Thread for basic questions

[dvgrn]

A impressively basic question: how GUYTU6J find those waves?

By creative use of a "repeat script" written by praosylen in 2016.

[KittyTac]

Is a rule table metacell possible?

Scary thought. You mean, a Life metacell with programmable elements that can support variables and permutation symmetry and 256 states and nearly unlimited numbers of rule lines?

It's certainly theoretically possible, but to be able to support a memory storage area big enough to hold the maximum number of rule lines, it would have to be so big and have such a high period that even StreamLife couldn't run it.

Of course, Golly can't run a rule table with the maximum number of rule lines on any existing computer, either. So if you set some very specific limitations, like no variables, symmetry=none only, and maximum of a thousand rule lines, then it's ... well, still extraordinarily impractical, since every cell would have to include its own rule-table processing circuitry, but ultimately just another kind of specialized Life computer.

Yes, I meant a cut-down version of rule-tables. It was kind of a silly question.

[Hunting]

A impressively basic question: how GUYTU6J find those waves?

By creative use of a "repeat script" written by praosylen in 2016.

Thanks!

[Bamboonium]

So what keeps just any pattern that lasts for long enough and eventually leads to a pattern that is stable from being a methuselah, or are there just infinite methuselahs? Also what separates something like a B-heptomino from this thing I found?

Regards,
Bamboonium

Code: Select all

x = 8, y = 5, rule = B3/S23
$3bo$2bobo$4bo$o4b3o!

[dvgrn]

So what keeps just any pattern that lasts for long enough and eventually leads to a pattern that is stable from being a methuselah, or are there just infinite methuselahs?

There are more or less infinite methuselahs, on a declining spectrum of interestingness. The umpteen bajillion ones with relatively large bounding boxes and/or relatively large populations are all about equally interesting.

Which is to say, they have to share one unit of interestingness between them all, so they're only one umpteen-bajillionth as interesting as the smallest-bounding-box smallest-population methuselah (for a given number of ticks of activity before stability).

Also what separates something like a B-heptomino from this thing I found?

The Thing-You-Found is made up of two polyplet clusters, except that the single cell on the left doesn't affect anything and might as well not be there, so what you're really found is another polyplet like your B-heptomino predecessor. This one is a predecessor of blockade, and after two ticks it shares common descendants with the stairstep hexomino, and the whole evolutionary sequence is somewhat confusingly known as "LoM", which is short for lumps of muck.

[Bamboonium]

Thank You! I am relatively new to game of life, and I'm still learning what things are. The Wiki links really help as well! :D

Regards,
Bamboonium

[jwk]

In this list of glider syntheses for still lifes of at most 16 bits, what do the left and middle columns represent (I assume they're identifiers of each unique still life) and how are they calculated / named?

[Apple Bottom]

In this list of glider syntheses for still lifes of at most 16 bits, what do the left and middle columns represent (I assume they're identifiers of each unique still life) and how are they calculated / named?

First of all, welcome to the forums!

That said, which list?

Pattern naming in general is described (rudimentarily) in this LifeWiki entry. Still lifes are sometimes identified by numbers, e.g. 16.1145 ("Eater hat bridge hat" in Mark Niemiec's database), but different pattern repositories (may) use different numbers of the same object. apgcodes unambiguously identify various classes of objects (including still lifes), and also encode the pattern itself, allowing it to be reconstructed from the code. If you're seeing objects being referred to using weird strings like xs16_4aarzwc453, those are apgcodes.

[jwk]

In this list of glider syntheses for still lifes of at most 16 bits, what do the left and middle columns represent (I assume they're identifiers of each unique still life) and how are they calculated / named?

First of all, welcome to the forums!

That said, which list?

Pattern naming in general is described (rudimentarily) in this LifeWiki entry. Still lifes are sometimes identified by numbers, e.g. 16.1145 ("Eater hat bridge hat" in Mark Niemiec's database), but different pattern repositories (may) use different numbers of the same object. apgcodes unambiguously identify various classes of objects (including still lifes), and also encode the pattern itself, allowing it to be reconstructed from the code. If you're seeing objects being referred to using weird strings like xs16_4aarzwc453, those are apgcodes.

Thanks for the welcome!

My bad, I was supposed to paste this list: https://github.com/ceebo/glider_synth/b ... ummary.txt

I found the apgcode page and understood it, so thank you for that. I was unable to find anything on the numbering system (e.g. 16.1145 as you said). Would you mind explaining it or pointing me towards somewhere I could understand it?

[Apple Bottom]

My bad, I was supposed to paste this list: https://github.com/ceebo/glider_synth/b ... ummary.txt

I found the apgcode page and understood it, so thank you for that. I was unable to find anything on the numbering system (e.g. 16.1145 as you said). Would you mind explaining it or pointing me towards somewhere I could understand it?

Oh yes, I think those are the numbers used in Mark Niemiec's database. Mark's site resides here, and there's a search interface here where you can paste in patterns as RLE.

There's (much) more data behind the site, but to my knowledge it's not public. So while it's possible to e.g. get the "Niemiec numbers" for all all 16-bit still lifes, the same isn't true for larger still lifes that Mark has also enumerated and numbered. Perhaps Mark can be persuaded to open access; storage space and bandwidth (read: transfer caps) are not nearly as much of an issue anymore in this day and age as they used to be.

