Page 1 of 2

What do you want out of (conway's) life this year?

Posted: April 9th, 2013, 12:05 pm
by LumpsOfMuck
I guess this can be considered a goal list for the year. The way I see it, goals help focus progress better than just progressing randomly and/or in an unorganized matter.

Now to state my wishes:

1. I want to see a new period osciollator found, because I would like to know whether life is truly omniperiodic.

2. I want to see some investigation done into the rules B38/S23 ,B3/S238, and ,B38/S238 . Because these rules are similar to life in their behavior (they are chaotic but not explosive), yet have their own unique nuances, such as the fact that at least one of them causes the pi heptomino to yield a pulsar, and the fact that the switch engine (unfortunately) is completely unstable in them. But he, where the switch engine fails, a new type of pattern may succeed. But we will never know unless we look.

Re: What do you want out of (conway's) life this year?

Posted: April 9th, 2013, 5:46 pm
by DivusIulius
LumpsOfMuck wrote:rules B38/S23 ,B3/S238, and ,B38/S238
Hmmm... do those rules have an official name?

Re: What do you want out of (conway's) life this year?

Posted: April 9th, 2013, 8:46 pm
by Tropylium
In case they do not, how about: "Pedestrian Life" for B38/S23 (since traffic lights do not occur in this rule), "Pulsar Life" for B3/S238 (given the frequent pulsars from pi heptominoes), and "Honey Life" for B38/S238 (since big beehives occur surprizingly frequently in this rule)?

Granted though, B38/S238 is also "pedestrian", and various other rules in the B3…/S23…8 range also produce big beehives more frequently than regular Life, but just throwing out some ideas.

Re: What do you want out of (conway's) life this year?

Posted: April 10th, 2013, 12:34 am
by Sphenocorona
Another possible name for B38/S23 (it's a bit silly but still descriptive): "SuperHighLife" (HighLife is B36/S23, and traffic lights are really rare in HighLife too - otherwise the Bomber would probably explode)

Re: What do you want out of (conway's) life this year?

Posted: April 10th, 2013, 12:39 am
by glider_rider
I would also like a new period of oscillator.
Also, I want 4c/11 technology.

Code: Select all

x = 30, y = 4, rule = B3/S23
b3o4b3o9b3o3b3o$o3bo2bo3bo7bo3bobo3bo$2ob2o2b2ob2o7b2ob2ob2ob2o$o10bo
7bo9bo!
It's a start.

Re: What do you want out of (conway's) life this year?

Posted: April 10th, 2013, 2:08 pm
by DivusIulius
I like all those names, they're not wild at all. :wink: Thanks!

Re: What do you want out of (conway's) life this year?

Posted: April 10th, 2013, 11:27 pm
by Sphenocorona
I imagine the discussion for B3/S238, B38/S23, and B38/S238 should probably get moved to another thread... (I discovered a small explosion in B38/S238 similar(?) to LoM in Life, and stairstep hexamino in B38/S238 is a methuselah that stabilizes in 3591 generations and outputs 38 gliders)

Now, back to the purpose of this thread:

For spaceships I would like to see some sort of addition to the table of spaceship speeds (like an oblique spaceship or a c/7 wickstrecher)

For oscillators I would like to see a new oscillator period. Especially period 34.

For software, I'd like to see a graphical search program that is at least a little bit better than WLS (maybe a version of gfind with some sort of basic GUI? probably won't happen though)

I would ask for Golly to implement SmoothLife but Golly would have to add support for a pseudo-continuous 'grid', so not sure how well that would work.

Re: What do you want out of (conway's) life this year?

Posted: May 7th, 2013, 8:46 am
by David
I want EVERYTHING. :wink:

Just kidding! I want B3/S23 to be proven to be omniperiodic, as well as new spaceship velocities. :)

Re: What do you want out of (conway's) life this year?

Posted: May 7th, 2013, 11:09 am
by MikeP
LumpsOfMuck wrote:1. I want to see a new period osciollator found, because I would like to know whether life is truly omniperiodic.
Well, we have period 43 and 53 oscillators now.

Re: What do you want out of (conway's) life this year?

Posted: May 7th, 2013, 8:39 pm
by pcallahan
This is more of a wish of the moment than of the year, but I'd like to see other significant catalysts come out of bellman. The snark glider reflector is almost miraculous, so it seems like it should be easy to get a new Herschel conduit or oscillator stabilization out of this kind of search. I tried two things that did not work, but it doesn't mean they can't:

(1)
It would be nice to find a small p1 or p2 stabilizer for the "ak-47" to get a true p94. There is a small p4 stabilization, but that makes it a p188 pseudo-gun, and there are some larger stabilizations that involve feeding glider output back, but nothing that just caps the ends like the p4.

