neighbour-counting conditions for birth and survival).
More flexibility means more complexity. Already Hensel notation has to be learned. A beginner doesn't immediately know what 'B35k/S234j' means, even when they know what 'B3/S23' means.

[tommyaweosme]

i get your point. im sorry for posting that. i was saying that conways game of life is b3cekainjr/s2cekain3cekainjr in hensel notation, and that every lifelike rule could be written in hensel notation, but hensel notation has more unique things in it.

Otherwise, someone could respond to your post by rejecting R1 Moore INT rulespace, saying "look, MAP rulesets allow to toggle any out of 2^9 = 512 distinguishable 3x3 conditions selectively; surely they must be more interesting, so why bother with Hensel notation?"

READ ALL THE WAY THROUGH AT THE END I REALIZE MY MISTAKE

that "repsonse" i agree with too. i might start making map rules instead of int rules. map rules are the new generation of conways game of life. int rules are so small, but map rules contain not just isotropic and totalistic rules, it also has close to infinite other possibilites. map just seems better. bigger always seems better. and it is. but simpler is better too. we need to stop wanting bigger stuff. i recognize my previous mistake.

[confocaloid]

[...] i might start making map rules instead of int rules. map rules are the new generation of conways game of life. int rules are so small, [...]

You might want to consider extending those to a multistate rulespace. Every non-isotropic two-state range-1 Moore-neighbourhood CA can be implemented using a RuleLoader ruletable. But of course RuleLoader supports way more than that. Hence one could say that ruletables must be more interesting, and MAP is just basically "too small".
On top of that, there are neighbourhoods other than range-1 Moore. Is there a gun in any two-state outer-totalistic CA using the range-1 hexagonal honeycomb neighbourhood? Does the answer change for non-strobing CA?
What is possible in three-state range-1 hexagonal outer-totalistic CA? What is possible with von Neumann neighbourhood instead? I believe these are underexplored.
(edited to clarify)

[iddi01]

Having fewer possibilities may be viewed as an advantage. Definitions of Life-like cellular automata are simpler (==> easier to understand and investigate the rulespace; no need to learn Hensel notation to be able to do rulegolfing). Still, there are enough degrees of freedom to have curiosities like Day & Night or 2x2 (for example).

I think INT rules are a perfect balance of flexibility and simplicity. The main disadvantage on those is having to learn Hensel notation, which isn't too hard for most people. (in fact, Hensel notation is quite easy for me now)

Hensel notation can also be used to make minor changes to OT rules and make them more interesting.

Even in plain Conway's Life (B3/S23), every year there are new discoveries. And many Life-like cellular automata remain nearly entirely unexplored, or at very least underexplored. I mean B34/S34, B36/S23, B38/S238, B356/S23, B2/S0, B345/S5, B3678/S34678, and many others.

Hence strictly speaking it's unnecessary to bother with non-totalistic CA or higher ranges; people will never run out of discoveries in Life-like cellular automata.

There will be much more discoveries in good INT rules if the same time and effort are put into them instead of OT.
For example some INT rules are proven both Turing-complete and omniperiodic not long after being discovered, which are unimaginable for OT.

Also, the first page in OCA forum contains on average 0.3 threads for (specific) OT rules, while above 5 threads for INT. If OT rules are really as good as you claim, then why would this happen?

Most OCA threads I see are INT, not Life-like, mainly because there's a lot more unexplored space in INT.

[confocaloid]

Most visible threads in this subforum are about CA expressible with Hensel notation (two-state isotropic CA using range-1 Moore neighbourhood on the square tiling), merely because

1. A lot of effort is put into investigating this particular rulespace, at the expense of all other rulespaces.
2. There are way way way more CA in this rulespace, compared to many other interesting rulespaces.

It's not because there's little left to be explored elsewhere. Other rulespaces are also unexplored, with an exception of extremely small "toy" rulespaces.

Instead, it is more or less a historical accident of people becoming interested in a rulespace that got a sufficiently readable notation early enough (Hensel notation is significantly less intuitive than just neighbour counts, but still readable).

