### Re: Thread for basic questions

Posted:

**January 26th, 2017, 11:48 am**It's a dirty lightspeed wave.

Forums for Conway's Game of Life

http://www.conwaylife.com/forums/

Page **12** of **66**

Posted: **January 26th, 2017, 11:48 am**

It's a dirty lightspeed wave.

Posted: **January 29th, 2017, 5:58 am**

Can we find a glider fleet that can synthesis itself?

Posted: **January 29th, 2017, 6:00 am**

GUYTU6J wrote:Can we find a glider fleet that can synthesis itself?

First, that would be an oscillator UNLESS it synthesizes itself in a different part of the universe, then, it would be an amazing spaceship.

In theory, it is possible.

Posted: **January 29th, 2017, 7:03 am**

What is the slowest spaceship in a b2a rule?

Posted: **January 29th, 2017, 8:32 am**

Are there any simple reflectors where if the glider is shifted 1 glider lane it passes by the reflector unobstructed?

Posted: **January 29th, 2017, 10:52 am**

Goldtiger997 wrote:Are there any simple reflectors where if the glider is shifted 1 glider lane it passes by the reflector unobstructed?

Simple ones, no. That would be nice to have.

Such things are called highway robbers, and I think currently the smallest and fastest ones are these -- 886 ticks to recover, or 951 ticks if you want to be able to use the output glider without adding another highway robber...!

Posted: **January 29th, 2017, 11:46 am**

Saka wrote:GUYTU6J wrote:Can we find a glider fleet that can synthesis itself?

First, that would be an oscillator UNLESS it synthesizes itself in a different part of the universe, then, it would be an amazing spaceship.

In theory, it is possible.

It's not even all that difficult in practice nowadays. The only difficulty is that no synthesis is going to get done unless the glider fleet runs into something, like a block. Otherwise it's just going to keep going out into empty space -- and how exactly is a block going to get out there ahead of the fleet?

-- Well, somehow we'll have to construct a series of one-time turners, such that the final result is a glider coming in at right angles to the beginning of the fleet, so it hits a leading glider and makes an appropriate block.

Once we have a block target to work with, it's not as hard as it looks, now that we have single-channel universal construction. Oddly enough, it might be easiest to have every glider in the fleet on the same lane -- except maybe one.

The Plan

Imagine a long line of gliders, divided into two pairs of identical sub-recipes with a target glider in between on a different lane (explained later): Recipe 1, Recipe 2, long-gap, target-glider, Recipe 1, Recipe 2. The long gap is just about as long as Recipes 1 and 2 combined.

- When recipe 1 hits a block, it single-channel constructs a 180-degree reflector on the input lane, with an extra output lane that can be blocked by triggering an eater seed.
- When recipe 2 comes along, it is reflected back towards the source of the glider stream.
- When recipe 2 has all been reflected, the first glider in recipe 2 hits the target glider, to make a block target.

Recipe 2 uses the block target to construct another 180-degree reflector, with a transparent output lane (e.g., an Fx77). - Also included in Reflectors #1 and #2 are a few one-time turners similar to the self-destruct mechanism in the Demonoid, that can temporarily block the output lane, and also create a new target glider at the right time.
- The next Recipe 1, Recipe 2 pair arrive eventually. They're reflected by the first reflector, and a copy escapes on the output lane.
- When Recipe 1 hits Reflector #2, the first glider triggers the one-time turner chain that sends a glider ahead, back to Reflector #1, in time to block the output lane.
- After that blocking eater is built at Reflector #1, another one-time glider heads back to Reflector #2, to be reflected by a 180-degree one-time turner (e.g., two blocks or a long boat) and travel back to shoot down the eater and release a new target glider parallel to the output lane.
- The eater destruction sends *another* glider -- let's call it Glider D -- back to Reflector #2, to be reflected 180 degrees again.
- Meanwhile, a second copy of Recipe #1 and #2 escapes from the output lane, following the first copy and the target glider.
- That completes the construction of a copy of the original glider fleet. With a little extra one-time circuitry, we could open and close another output channel, resulting in two copies of the fleet heading off in opposite directions.
- Now we just have to get rid of the glider loop that's been left behind.
- Glider D triggers a one-time circuit the builds an eater in the input lane of Reflector #2, and sends another glider D' back to Reflector #1.
- Glider D' triggers a self-destruct mechanism at Reflector #1, which produces a glider D''.
- Glider D'' travels back to trigger the self-destruct mechanism at Reflector #2.
- Done (finally).