All you need to stabilize ak-47 is to eat a traffic-light from far enough away to avoid the main engine. It seems like there should be a way to collide either a block or blinker with a traffic light and then use a catalyst to convert the explosion back into just the original block or blinker (at the right phase). I didn't find this with bellman, but I'm just getting the hang of it.

(2)
I thought it would be worth seeing what bellman could do with a pi heptomino. E.g., if it could be turned 90-degrees with appropriate spacing there might be a new oscillator. Converting it to a Herschel would also be nice. I did one search and found something that appeared very exciting. It shifted the pi to the left and converted it into a clean honey-farm reaction. Unfortunately, the still life is impossible as far as I can tell (you either get 3 neighbors around an empty cell or more than three around a live cell). I'm attaching the gif output from bellman.

Re: What do you want out of (conway's) life this year?

Posted: May 8th, 2013, 12:59 am
by dvgrn
pcallahan wrote:Unfortunately, the still life is impossible as far as I can tell (you either get 3 neighbors around an empty cell or more than three around a live cell). I'm attaching the gif output from bellman.
Yup, there seems to be a problem with that one "don't care" cell in the middle of the bottom row. Here's some RLE, in case anyone has trouble translating the GIF -- unstable but correct on the left, stable but wrong on the right:

Code: Select all

x = 47, y = 14, rule = B3/S23
2b2o28b2o$3b2o28b2o$2b2o28b2o3$9bo2b2ob2o22bo2b2ob2o$5b2obobo2bob2o18b
2obobo2bob2o$2obo2bobob2obo16b2obo2bobob2obo$2ob2o4bo3bo16b2ob2o4bo3bo
$4bob3o2b2ob2o19b4o2b2ob2o$4o8bobo15b4o8bobo$o3b3o4bo3bo14bo2bob4o2bo
3bo$3bo2bo4b2ob2o18b2o2bo2b2ob2o$4b2o!
Definitely a promising-looking almost-result, though. I'll be really interested to see what else Bellman can find, with the help of multiple naturally-evolved large neural nets...!

Re: What do you want out of (conway's) life this year?

Posted: May 9th, 2013, 6:54 pm
by MikeP
pcallahan wrote:All you need to stabilize ak-47 is to eat a traffic-light from far enough away to avoid the main engine. It seems like there should be a way to collide either a block or blinker with a traffic light and then use a catalyst to convert the explosion back into just the original block or blinker (at the right phase).
I took a different approach, which was to search for a catalyst which eats the traffic light directly. I've found a lot of exotic eaters with Bellman. Most of them are boring, but in situations like this they're just the thing. It wouldn't surprise me if there are other uses for this traffic light eater too.

Code: Select all

#C "AK-94" period 94 glider gun
#C Mike Playle, 9th May 2013
#C RIP Mikhail Kalashnikov, died 23 Dec 2013, aged 94.
x = 38, y = 25, rule = B3/S23
13b2o7bo$5bo7b2o5b3o6b2o$5b3o11bo9bo$8bo10b2o6bobo$7b2o18b2o4$2b2ob2o$
o2bob2o28b2o$2obo26b2o2bo2bo$3bo26bobo2b2o$3b2o28b2o$b2o2bobo12bo13bo$
o2bo2b2o11bobo12bob2o$b2o15bo3bo8b2obo2bo$18bo3bo8b2ob2o$18bo3bo$19bob
o$20bo$29b2o$17b2o10bo$18bo11b3o$15b3o5b2o7bo$15bo7b2o!
Here's the input file which found this catalyst:

Code: Select all

#S first-encounter 24
#S last-encounter 42
#S repair-interval 10
#S stable-interval 10
#S max-live 200
#S max-active 8
#P 9 11

            ????????????
            ????????????
            ????????????
     ...    ????????????
    ..@..   ????????????
   ..@.@..  ????????????
   .@...@...????????????
   .@. .@.  ????????????
   .@...@.  ????????????
   ..@.@..  ????????????
    ..@..   ????????????
  ......    ????????????
  .@@.      ????????????
....@.  ....????????????
.@@@..  .@@.
.@...   .@@.
...     ....
[edit: added dedication]

Re: What do you want out of (conway's) life this year?

