The Online Life-Like CA Soup Search

For general discussion about Conway's Game of Life.

Re: The Online Life-Like CA Soup Search

Mats wrote:Well... A "normal" methuselah doesn't really stabilize eiter. It's bounding box for generation n is expanding forever. That's not really stable. So it's a matter of definition...

Well, methuselahs don't "stabilize" per se, but the generation count is decided when the population becomes cyclical.
Axaj

Posts: 232
Joined: September 26th, 2009, 12:23 am

Re: The Online Life-Like CA Soup Search

I have run manual verification of the three unvalidated Methuselahs in the high 50 table.

My results are,

Ed, 27070 is 27057
Ed, 26780 is 26106
Schneelocke 26650 is 26188

The latter two were high estimates because a glider escaped from the active area shortly before stability.
ians

Posts: 20
Joined: March 30th, 2011, 8:08 pm

Re: The Online Life-Like CA Soup Search

I have now confirmed these. The 27057-generation pattern is of particular interest, because it leads to the following 17-cell 27014-generation pattern:
`#C 17-cell 27014-generation methuselah#C found by "Ed" (March 8, 2011)x = 29, y = 17, rule = B3/S235bo\$b2o2b2o\$3bo\$2bo8\$o\$o\$2o24bo\$27b2o\$24bobo\$24bo!`

The century predecessor can be replaced by either of these:
`x = 16, y = 4, rule = B3/S234b2o5bo2b2o\$2o2bo6bo2bo\$2bo8bobo\$bo!`
-Matthias Merzenich
Sokwe
Moderator

Posts: 1473
Joined: July 9th, 2009, 2:44 pm

Re: The Online Life-Like CA Soup Search

Fascinating.
I see that the original soup search at generation 58 minus the irrelevant gliders, is identical to the 17 cell at generation 15.
The 43 difference is the same as 27057 minus 27014, and the century predecessors make it coincide with the evolution
of the area to the lower right.
I take it that the "century" is a known, frequently occurring construction and that you know about it.
ians

Posts: 20
Joined: March 30th, 2011, 8:08 pm

Re: The Online Life-Like CA Soup Search

ians wrote:I take it that the "century" is a known, frequently occurring construction and that you know about it.

Indeed, although the "predecessors" that I posted actually evolve into generation 2 of century rather than century itself. Century is among several commonly occurring, unstable patterns (such as the B-heptomino and R-heptomino). Patterns like this are interesting, as they can often be manipulated so as to reappear in a different location and with a orientation (this is the basic idea behind Herschel loops). For example A period-246 gun by David Buckingham, and a period-90 oscillator (diuresis) by David Eppstien are based on century:
`x = 109, y = 42, rule = B3/S2334b2o\$33bo2bo\$36bo\$36bo\$30bo3bobo\$29bo3bobo\$29bo4bo\$30b4o\$69bobo2bobo20bobo2bobo\$65b2obo2bo2bo2bob2o12b2obo2bo2bo2bob2o\$16bob3o48bobo2bobo20bobo2bobo\$15bo5b2o\$18b4o67b2o\$16b2o70bobo\$2o85b2o\$bo86b2o\$bobo\$2b2o37b2o\$41bo\$39bobo\$39b2o\$88b2o\$17b2o68b2o\$16bobo69bobo\$15b2o72b2o\$16b2o\$69bobo2bobo20bobo2bobo\$29b2o34b2obo2bo2bo2bob2o12b2obo2bo2bo2bob2o\$29b2o38bobo2bobo20bobo2bobo3\$43bobo\$44b2o\$44bo\$4b2o\$4b2o2b2o\$2b2obo2b2o\$2b3o3\$3b2o\$3b2o!`

Note that In these cases, the actual pattern known as "century" never occurs, but generation 2 of century does. The following is the actual pattern that "century" refers to:
`..ooooo..o..`

Much of this information can be found on the Life wiki; however, it can be somewhat difficult to navigate if you do not already know what you are looking for.
-Matthias Merzenich
Sokwe
Moderator