That Awkward Problem, Revisited

But again, how do we know there's a target block out there in empty space? Is this a Sparse Life universe, and we're just going to hope the glider stream encounters something in exactly the right location eventually?

Maybe the first part of Recipe #1 could be designed to converge to the same target, starting from several possible initial pieces of junk -- but it seems unlikely that we could handle more than a half-dozen cases, and there are a lot more block and blinker placements than that, all about equally likely.

There's Probably A Simpler Way

Did I make my design unnecessarily complicated? It would be nice to be able to avoid the large gap in the middle of the glider fleet, but I don't see how to build a functional glider loop to catch the second half of the fleet, without having to wait for the second reflector's construction to be completed.

You could build Reflector #2 first with no problem, since it's transparent to the incoming stream, but then how do you get a target to the location of Reflector #1? Could build and trigger a Cordership seed before starting to build Reflector #2, I suppose -- but Corderships are so slow that you might still need to wait around for a while before sending the recipe to build Reflector #1.

-- That could work, I guess, but I think it wouldn't be easy to get the spacing right between the output pairs of recipes. Somebody want to do the math on that option?

Posted: **February 1st, 2017, 11:25 pm**

How can I find a known synthesis?

Perhaps more broadly, is there a list of resources for finding known synthesis? These are the ones I know of:

Sometimes searching for an object by name (or other reference) just doesn't turn up any results and it's not in one of the other collections. For example, I posted a synthesis of a 17-bit still life on the Unsure discoveries thread. It's more than likely suboptimal, but I don't know how to find out what the current optimal synthesis is, nor do I have enough experience with glider synthesis to work out which smaller objects I should look at to find the "standard techniques" which will solve this particular problem.

Many thanks to all those who have contributed to the above resources.

Perhaps more broadly, is there a list of resources for finding known synthesis? These are the ones I know of:

- LifeWiki which contains syntheses for many well known objects and for other notable patterns.
- Mark Niemiec's Life Page - A fantastic database containing all small objects with population up to 18 cells as well as many more up to 25 cells. Syntheses for all objects up to 15 cells and for many larger objects are included. In the case of 16 and 17 cell objects, those without syntheses in the database can reportedly "be synthesized using standard techniques from smaller still-lifes."
- Chris Cain's Glider Synth project which includes syntheses for almost all still life objects with a population up to 16 cells and some larger ones included as well.
- These forums - Syntheses for many different objects are littered throughout the discussion threads (and concentrated in a few in particular) and if you're lucky you can find what you're looking for with the forum search.

Sometimes searching for an object by name (or other reference) just doesn't turn up any results and it's not in one of the other collections. For example, I posted a synthesis of a 17-bit still life on the Unsure discoveries thread. It's more than likely suboptimal, but I don't know how to find out what the current optimal synthesis is, nor do I have enough experience with glider synthesis to work out which smaller objects I should look at to find the "standard techniques" which will solve this particular problem.

Many thanks to all those who have contributed to the above resources.

Posted: **February 3rd, 2017, 9:23 am**

Here's a challenge: The smallest breeder we know has 23 cells. What's the smallest breeder in terms of bounding box?

Posted: **February 3rd, 2017, 9:35 am**

gameoflifemaniac wrote:Here's a challenge: The smallest breeder we know has 23 cells. Find the smallest breeder in terms of bounding box.

Which one has 23 cells? And this is the basic questions thread not the basic challenges thread. Reformat your post.

Posted: **February 3rd, 2017, 10:06 am**

Saka wrote:Which one has 23 cells?

I think he means the switch engine ping-pong.

There certainly is some kind of quadratic growth pattern that is smaller than the 27X27 Max.

Posted: **February 3rd, 2017, 10:09 am**

Yes

Posted: **February 5th, 2017, 4:18 am**

I've been getting back into Life and CA study after a long hiatus and have returned to see a lot of rulestrings of the form B3/S23aqz... with letters and dashes added to the B.../S... format. Is there a script for running these? Where does this format originate and what does it mean?

Posted: **February 5th, 2017, 5:16 am**

twinb7 wrote:I've been getting back into Life and CA study after a long hiatus and have returned to see a lot of rulestrings of the form B3/S23aqz... with letters and dashes added to the B.../S... format. Is there a script for running these? Where does this format originate and what does it mean?

It's Alan Hensel's "neighbors2" rule notation for isotropic non-totalistic rules, and it has been built in to Golly 2.8 while you weren't looking. There is a script, but it's better to just download the latest Golly.