I think INT rules are a perfect balance of flexibility and simplicity. The main disadvantage on those is having to learn Hensel notation, which isn't too hard for most people. (in fact, Hensel notation is quite easy for me now)

For you, and for other active forum members, maybe.
However, people who post here are a small subset of people who read pages on this website (wiki + forum + ref). And those people are a small subset of people interested in cellular automata.

The Hensel notation is already fairly hard to learn; it's just that people don't bother to register here merely to complain about it.

There is some inevitable complexity when you go to larger rulespaces. Life-like cellular automata are simple enough to have an intuitive rulestring notation, yet interesting enough to warrant in-depth exploration.

Most OCA threads I see are INT, not Life-like, mainly because there's a lot more unexplored space in INT.

Also, the first page in OCA forum contains on average 0.3 threads for (specific) OT rules, while above 5 threads for INT. If OT rules are really as good as you claim, then why would this happen?

There will be much more discoveries in good INT rules if the same time and effort are put into them instead of OT.

Personally, I think there is already a bit too much effort put into INT rulespace, at the expense of other rulespaces.

Having many threads about specific non-totalistic CA, each getting a few replies and then sinking away from the first page, without an obvious explanation why those are interesting specifically, at some point begins to feel like spam.

[H. H. P. M. P. Cole]

Most visible threads in this subforum are about CA expressible with Hensel notation (two-state isotropic CA using range-1 Moore neighbourhood on the square tiling), merely because

1. A lot of effort is put into investigating this particular rulespace, at the expense of all other rulespaces.
2. There are way way way more CA in this rulespace, compared to many other interesting rulespaces.

I second that comment. INT is overrated. Which other rulespaces could we explore in instead?

Offhand I'm thinking about INT Hexagonal, and/or totalistic HROT (i.e. HROT weighted neighbourhoods with weights of only 0 and 1).

[hotdogPi]

Golly doesn't yet support INT hexagonal, which makes it a bit harder to work with.

[confocaloid]

> Which other rulespaces could we explore in instead?
Other neighbourhoods with neighbour-counting conditions also can be interesting, despite their simplicity. What is known about the "knight's neighbourhood" (8 cells reachable by a knight's move)?

Two-state isotropic R1 hexagonal is an interesting rulespace. There are already several known interesting rulesets. The rulestring notation is simpler to learn (just three different letters multiplied by two different neighbour-counts).
Not yet supported in Golly directly, though, and the hex grid display is only supported via Lua overlay rather than natively. LifeViewer does support both the notation and the hex grid display.

[...]
Several interesting isotropic hex rules (which were explored to some extent):

• B2o/S2m34H (Catagolue)
• B2o3m56/S2-p4oH
• B2o45/S2o45H (Catagolue)
• B2/S2-mH (Catagolue)
• B2o3-o4-m56/S2m3o4-o56H (Catagolue)
• B2-m3-o/S2m3-p4-m5H (Catagolue)
• HexInverseFire B2-m4-m56/S2m34m56H (Catagolue)

[...]

[b-engine]

Golly doesn't yet support INT hexagonal, which makes it a bit harder to work with.

Why not using LifeViewer? I primarily use the on-forums LifeViewer to edit patterns. I only recently downloaded Golly a few months ago.

Code: Select all

# Click SHOW IN VIEWER to open LifeViewer
x = 5, y = 4, rule = B2o/S2m34H
o2$4bo$b4o!

[tommyaweosme]

B34e5y8/S234c - tinleylife
natural 2c/6

Code: Select all

x = 10, y = 13, rule = B34e5y8/S234c
2$2b2o$bo3bobo$2b4obo2$2b4obo$bo3bobo$2b2o!

natural p3

Code: Select all

x = 3, y = 5, rule = B34e5y8/S234c
bo$3o2$3o$bo!

pentdecalethon is a lot more common

[b-engine]

B34e5y8/S234c - tinleylife
...
pentdecalethon is a lot more common

Really?

Code: Select all

x = 4, y = 3, rule = B34e5y8/S234c
b2o$o2bo$b2o!