A 'life-emerging' INT rule?

[rabbit]

The latest version I have currently has the ability to self-replicate, but the process interferes with the first replicator, leading to a collapse at about generation 30,000 for this starting configuration. The behavior of large random soups (1024x1024, 2048x2048) is still undefined.

This pattern appears to grow chaotically forever:

Code: Select all

x = 580, y = 7, rule = EvoPhotons-pre
.B499.B$3B497.3B$.B499.B4$79.A499.A!

Around 5000-6000 generations in, a regular stairstep formation starts growing like so:

Code: Select all

x = 105, y = 11, rule = EvoPhotons-pre
28.BA4$BA4$103.B$103.2B$103.B!

[speedydelete]

Thanks for the suggestion! For these rules, what do you think could be the common stable splitter? Dot cannot work since the spaceship fails with S0, and the minrule B2ce/S01 is already quite expanding with natural puffers present.
In your reference topic, I noticed the rule having near-splitters and small stable reflectors. However they do not seem to be too common, especially since a ship-emitting circuitry requires quite some objects to work, and would therefore be quite rare. My consideration has been that I get a stable 0&180 splitter and loop Ts between two of these splitters with bypass collision. This goal has lead to rule byproducts which shift the dot by 1 or 2, and I actually found it useful if we are expecting the rule to be not too complex with all the expansions and apgsearchable(?), since shifting the dot makes guns impossible as the 2 splitters will collide and eliminate everything eventually. However, a shift requires the bypass to also change some phase(?) so that multiple Ts can still loop.
A benefit of using diagonal ships is that they are less likely to have symmetry explosion, which could be another problem in natural soups. Another is that for diagonally asymmetric ships such as the banana spark you demonstrated in the large rulespace, if we have a diagonally symmetric stable splitter the splitter's ability is doubled. I will look further into your suggested rulespace and try later back home.

Instead of just using Ts, I think we should branch out to other 3-cell ships, here are all the possible patterns for period 3 (they can only have C1 or D2 symmetry):

Code: Select all

x = 69, y = 57, rule = B/S012345678
9o3b9o3b9o3b9o3b9o$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$ob2o4bo3bob2o4bo3bob2o
4bo3bobobo3bo3bob2o4bo$obo5bo3bo3bo3bo3bo7bo3bo2bo4bo3bo7bo$o7bo3bo7b
o3bobo5bo3bo7bo3bo3bo3bo$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$o7bo3bo7bo3bo7b
o3bo7bo3bo7bo$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$9o3b9o3b9o3b9o3b9o4$9o3b9o
3b9o3b9o3b9o$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$ob2obo2bo3bob2o4bo3bobobo3b
o3bobobo3bo3bobobo3bo$o7bo3bo7bo3bo4bo2bo3bo5bobo3bo7bo$o7bo3bo4bo2bo
3bo7bo3bo7bo3bobo5bo$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$o7bo3bo7bo3bo7bo3bo
7bo3bo7bo$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$9o3b9o3b9o3b9o3b9o4$9o3b9o3b9o
3b9o3b9o$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$ob2o4bo3bobobo3bo3bobobo3bo3bob
obo3bo3bobo2bo2bo$o4bo2bo3bo7bo3bo7bo3bo7bo3bo2bo4bo$o7bo3bo2bo4bo3bo
4bo2bo3bo5bobo3bo7bo$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$o7bo3bo7bo3bo7bo3bo
7bo3bo7bo$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$9o3b9o3b9o3b9o3b9o4$9o3b9o3b9o
3b9o3b9o$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$obo2bo2bo3bobo3bobo3bobo3bobo3b
obo5bo3bobo5bo$o7bo3bo3bo3bo3bo7bo3bo2bo4bo3bo2bo4bo$o2bo4bo3bo7bo3bo
3bo3bo3bo4bo2bo3bo7bo$o7bo3bo7bo3bo7bo3bo7bo3bo4bo2bo$o7bo3bo7bo3bo7b
o3bo7bo3bo7bo$o7bo3bo7bo3bo7bo3bo7bo3bo7bo$9o3b9o3b9o3b9o3b9o4$9o3b9o
3b9o3b9o3b9o3b9o$o7bo3bo7bo3bo7bo3bo7bo3bo7bo3bo7bo$obo5bo3bobo5bo3bo
bo5bo3bobo5bo3bobo5bo3bobo5bo$o3bo3bo3bo3bo3bo3bo3bo3bo3bo3bo3bo3bo4b
o2bo3bo5bobo$o2bo4bo3bo7bo3bo7bo3bo7bo3bo3bo3bo3bo3bo3bo$o7bo3bo2bo4b
o3bo4bo2bo3bo5bobo3bo7bo3bo7bo$o7bo3bo7bo3bo7bo3bo7bo3bo7bo3bo7bo$o7b
o3bo7bo3bo7bo3bo7bo3bo7bo3bo7bo$9o3b9o3b9o3b9o3b9o3b9o!