(The new Lua overlay feature in the Golly 2.9 beta might be even more interesting than the isotropic rule support, so maybe just download that instead.)

Posted: **February 6th, 2017, 10:37 pm**

How do I get ntgfind and how hard is it to compile (i've heard it's a different compilation for every rule, just like ntzfind)?

sorryifthisisareallynewbiequestionhhhhhhhhh

idon'tknowwhichtopictopostthisineither

sorryifthisisareallynewbiequestionhhhhhhhhh

idon'tknowwhichtopictopostthisineither

Posted: **February 7th, 2017, 12:40 pm**

drc wrote:How do I get ntgfind and how hard is it to compile (i've heard it's a different compilation for every rule, just like ntzfind)?

gfind for non-totalistic rules is not yet in a polished state. Here are the instructions to compile gfind for a given isotropic rule specified in the version of Hensel notation used in Golly:

- Read this post by EricG in the adapting gfind thread
- Download the file SimpleInverseFire-v2.c and the script For-gfind-from-HenselNotation.py
- Convert rule specification from Golly's variant to the one EricG used: swap 'r' and 'y'; replace 'n' with 'v'
- Run the python script and specify the required rule.
- Follow EricG's instructions, steps 2 & 3EricG wrote:2) Open SimpleInverseFire-v2.c, and search for the string "EDG Patch" until you find the long list of RuleTab entries. Delete these entries, and replace them with the output from one of the gfind-related scripts.

3) Rename and recompile the new version of SimpleinverseFire-v2.c. When the proram asks for a rule, type in any totalistic rule (eg B3/S23). - Profit?

There's a bug in the script with S8: remove any duplicate mME values if you use S8.

There's also no symmetry checking for the isotropic rules in gfind. I try to specify a totalistic rule to gfind which matches the allowed symmetries of the isotropic rule. Due to this you may get results which are non-working ships in some cases.

Posted: **February 8th, 2017, 11:34 am**

calcyman wrote:I predict that someone will find a synthesis for a 2017-bit still life, using no more than 50 gliders, before the end of this year.

That will be more than 40 bits per glider.Have we found any synthesis that is so effective?

Posted: **February 8th, 2017, 12:11 pm**

GUYTU6J wrote:calcyman wrote:I predict that someone will find a synthesis for a 2017-bit still life, using no more than 50 gliders, before the end of this year.

That will be more than 40 bits per glider.Have we found any synthesis that is so effective?

Our most-efficient known synthesis is the eight-glider-per-bit (40/5) synthesis of omnibus, I'm pretty sure.

Unless we can find some repeatable component somehow fireable from guns, we can't reach that goal.

Posted: **February 8th, 2017, 1:49 pm**

BlinkerSpawn wrote:GUYTU6J wrote:calcyman wrote:I predict that someone will find a synthesis for a 2017-bit still life, using no more than 50 gliders, before the end of this year.

That will be more than 40 bits per glider.Have we found any synthesis that is so effective?

Our most-efficient known synthesis is the [eight-bit-per-glider] (40/5) synthesis of omnibus, I'm pretty sure.

Where is that 5-glider synthesis, anyway? I'm not finding it with any Google or conwaylife.com search on "omnibus".

BlinkerSpawn wrote:Unless we can find some repeatable component somehow fireable from guns, we can't reach that goal.

The trick that came to my mind was to synthesize something like a boatstretcher -- e.g., start with a crab and add a boat just behind it... The crab synthesis is only 18 gliders, but is there any hope of sneaking a boat in there after the crab is constructed, maybe with colliding *WSSes?

Then look for a way to kill off the crab without starting a chain reaction back down the long^k boat. That would give a still life with N bits for any N, using probably somewhere around 50 gliders if it works -- assuming you can use a tub instead of a boat for even bit counts, or add a bit at the other end.

Could do something similar with a linestretcher -- there have been Corderrakes that can pull off that trick for a long time... but that would be a lot more than 50 gliders to construct, and stopping at a specific length would be trickier, so you'd probably just have to decorate one end of the line or the other with a few extra cells instead, to get to 2017.

-- Hmm, and look who the last person was who was thinking about synthesizing tubstretchers, within a week of the above prediction...!

Posted: **February 9th, 2017, 6:08 am**

dvgrn wrote:-- Hmm, and look who the last person was who was thinking about synthesizing tubstretchers, within a week of the above prediction...!

Okay, you got me; it's a fair cop...

