Switch engine Methuselahs

[TWLife]

I've recently been working on Switch engine Methuselahs, the're quite interesting, so here's a 35 cell Methuselah that lasts for over 1.5 million gens. however, it doesn't start within a 20x20 box, so it's shouldn't really be considered a Methuselah.

Code: Select all

x = 170, y = 180, rule = B3/S23
167b3o74$50b4o$54bo$54bo$55b2o5$42bo$42bo$42bo$42bo$43b2o$45bo$45bo70$
2b4o$6bo$6bo$7b2o12$o$o$o$o$b2o$3bo$3bo!

And the thing in the bottom left corner isn't the lowest amount of cells required to produce the SE, I just don't know where to find them. So really, it should be about 28 cells or something...

[dvgrn]

Are you familiar with Nick Gotts' weblog? His ark patterns seem somewhat along these lines, and it looks like many of his patterns have similarly long "eventful histories".

...the thing in the bottom left corner isn't the lowest amount of cells required to produce the SE, I just don't know where to find them.

Your SE predecessors are 8 cells, and all of Nick Gotts' switch engines seem to get started with 8-cell predecessors, so I strongly suspect you won't find anything with 7 cells or less.

But I think you mean that you don't really need the top switch engine in the pair at the bottom, since you just end up with a single block-laying SE once the debris settles. Probably just a block or blinker somewhere will allow the block-laying switch engine to escape, but I don't seem to have an example handy.

[towerator]

Actualy "arkoids" aren't considered methusalehs for some...
it's because they are basically soups continuously shot by dirty backrakes.

[TWLife]

Are you familiar with Nick Gotts' weblog? His ark patterns seem somewhat along these lines, and it looks like many of his patterns have similarly long "eventful histories".

...the thing in the bottom left corner isn't the lowest amount of cells required to produce the SE, I just don't know where to find them.

Your SE predecessors are 8 cells, and all of Nick Gotts' switch engines seem to get started with 8-cell predecessors, so I strongly suspect you won't find anything with 7 cells or less.

But I think you mean that you don't really need the top switch engine in the pair at the bottom, since you just end up with a single block-laying SE once the debris settles. Probably just a block or blinker somewhere will allow the block-laying switch engine to escape, but I don't seem to have an example handy.

actually, I think there is a 9 or 10 cell pattern that also results in a single switch engine

[dvgrn]

actually, I think there is a 9 or 10 cell pattern that also results in a single switch engine

Ah, yes -- thanks for jogging my memory (I keep most of my memory online these days, it seems.) So a somewhat equivalent pattern would be

Code: Select all

#C 29-cell methuselah becomes boring at 3,256,766 ticks
x = 163, y = 163, rule = B3/S23
160b3o74$43b4o$47bo$47bo$48b2o5$35bo$35bo$35bo$35bo$36b2o$38bo$38bo69$
6bo$4bob2o$4bobo$4bo$2bo$obo!

Luckily the 10-cell switch-engine predecessor leaves some exhaust that suppresses the first high-period interaction between the two c/12 puffers, so the pattern stays interesting for twice as long. Are there similar high-period interactions between just two switch engines, or do you need all three?

[TWLife]

well, I don't actually know... but I think 3 gives more possibilities....

[NickGotts]

I must get round to updating that ancient "blog" (I didn't actually get the idea of a blog when I created it)! Yes, I 've done quite a bit, most of it I'm afraid never written up anywhere, on patterns that keep on producing novelty for a long time. The smallest (in terms of number of initial on-cells) I've found that go on as far as I've been able to trace them (although I suspect they don't last forever), have 32 cells, like this one:
#lj2ah-ohhsf-w344s1250
x = 386, y = 1273, rule = B3/S23
347bo$347bo$347bo$347bo$345b2o$344bo$344bo12$384bo$383bobo2$382bo2bo$
382b2o$382bo1227$31bo$31b2o$31bo2bo2$32bobo$33bo13$o3bo$bobo$2bo2bo$5b
o$5bo!