It's probably better to use one of the smaller patterns in the top row. We should EPE each one of those up to p3 and systematically figure out which ones are the best.

As for the B2ce/S1 ship, dominoes or blocks should be used for the still life, I think. Also, yes, the B2-ak5j/S12-k rulespace is not very conducive to natural life... it is probably not good for this purpose.

Also, I was incorrect earlier, the B2ce/S1 ship should be tied with several others as holding the versatility record without B2a. Of course, B2a could also be considered too... notice the B2ac3q/S0 and B2-ek/S0 rulespaces.

Edit 2: may I also ask how do I get more non-expanding rules? The rules are too expansive and causes the whole thing to break down.

That's just luck... Sometimes, you find a transition that makes the rule less explosive, so you could try that.

I'll explain in my next post, after I clean it up, the EPE splitter-searching techniques I've developed.

[NNlk05]

I think a V ship has the biggest rule ship out of all (non light speed) ships in inf (IRCC 2^90).

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

It's probably better to use one of the smaller patterns in the top row. We should EPE each one of those up to p3 and systematically figure out which ones are the best.

Done:

2o2$o!.txt

(314 Bytes) Downloaded 5 times

bo$obo!.txt

(2.01 KiB) Downloaded 2 times

How are we then supposed to "systematically figure out which ones are the best"?

EDIT: I don't know how to do EPE wizardry, but looking at the bare rules B2c3ar4a/S01e2e3ijn seems to be the best base, given that it has a 180-degree reflector:

Code: Select all

x = 6, y = 4, rule = B2c3ar4a/S01e2e3ijn
o$4bo2$4b2o!

This allows for oscillators of period 6n:

Code: Select all

x = 9, y = 32, rule = B2c3ar4a/S01e2e3ijn
2b2o2$2bo$o3bo4$2b2o2$2bo$o4bo4$2b2o2$2bo$o5bo4$2b2o2$2bo$o6bo4$2b2o2$
2bo$o7bo!

Census

[speedydelete]

It's probably better to use one of the smaller patterns in the top row. We should EPE each one of those up to p3 and systematically figure out which ones are the best.

How are we then supposed to "systematically figure out which ones are the best"?

EDIT: I don't know how to do EPE wizardry, but looking at the bare rules B2c3ar4a/S01e2e3ijn seems to be the best base, given that it has a 180-degree reflector:

Code: Select all

x = 6, y = 4, rule = B2c3ar4a/S01e2e3ijn
o$4bo2$4b2o!

Here's some ad-hoc criteria:

• Lower period and higher speed is better, the closer you can pack spaceships together, the better!
• Higher versatility (number of rules it works in) is better
• Lower explosive-ness of rulespace is better (note that this does not apply to the B2ce/S1 ship because even though the minrule B2ce/S1 is explosive, this is because of the replicator, which is fairly easy to suppress)
• Small still lifes: D8 symmetry is preferred, though D4 also works

The 180 degree reflector is not that useful in my opinion.

For the final rule, I would like it to also have B3/S23-level depth in conduits and catalysts if possible, but this is admittedly just a personal preference.

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

I managed to find three rules with 3 small spaceships, two orthogonal and one diagonal each. Could this be useful in making Demonoids?

Code: Select all

x = 13, y = 4, rule = B2cei3eiy4a/S12c3ae4i
bo4bo3b2o$obo3bo5bo2$6bo!

Code: Select all

x = 13, y = 4, rule = B2ce3eij/S12ck3an4i
bo4bo3b2o$obo3bo5bo2$6bo!

Code: Select all

x = 13, y = 4, rule = B2ce3eij6c/S12ck3ae5y
bo4bo3b2o$obo3bo5bo2$6bo!

(I do not know how to find a splitter.)

(EDIT: If you ever find rules based on these, please credit me for the first step of the rule construction.)