It's relatively straightforward to halt a tubstretcher using two gliders, and add a bit with a further two gliders, so a clean 46-glider tubstretcher synthesis would be sufficient to build a long^1006 boat in 50 gliders.

Posted: **February 9th, 2017, 11:51 am**

calcyman wrote:dvgrn wrote:-- Hmm, and look who the last person was who was thinking about synthesizing tubstretchers, within a week of the above prediction...!

Okay, you got me; it's a fair cop...

It's relatively straightforward to halt a tubstretcher using two gliders, and add a bit with a further two gliders, so a clean 46-glider tubstretcher synthesis would be sufficient to build a long^1006 boat in 50 gliders.

The tubstretcher-destruction and bit-adding can be combined, so we have 48 gliders to work with:

Code: Select all

`x = 21, y = 20, rule = B3/S23`

obo$b2o$bo5bo$8b2o$7b2o$16b2o$16bobo$16bo$19bo$19b2o$19b2o$18b2o$9b2o

6b2o$9bobo3b2o2bo$9bo5b3o$12bo$12b2o$16bo$15bobo$16bo!

Posted: **February 10th, 2017, 4:58 am**

GUYTU6J wrote:Can we find a glider fleet that can synthesis itself?

If a single glider counts, yes.

If you mean multiple, it would probably be a big complex spaceship. Which would be pretty nice.

Posted: **February 12th, 2017, 8:58 am**

BlinkerSpawn wrote:GUYTU6J wrote:

That will be more than 40 bits per glider.Have we found any synthesis that is so effective?

Our most-efficient known synthesis is the eight-glider-per-bit (40/5) synthesis of omnibus, I'm pretty sure.

Unless we can find some repeatable component somehow fireable from guns, we can't reach that goal.

And now we have a 40-glider synthesis of a 2017-bit still-life (thanks Kazyan and BlinkerSpawn!):

Code: Select all

`x = 2165, y = 2166, rule = B3/S23`

o$b2o$2o$6bobo$7b2o$7bo1997$2150bobo$2150b2o$2151bo$2019bo$2020bo138bo

$2018b3o137bo$2158b3o3$2161bo$2159b2o$2010bo149b2o$2011bo4bobo$2009b3o

5b2o$2017bo4$2134bo$2132b2o$2133b2o2$2010bo9bo$2011bo9bo$2009b3o7b3o3$

2039bo$2029bo10bo92bo$2027bobo8b3o90b2o$2028b2o102b2o4$2025bo$2026bo$

2024b3o$2124bo$2124bobo$2124b2o6$2058bo$2056bobo$2057b2o14$2101bo$

2100bo$2100b3o14$2094bo$2094bobo$2078bo15b2o$2079b2o$2078b2o12$2101bo$

2100b2o$2100bobo19$2043b2o$2042bobo$2044bo2$2129b2o$2128b2o$2130bo6bo$

2035b2o99b2o$2034bobo99bobo$2036bo13$2016b3o10b2o$2018bo9bobo$2017bo

12bo2$2013b2o$2014b2o140b3o$2013bo142bo$2157bo2$2005b2o153bo$2006b2o

151b2o$2005bo153bobo4$2140b3o13b2o5b2o$2140bo15bobo3b2o$2141bo14bo7bo

2$2024bo134b3o$1993bo30b2o133bo$1993b2o24b2o2bobo134bo$1992bobo23bobo

136bo$2020bo135b2o$2156bobo$2152b2o$2151b2o$2153bo!

Posted: **February 12th, 2017, 12:04 pm**

calcyman wrote:And now we have a 40-glider synthesis of a 2017-bit still-life (thanks Kazyan and BlinkerSpawn!)\

Now 39, by chris_c. Congratulations to all of you, and thank you Calcyman for making my prediction come true as well as yours. It worked out much better than my prediction about never again seeing a continuous range of bit counts from xs4 to xsN in Catagolue C1 symmetry.

-- That prediction is nicely back in the money again now, though, if I can just figure out how to disqualify the less-than-three-month period last year when I forgot to feed the shoggoth. Nowadays it's going to be a really tough job filling all the gaps up to 56 bits, before some even bigger even-bit-count symmetric thing shows up.

I [was] right in principle, and Catagolue only just happened to prove me wrong in practice.

Posted: **February 13th, 2017, 10:54 am**

GUYTU6J wrote:Saka wrote:Which one has 23 cells?

I think he means the switch engine ping-pong.

There certainly is some kind of quadratic growth pattern that is smaller than the 27X27 Max.

Ah! Max! You're absolutely right.