Posted: May 9th, 2013, 7:12 pm
by pcallahan
Wow! Thanks. At this point, I'd have to be greedy it to ask for anything else out of Life this year. Time to start thinking about 2014.

BTW, I did try a bellman search for a direct eater, but obviously I don't have the hang of it yet. This is really a beautiful result even if it isn't quite as canonical as a stable reflector.

Re: What do you want out of (conway's) life this year?

Posted: May 9th, 2013, 8:03 pm
by MikeP
I've been tinkering with this algorithm on and off for the last 15 years and I think I'm just starting to get a feel for what it can do! There is a huge deep sea of unexplored search space just waiting to be plumbed.

My main wish for this year is for someone else to find a significant result using this technique - either by running Bellman, or by devising a program of their own that works on similar principles.

Re: What do you want out of (conway's) life this year?

Posted: May 9th, 2013, 8:29 pm
by pcallahan
It's reasonably well documented and usable (at least compared to my gencols and ptbsearch tools). The animated gifs are really nice (I used to copy-paste matrices of results into XLife, but I'm not set up for that anymore). I think the part I haven't figured out is how many unknown cells to use. With too many, it seems to go off without finding anything, and after an hour or so, I don't know if it's worth waiting longer.

I also don't have the time to delve any deeper. If I had the kind of time I used to have, I would probably try to catalog small catalysts, e.g. by starting with some small active patterns and see which still-lifes can interact and self-repair, independent of what else they do. One hard part would be eliminating ones that are essentially duplicates. Given a set of small catalysts, it might be more efficient to try to use them in combination than to search for a new one every time. I'm not even certain that's true, since I've been surprised to see some of these reactions. It almost seems like any still life could be a catalyst combined with just the right explosion.

Re: What do you want out of (conway's) life this year?

Posted: May 9th, 2013, 10:46 pm
by dvgrn
pcallahan wrote:...Time to start thinking about 2014.
Well, okay: in 2014 I'd like to see a finished B3/S23 replicator pattern. For this year I'll be satisfied with a completed one-armed Geminoid self-constructing spaceship.
pcallahan wrote:This is really a beautiful result even if it isn't quite as canonical as a stable reflector.
MikeP wrote:It wouldn't surprise me if there are other uses for this traffic light eater too.
Yes, it's impressive how simple and compact that eater is -- it looks like it should be somewhere in Dean's eater collection, but of course it can't eat a glider, only a pre-traffic light...

It would certainly surprise me if there weren't other uses. Let's see, maybe if you had a couple of spare glider sources from a high-period Herschel-based gun, but both of them were the wrong phase and/or color to be reflected by Snarks to get the glider phase you want, then you could use this as a rephaser:

Code: Select all

#C Suppressed traffic light + glider reaction;
#C Snark reflector shown for comparison
x = 56, y = 30, rule = B3/S23
bo52bo$o52bo$3o50b3o3$10b2o$5b2o2bo2bo$5bobo2b2o24bo$8b2o26b3o$9bo29bo
$9bob2o25b2o$6b2obo2bo$6b2ob2o4$4b2o$4bo43b2o$5b3o33b2o5bobo$7bo33b2o
7bo$50b2o2$37bo$36bobob2o$36bobobobo$33b2obobobobo2bo$33bo2bo2b2ob4o$b
3o31b2o4bo$bo39bobo$2bo39b2o!
The large still life eats three out of four blinkers from a traffic light, even without the fishhook eater -- that's bound to come in handy somewhere. I think up to now we've only had catalysts that could eat two blinkers at a time (a ship) or one at a time (fishhook eater), but there wasn't a way to reach back far enough into the active reaction to pick off that innermost blinker.

So it seems likely that this will allow some previously impossible Herschel conduits to be completed, among other things. I'm tempted to add it to catgl or ptbsearch and see what uses come up in a search...!

Re: What do you want out of (conway's) life this year?

