ConwayLife.com

I am beginning to think completing this G10-to-T is literally impossible:

x = 144, y = 121, rule = B3-qr4k5cek6c/S23-a4er56aek
bo$2bo$3o51$131bo4bo$130bobo2bobo$127bo3bo4bo$126bobo$123bo3bo10bo$122b
obo12bobo$119bo3bo14bo$118bobo$115bo3bo20bo$114bobo22bobo$111bo3bo24b
o$110bobo$107bo3bo30bo ...

Circuitry collection, a lot of it is new:


T -> signal:
x = 11, y = 9, rule = B3-qr4k5cek6c/S23-a4er56aek
9b2o$8bobo$8b2o$bo$2o$bo$9bo$8bobo$9bo!

G -> signal:
x = 10, y = 15, rule = B3-qr4k5cek6c/S23-a4er56aek
bo$2bo4b2o$3o4bo$9bo$8b2o$6bo$6b2o$9bo$2bo5b2o$bobo$2bo2$5bo$4bobo$5b
o!

2 signals ...

I think this is all of the 2T collisions:


nothing:
x = 12, y = 12, rule = B3-qr4k5cek6c/S23-a4er56aek
10bo$9b3o8$bo$2o$bo!
x = 12, y = 11, rule = B3-qr4k5cek6c/S23-a4er56aek
9bobo$10bo$10bo6$bo$2o$bo!
x = 12, y = 9, rule = B3-qr4k5cek6c/S23-a4er56aek
9bobo$10bo$10bo4$bo$2o$bo!
x = 12, y = 9, rule ...

Diode:

x = 55, y = 19, rule = B3-qr4k5cek6c/S23-a4er56aek
2o$3o$bo3bo$4bobo$5bo3bo$8bobo42b2o$9bo3bo38b3o$12bobo34bo3bo$13bo3bo
30bobo$16bobo26bo3bo$17bo3bo22bobo$20bobo18bo3bo$21bo3bo14bobo$24bobo
10bo3bo$25bo3bo6bobo$28bobo6bo$29bo3bo$32bobo$33bo!

Generally it is fairly simple to connect up ...

I think this is the simplest possible orthogonal wire:

x = 43, y = 34, rule = B3-qr4k5cek6c/S23-a4er56aek
5bo7bo7bo7bo7bo$4bobo5bobo5bobo5bobo5bobo$5bo7bo7bo7bo7bo$2o7bo7bo7bo
7bo7bo$3o5bobo5bobo5bobo5bobo5bobo$bo7bo7bo7bo7bo7bo2$38bo$37bobo$38b
o2$41bo$40bobo$41bo2$38bo$37bobo$38bo2$41bo$40bobo ...

I would recommend basing things on speedydelete's apgweb.


I wouldn't recommend doing that, at least for a few months. It's quite outdated and extremely slow, broken, and unoptimized, primarily due to being written in TypeScript and being run as JS, not as wasm. It did work at some point, but I ...

How suspicious...



On Feb. 23, 2008, Dave Greene wrote:
Every sufficiently large and ancient B3/S23 universe observed so far
turns out to be full of this weird agar, which expands in all
directions through random ash at c/137. IceNine seems to absorb
blocks and blinkers particularly well, due to ...

MDA wrote: September 26th, 2026, 2:57 pm Some notes:

- Period maps are only shown up to period 127.

- The period map marks p1 phoenix oscillators as unknown when they should be marked as impossible.

The first one is intentional to reduce server load, I'm going to fix the second.

The server is back up now, everything seems to work, but I could've broken something.

Changes:
- Fixed the bug that caused data to be wiped.
- It's now actually usable on mobile!
- The period maps are now much better and show the c/2o-c/3d and c/2o-c/4d speed limit lines.

Due to a bug, the INT obliques file was wiped. The last backup appears to be at 2026-09-25 18:00:00 UTC, so all oblique ships submitted after that date are lost. The server is temporarily down while I fix this and also do a few other things that I should've done a while ago. After it gets back ...

Similar to any of those two I mentioned(Ikpx2 or Qfind) except with either A* search or dijkstra’s algorithm, although it could be made,maybe it’s either better or worse, idk, but if you combine the gfind algorithm(doesn’t build the full graph) and A* search, it could probably either be better or ...