Posts: 1473
Joined: July 9th, 2009, 2:44 pm

Re: The Online Life-Like CA Soup Search

Thank you for the above.
I put the first generation century in Golly and ran it. I can see that memorizing the sequence would be valuable.
I have never studied Life seriously, rather tinkered with it on and off over the years.
My introduction to Life was the article in Martin Gardner's Mathematical Games in Scientific American in 1970.
In 1982 I wrote Life in TI9900 assembler on an expanded TI99/4A. I used integer arithmetic on a 32768 square
and kept coordinate pairs in a table. For the first time, I saw the complete generation of the R-pentomino.
ians

Posts: 20
Joined: March 30th, 2011, 8:08 pm

Re: The Online Life-Like CA Soup Search

Nathaniel,

Will you give us a Soup Search script with source to save locally?
Since your concern is tampering with the source in order to create false uploads,
you could omit the network source but publish the search.
I feel that we might be missing some gems below the arbitrary 20,000 generations limit.
Anyone finding a worthy soup could post the rle file to a forum.
ians

Posts: 20
Joined: March 30th, 2011, 8:08 pm

Re: The Online Life-Like CA Soup Search

The unverified, Schneelocke 26,960 in the 50 high table - my manual number is 26945.

Why do not the owners themselves send in the actual number and thus help others verify?

Ian
--
ians

Posts: 20
Joined: March 30th, 2011, 8:08 pm

Re: The Online Life-Like CA Soup Search

Golly 2.2 running on Windows XP Professional 32 bits doesn't save the Soup Search files over 20k.
Golly has reported one file in the 20-21k generations range and one in 24-25.
I have searched the whole C drive for "Methus" and find only the Golly Methuselah folder.
They are being saved in ~/.golly on my ubuntu 10.04 64 bit machine.

Ian
--
ians

Posts: 20
Joined: March 30th, 2011, 8:08 pm

Re: The Online Life-Like CA Soup Search

If the script is using getdir("data") to specify the location then the files should be here:

C:\Documents and Settings\your-user-name\Application Data\Golly\

I think this is normally a hidden folder so you need to turn on the appropriate Windows option to see/search those folders.

Andrew
Moderator

Posts: 748
Joined: June 2nd, 2009, 2:08 am
Location: Melbourne, Australia

Re: The Online Life-Like CA Soup Search

Got them thank you.
I didn't know about hidden files and folders in XP.

Ian
--
ians

Posts: 20
Joined: March 30th, 2011, 8:08 pm

Re: The Online Life-Like CA Soup Search

Here's one of my saved Methuselahs that is an underestimate.
Of course I had hoped that it would run for another two thousand generations
It was saved as 20490 and actually stabilizes at 21015.

`x = 20, y = 20, rule = B3/S239bo5b2o\$2o10bobobob2o\$ob2obobo3bo6bo\$bo4bo2bobo\$b2o8bo6bo\$b2obo2bobob3ob4o\$2b3o4b2o2b2obobo\$3bo9bo2bo\$2bo2b2o3b2o2bo3bo\$bobob2o2bo6b4o\$6bobo2bo4bo2bo\$bob2ob3o3b2obobo\$ob4obobo2bo2bo2b2o\$b4obo2bo2b4o2bo\$b2obo2b2o2b3o2b2o\$o7bobo3b3obo\$3bo3bo5bo\$2b2o4b5o4b2o\$b2o2b2o2bo3b3o\$4o7b2o2bobobo!`
ians

Posts: 20
Joined: March 30th, 2011, 8:08 pm

Re: The Online Life-Like CA Soup Search

I make Schneelocke's 26630 at number 47, 26612.
ians

Posts: 20
Joined: March 30th, 2011, 8:08 pm

Re: The Online Life-Like CA Soup Search

I have finally found an 18-cell methuselah that breaks the 40K barrier, a goal I have had since my earlier post late last year:
`#C 18-cell 40514-generation methuselah#C based on Richard Wobus' 15-cell marvelx = 78, y = 54, rule = B3/S2377bo\$77bo\$77bo21\$3o20\$3bo\$3bo\$3bo5\$20b3o\$9b3o10bo\$22bo\$21bo!`

The bounding box is only slightly larger at 78 x 54.
knightlife

Posts: 566
Joined: May 31st, 2009, 12:08 am

Re: The Online Life-Like CA Soup Search

Even though the Soup Search itself is down, could the table of Methuselahs found (and maybe even the census itself, depending on how much work needs doing to it) still be left on the site, as a reference?