Posted: May 10th, 2013, 12:30 am
by Sokwe
Wow! That worked much better than I expected.
MikeP wrote:It wouldn't surprise me if there are other uses for this traffic light eater too.
Indeed, another great thing about this new catalyst is that it can replace a spark-based pre-pulsar pusher found by Jason Summers:

Code: Select all

x = 22, y = 54, rule = B3/S23
10b2o$9bo2bo2b2o$10b2o2bobo$12b2o$12bo$9b2obo6bo$9bo2bob2ob2ob2o$11b2o
b2o3bo4$11b2ob2o3bo$9bo2bob2ob2ob2o$9b2obo6bo$12bo$12b2o$10b2o2bobo$9b
o2bo2b2o$10b2o7$4b2o6b2o$4bobo6bo$10b3o$6b2o2b2o$7b3ob2o2bo$12bo3bo$3b
2o6bo4bo$2b4o6b4o$bo4bo$2o4b2o$2o4b2o11bo$bo4bo7b2ob2ob2o$2b4o8b2o3bo$
3b2o2$3b2o$2b4o8b2o3bo$bo4bo7b2ob2ob2o$2o4b2o11bo$2o4b2o$bo4bo$2b4o6b
4o$3b2o6bo4bo$12bo3bo$7b3ob2o2bo$6b2o2b2o$10b3o$4bobo6bo$4b2o6b2o!
Here is a new p22 pre-pulsar shuttle and an improved p25 pre-pulsar shuttles:

Code: Select all

x = 25, y = 53, rule = B3/S23
2b2o18b2o$bo2bo2b2o8b2o2bo2bo$2b2o2bobo8bobo2b2o$4b2o14b2o$4bo16bo$b2o
bo9bo6bob2o$bo2bob2o5b2o3b2obo2bo$3b2obo7bo3b2ob2o4$3b2obo7bo3b2ob2o$b
o2bob2o5b2o3b2obo2bo$b2obo9bo6bob2o$4bo16bo$4b2o14b2o$2b2o2bobo8bobo2b
2o$bo2bo2b2o8b2o2bo2bo$2b2o18b2o8$5b2o$6bo$4bo$4b5o$8bo9b2o$2b4o7b2o2b
o2bo$2bo2bo7bobo2b2o$16b2o$17bo$10bo6bob2o$b2o6b2o3b2obo2bo$2o8bo3b2ob
2o4$2o8bo3b2ob2o$b2o6b2o3b2obo2bo$10bo6bob2o$17bo$16b2o$2bo2bo7bobo2b
2o$2b4o7b2o2bo2bo$8bo9b2o$4b5o$4bo$6bo$5b2o!
MikeP wrote:I've found a lot of exotic eaters with Bellman. Most of them are boring
If these "boring" catalysts have novel reactions, please post them.

Re: What do you want out of (conway's) life this year?

Posted: May 10th, 2013, 1:32 am
by pcallahan
Here's a catalyst I found with Bellman that I wouldn't exactly call boring, but it doesn't seem very useful. It came up in my unsuccessful attempts to make a traffic light eater. It starts with a pre-traffic-light and block and turns it into two loaves. Since doesn't even restore the block, I doubt there is any application.

Code: Select all

x = 10, y = 17, rule = B3/S23
3o$bo4$b2o$b2o3$3bo2bo$3b4o2bo$o6b3o$5obo$4b2o2b2o$2o7bo$2o6bo$8b2o!

Re: What do you want out of (conway's) life this year?

Posted: May 10th, 2013, 2:03 am
by Sokwe
pcallahan wrote:Here's a catalyst I found with Bellman that I wouldn't exactly call boring, but it doesn't seem very useful.
This could be useful:

Code: Select all

x = 20, y = 16, rule = B3/S23
5bo$5b3o$8bo4b2ob2obo$7b2o5bobob2o$14bo$12b2obo$13bo2bo$13bob2o$12b2ob
o$7b3o5bo$7bobo5bob2o$7b3o4b2ob2o$2b2o$bobo$bo$2o!
It's extremely disappointing that this doesn't work with the p24 blinker-to-traffic light.

Edit: I just realized that, in terms of minimum population, the p25 pre-pulsar shuttle beats the previous record holder by 2 cells.

Re: What do you want out of (conway's) life this year?

Posted: May 10th, 2013, 5:31 am
by MikeP
Sokwe wrote:
MikeP wrote:I've found a lot of exotic eaters with Bellman. Most of them are boring
If these "boring" catalysts have novel reactions, please post them.
Since I started experimenting with this I must have seen many hundreds of partial patterns of the form "evolving pattern encounters catalyst and subsequently dies out". I'm afraid I don't have complete records of all the patterns I've discarded for being relatively uninteresting.

Re: What do you want out of (conway's) life this year?

Posted: May 10th, 2013, 1:25 pm
by Kasuha
This almost looks like we're on the verge of discovering missing oscillator periods by "capping" already known reactions with stable eaters.
For instance, anyone would try to make a stable endpoint for this p19 wick?