Can someone make an A* search or Dijkstra's Algorithm Based pattern searcher?(somewhat like Qfind or Ikpx2)
I was thinking about this while going back from school so yeah


How would such a program work? gfind's algorithm (which is used in qfind) already implements the "shortest path" property ...

Hdjensofjfnen wrote: September 21st, 2026, 12:09 am Not every still life allows a solution: for example, the ørb (15 cells) is unsolvable (given my search setup). Can anyone find smaller?

Are you sure you have set it up right? I believe that theoretically there could be one hundreds of cells big but none smaller, even for small objects.

I think moving this to https://conwaylife.com/forums/viewtopic.php?t=5471 is a good idea, can that be done?

TheWayOfTheCon wrote: September 17th, 2026, 12:35 am I haven't been able to do ./install on my Windows computer, even with Nodejs downloaded seperately.

Yes, VLS cannot be used on Windows without the use of Windows Subsystem for Linux (WSL). Even if you got the ./install working, the program itself would probably fail.

I'm currently working on making:
- Better symmetry specification and searching
- Tree search (breadth-first)
- CDCL - I think this plus tree search will make it faster than ikpx2!


VLS now works for most commands, but when I try "./vls B/S B1e2345678/S012345678 periodic 'c/2o' 3 3", this happens ...

BREAKING CHANGE:

For spaceship searches the bounding box now includes the location of the spaceship AT generation <period>, so "./vls B3/S23 periodic c/4d 3 4" no longer finds the glider (but "4 4" does).

I've also added input files! These are kind of a mini programming language that you use to ...

Sigh... yet another unconnectable (on the pi's side) H-to-P...
x = 35, y = 24, rule = LifeHistory
27.A$25.3A$8.A15.A$8.3A13.2A$11.A9.5B7.2A$10.2A3.B4.4B9.A$10.8B.7B4.B
A.A$12.17B.B2A$12.19B$11.20B$9.21B$7.23B$7.2BE19B$6.3BEBE16B$7.2B3E15B
$6.5BE14B$5.10B.9B$4.4B10.9B$3.4B12.6B.2A$2.4B12.8B2A ...

splitterrules wrote: September 14th, 2026, 12:04 pm I tried running "./vls B3/S23 periodic 'c/4d' 4 4" and it gave me these errors:
Screenshot 2026-09-14 120222.png

Should be fixed now, update your VLS using "./update". Also, you seem to be using an old version of gcc, perhaps you should update it too.

erictom333 wrote: September 13th, 2026, 10:53 pm I call these rules linear rules, *snip*

Very nice work!

Does this idea generalize in more than one way (other than addition modulo state count) to multistate grids?

I've updated the program and the above post to use a different option format.

To update VLS, just do "./update" in the lifeweb directory.

Also, the goals of VLS are to be as fast and user-friendly as possible, those are the two most important things.

This is a new supercharged version of lifesrc that I wrote in C and TypeScript. I think it is generally faster than all other lifesrc-style programs. (LLS does not count here because it is not really a lifesrc, however I intend to surpass LLS at some point.) It is as general as LLS, the "variable ...

And finally a side note on replicators. Notice that for B1357/S1357 the rule is its own XOR symmetries, this is also true for the von Neumann replicator rules:

Screenshot 2026-09-07 141313.png

Therefore, the rules are also maximally XOR symmetric, because in MAP a rule cannot have more than 512 ...

Now this is the interesting part!


x = 80, y = 25, rule = B1e2c3eq4aikt5eq6c7e/S01c2-c3-eq4-aikt5-eq6-c7c8:T80,25
20bobobobobobo5b12ob2ob2o5bo2bo2bo3bo4bo$20bobobobobobo5b12ob2ob2o5bo
2bo2bo3bo4bo$20bobobobobobo5b12ob2ob2o5bo2bo2bo3bo4bo$20bobobobobobo5b
12ob2ob2o5bo2bo2bo3bo4bo ...

These are the rotations and reflections of patterns, see the wiki for a deeper explanation. For square-grid CA, they are: C1, C2, C4, D2|, D2-, D2\, D2/, D4+, D4x, and D8. Because the transition function is always centered on a cell, there is no need to distinguish between even and odd, only the odd ...