B34a/S2-i3-a45

[KittyTac]

This cellular automaton isn't interesting for its engineering potential (it explodes), but rather for the interesting structure of the "crystal", reminiscent of 1-dimensional cellular automata, with particles interacting and whatnot.

[LaundryPizza03]

Could you please show this property using an example seed?

[KittyTac]

Code: Select all

#CXRLE Pos=-13,-12
x = 8, y = 8, rule = B34a/S2-i3-a45
ob2o2b2o$obob2o$bo3bo$2obo$o2b2obo$bob4o$bo2bo$3obo2bo!

This is a sample seed.

[dvgrn]

This is a sample seed.

Nice find -- the rule, not the particular seed. Seems like this will have a tendency to "go boring" after a very very very long time -- all the chaotic stuff is eventually confined to the four diagonals, and the width of the chaotic region seems to like to drop to ~60 cells diagonally every now and then, but then widens out again.

The narrow sections seem as if they'll act as bottlenecks on novelty generation. If you watch a 250x250 rectangle moving at c/3 to track one of the diagonals, eventually it's bound to repeat, for the usual reason that there are "only" 2^302500 possibilities. When it does repeat, it's hard to see how any interactions from outside the region could save it from falling into a cycle and repeating the same way indefinitely.

Like the longer-period line puffers in Conway's Life, it may be far beyond our current computational reach to find the actual period of one of these crystal puffers. Golly could certainly try, though.

Would just need a script to select and chop out a 500x500 region (let's say) at the boundary, maybe every few thousand ticks. Discard the rest of the pattern -- g.clear(1) -- to keep memory use from getting too large.

Then after running a suitably long time, start checking hashes of the 500x500 rectangle every three ticks, and see if it settles into a repeating pattern with a period short enough to find.

[KittyTac]

Thanks, I found it while looking for GoL-like rules to feed to apgsearch. Found something that is W110-like instead.