Code: Select all

x = 24, y = 65, rule = B3/S23
15bobo$15bobo$15b3o$22bo$11bo9bobo$10bobo9bo$11bo$16b3o$16bobo$16bobo
2$13bobo$13bobo$13b3o$20bo$9bo9bobo$8bobo9bo$9bo$14b3o$14bobo$14bobo2$
11bobo$11bobo$11b3o$18bo$7bo9bobo$6bobo9bo$7bo$12b3o$12bobo$12bobo2$9b
obo$9bobo$9b3o$16bo$5bo9bobo$4bobo9bo$5bo$10b3o$10bobo$10bobo2$7bobo$
7bobo$7b3o$14bo$3bo9bobo$2bobo9bo$3bo$8b3o$8bobo$8bobo2$5bobo$5bobo$5b
3o$12bo$bo9bobo$obo9bo$bo$6b3o$6bobo$6bobo!
Or perhaps for this p41 wick?

Code: Select all

x = 21, y = 141, rule = B3/S23
8bo3bo$8bo3bo$7bobobobo$8bo3bo$8bo3bo2$3o15b3o2$8bo3bo$8bo3bo$7bobobob
o$8bo3bo$8bo3bo20$8bo3bo$8bo3bo$7bobobobo$8bo3bo$8bo3bo2$3o15b3o2$8bo
3bo$8bo3bo$7bobobobo$8bo3bo$8bo3bo20$8bo3bo$8bo3bo$7bobobobo$8bo3bo$8b
o3bo2$3o15b3o2$8bo3bo$8bo3bo$7bobobobo$8bo3bo$8bo3bo20$8bo3bo$8bo3bo$
7bobobobo$8bo3bo$8bo3bo2$3o15b3o2$8bo3bo$8bo3bo$7bobobobo$8bo3bo$8bo3b
o20$8bo3bo$8bo3bo$7bobobobo$8bo3bo$8bo3bo2$3o15b3o2$8bo3bo$8bo3bo$7bob
obobo$8bo3bo$8bo3bo!
Note: Both patterns are taken from Jason Summers' collection. There are wicks for all periods which are not yet covered by finite oscillators but these two seemed to be the best potential candidates to me...

Re: What do you want out of (conway's) life this year?

Posted: May 13th, 2013, 7:55 pm
by pcallahan
Here's another interesting but useless catalyst from Bellman. I was hoping to find a smaller century eater. Instead I found a conversion to a pond, loaf, block constellation that doesn't leave much room for other catalysts. But it does something I have never seen before. It removes one end of a double fishhook eater and then restores it later.

Code: Select all

x = 6, y = 14, rule = B3/S23
3b2o$b3o$2bo6$4b2o$3bobo$3bo$b3o$o$2o!

Re: What do you want out of (conway's) life this year?

Posted: May 31st, 2013, 1:18 am
by knightlife
A p19 oscillator would be nice. I was impressed by this p17 oscillator years ago (by Dean Hickerson):

Code: Select all

x = 15, y = 13, rule = B3/S23
5bo$4bobo$4bobo3b2o$b2obob2o3bo$2bobo6bob2o$o2bobob3obo2bo$2obobo4bobo
$3bobo2b2o2b2o$3bobo3bobo$4b2obobobo$6bobobo$6bobo$7b2o!
Maybe there is a similar oscillator with period 19, although stabilization for the p19 wick posted above might be easier to find.

Re: What do you want out of (conway's) life this year?

Posted: June 8th, 2013, 9:47 am
by codeholic
pcallahan wrote:I was hoping to find a smaller century eater.
Here is a small century eater for the bistable switch.

Code: Select all

x = 75, y = 61, rule = B3/S23
2bo$obo$b2o17$57bob2o$57b2obo2$55b5o$54bo2bo2bo$44bo9b2o3bobo$42bobo
15b2o$43b2o15$73b2o$73b2o4$61bo$60bobo$61b2o9$49b2o$48bobo$48bo$47b2o!

Re: What do you want out of (conway's) life this year?

Posted: June 10th, 2013, 6:26 am
by Freywa
Kasuha wrote:This almost looks like we're on the verge of discovering missing oscillator periods by "capping" already known reactions with stable eaters.
For instance, anyone would try to make a stable endpoint for this p19 wick?

Code: Select all

x = 24, y = 65, rule = B3/S23
...
I'm working on finding the stabilisation for that wick using Bellman. Unfortunately the tile-wrapping issue may need to be resolved first.