Anyhow, as for the numbers as such, I think they're largely arbitrary, an artifact of the tools Mark used to enumerate these still lifes. They don't mean anything as such (unlike apgcodes).

[Ian07]

What's the deal with the Octagon 2 and 4 oscillators? What happened to 1 and 3?

[dani]

What's the deal with the Octagon 2 and 4 oscillators? What happened to 1 and 3?

Octagon 1 is the pond, and octagon 3 does nothing of interest. Rather than being named after the order they were found, they were named after the size octagon they are.

[dani]

What's the maximum lifespan of a pattern with 1 cells but not infinite in golly?

A cell in //256 stabilizes at 255 {fixed}

But obviously ruletables could make this better.

[praosylen]

What's the maximum lifespan of a pattern with 1 cells but not infinite in golly?

This stabilizes at generation 255:

Code: Select all

x = 88, y = 100, rule = B3/S23
22bo$22bo$22bo6$29b3o4$21b3o2$38bo$38bo$38bo2$30bo$30bo$30bo2$45b3o2$
8b2o$2bo5b2o$obo34b3o$b2o$54bo$54bo$54bo2$15b2o29bo$15b2o29bo$6b2o38bo
$6b2o$61b3o4$22b2o29b3o$22b2o$13b2o55bo$13b2o55bo$70bo2$62bo$62bo$29b
2o31bo$29b2o$20b2o55b3o$20b2o3$69b3o2$36b2o48bo$36b2o48bo$27b2o57bo$
27b2o$78bo$78bo$78bo2$43b2o$43b2o$34b2o$34b2o$85b3o4$50b2o$50b2o$41b2o
$41b2o5$57b2o$57b2o$48b2o$48b2o5$64b2o$64b2o$55b2o$55b2o7$62b2o$62b2o!

But obviously ruletables could make this better.

Wrong RLE, I believe.

[AforAmpere]

What's the maximum lifespan of a pattern with 1 cells but not infinite in golly?

Probably at least this many steps:

Code: Select all

2^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^3

You could make one cell print out a pattern in KnuthArrows, modifying the rule to not have results collapse back into a spaceship. Then you could use up to around 200 more states, printing out a pattern that would be a valid pattern in that rule, which would evolve for over the time above. However, this is probably a weak lower bound, as I'm sure there is a way to make a computer with one cell that calculates something like Loader's Number, which would blow the previous example out of the water.

[bprentice]

Consider a rule that runs every possible Life pattern and is also omni-periodic. Would this forum drop its study of Life in favor of such a rule? If not why not?

This leads to a more general question. Are there any rules that have been discussed in this forum that deserve their own board?

Brian Prentice

[KittyTac]

Consider a rule that runs every possible Life rule and is also omni-periodic. Would this forum drop its study of Life in favor of such a rule? If not why not?

What do you mean by "runs every possible Life rule"?

[bprentice]

KittyTac,

Sorry, I meant every possible Life pattern!

Brian Prentice

[KittyTac]

KittyTac,

Sorry, I meant every possible Life pattern!

Brian Prentice

Like extendedlife? As you can see, we did not drop research for Life. Why? We need omniperiodicity specifically in Life. For science!

[wwei23]

The answer is no. I made a rule called Symbiosis once. It's two life states that keep each other permanently alive, but cels can only be born from the same color. As a result, in a few weeks, periods not known in Life were discovered. Did the forum drop life in favor of Symbiosis?

[bprentice]

Did the forum drop life in favor of Symbiosis?

I am aware of your Symbiosis rule of course, it is one of the reasons I started this discussion. Why did this forum not drop Life in favor of Symbiosis?

We need omniperiodicity specifically in Life. For science!

Why specifically? We have Symbiosis, a better rule.

Brian Prentice

[dani]

Symbiosis is not like life, in that Life is much simpler and 2-state.

I don't think this is a valid concern, however. There aren't any boards for rules that were very popular back in the day (tlife and Snowflakes, i.e.)

[bprentice]

Symbiosis is not like life, in that Life is much simpler and 2-state.

Are you equating "simpler and 2-state" to better? I believe there needs to be a serious discussion on what "better" means.

There aren't any boards for rules that were very popular back in the day (tlife and Snowflakes, i.e.)

My point exactly! Do either tlife or Snowflakes or any other rule that is discussed on this forum deserve their own board?

Brian Prentice

[M. I. Wright]

Life has precedent. I think that's about it, really: it was the first CA simple enough to (with some help from Martin Gardner et al) introduce the concept to the masses, which resulted in more foundational work being done on it, and from then on it's just been a positive-feedback loop of people being drawn to it and doing relatively-advanced work simply because people before them've done so as well.

Had Conway landed a dart on B36/S23, you might now be asking why we don't drop Life in favor of "LowLife" (which might not have our easy replicator but it looks to be a bit more lively overall and hey look at this one methuselah that almost gives birth to itself a few times you think we could make any conduits out of it?), and the reason would probably still be one of precedent without much more to it. There aren't many other reasons a single CA would deserve such a large following.

[KittyTac]

Also: 2-state rules are easier to search and simulate large objects in due to better performance.