Lewis

Posts: 320
Joined: March 17th, 2009, 5:26 pm
Location: UK

Re: The Online Life-Like CA Soup Search

Lewis wrote:Even though the Soup Search itself is down, could the table of Methuselahs found (and maybe even the census itself, depending on how much work needs doing to it) still be left on the site, as a reference?

I'll copy it over to a permanent location on the new server shortly. In the meantime, it can be viewed here: http://tripleplaycan.brinkster.net/soup/

Nathaniel

Posts: 453
Joined: December 10th, 2008, 3:48 pm

Re: The Online Life-Like CA Soup Search

Hello,

An (in)auspicious first post, possibly: when do you think the soup server will be back? I am looking forward to contributing cycles.

KR,

Posts: 4
Joined: June 15th, 2011, 5:44 am

Re: The Online Life-Like CA Soup Search

Hello,

An (in)auspicious first post, possibly: when do you think the soup server will be back? I am looking forward to contributing cycles.

KR,

Unfortunately, the search script itself was fairly primitive in nature, and Nathaniel has said that he will not be bringing it back in the near future. If you have ideas for how to write a better script, then maybe it will come back.
137ben

Posts: 343
Joined: June 18th, 2010, 8:18 pm

Re: The Online Life-Like CA Soup Search

Unfortunately, while I grasp the concept of Life just fine, I am clueless about the intricacies of many things discussed or built. Neither do I know python. I do however like to watch patters evolve and like distributed search programs so I can donate cycles.

is there another program active that performs useful Life activity that I could donate cycles to?

EDIT: stand-alone search scripts that I could run on my machines, of which any useful output i'd have to send somewhere, would work too.

Thanks.

Posts: 4
Joined: June 15th, 2011, 5:44 am

Re: The Online Life-Like CA Soup Search

BadKarma wrote:Stand-alone search scripts that I could run on my machines, of which any useful output i'd have to send somewhere, would work too.

There is a game of life census screensaver which does the same as what the Soup Search script did but is limited to the standard life rule. I brought up the idea in this topic (http://www.conwaylife.com/forums/viewtopic.php?f=7&t=671), along with a download link. The screensaver can generate results on its own, or several result files from different computers can be combined into one larger set of statistics, what I'm currently doing.
If you're running the program on several computers, I'd recommend setting each one to run soups of different densities so the resulats can be merged easily.

Lewis

Posts: 320
Joined: March 17th, 2009, 5:26 pm
Location: UK

Re: The Online Life-Like CA Soup Search

Had a look: ran the screensaver. Interesting. It is 'truly' a census, correct? Thank you for the suggestion!

I must say that I find the methuselah_density script from nathaniel's blog more interesting in that it actually generates 'starting positions' (.rle's) for insteresting object (outliers). More 'fun factor', if you will. Not to disregard your suggestion at all, mind.

Posts: 4
Joined: June 15th, 2011, 5:44 am

Re: The Online Life-Like CA Soup Search

BadKarma wrote:I must say that I find the methuselah_density script from nathaniel's blog more interesting in that it actually generates 'starting positions' (.rle's) for insteresting object (outliers). More 'fun factor', if you will. Not to disregard your suggestion at all, mind.

The LifeSS Screensaver saves the starting seeds for any patterns which go over 20,000 generations or have a final population of over 3000 cells. (The files are saved somewhere in C:\lifess\_found ). It also saves any soup which turns up a new pattern it hasn't encountered yet. When the edge is set to 20x20, many of the methuselahs can be trimmed down to about 30 cells (sometimes less).

Lewis

Posts: 320
Joined: March 17th, 2009, 5:26 pm
Location: UK

Re: The Online Life-Like CA Soup Search

Awesome. Will activate it, then!

Posts: 4
Joined: June 15th, 2011, 5:44 am

Re: The Online Life-Like CA Soup Search

Are there any old copies of the soup search script that I could try to modify for use with 22da, which is a hexagonal rule? (I want to do a soup search in 22da for new oscillators and/or spaceships, and I do not want to start over from scratch. I do not want an already compiled program, either.)
c0b0p0

Posts: 645
Joined: February 26th, 2014, 4:48 pm

Previous