Page 112 of 209
Re: Thread for basic questions
Posted: June 10th, 2021, 9:50 am
by wwei47
How would I automatically detect and remove wicks from a 50-by-100 torus?
Re: Thread for basic questions
Posted: June 10th, 2021, 9:57 am
by dvgrn
wwei47 wrote: ↑June 10th, 2021, 9:50 am
How would I automatically detect and remove wicks from a 50-by-100 torus?
... "and remove" means you want to keep whatever else is in the 50-by-100 torus that is not a wick? Are you only worried about stationary wicks, or are moving waves also worth worrying about?
... In other words, do you want to disallow infinite connected objects from some kind of soup search experiments? Why will your soups-or-whatever be producing wicks like that, often enough to bother worrying about? A little more context would be very helpful here.
Re: Thread for basic questions
Posted: June 10th, 2021, 10:06 am
by wwei47
dvgrn wrote: ↑June 10th, 2021, 9:57 am
... "and remove" means you want to keep whatever else is in the 50-by-100 torus that is not a wick? Are you only worried about stationary wicks, or are moving waves also worth worrying about?
... In other words, do you want to disallow infinite connected objects from some kind of soup search experiments? Why will your soups-or-whatever be producing wicks like that, often enough to bother worrying about? A little more context would be very helpful here.
Yes, keep everything else, removing moving waves too. The soups would be the board on lifecompetes.com. And I sometimes intentionally create wicks and waves on the lifecompetes server so I know they need to be handled. Sometimes, the entire board consists of nothing but agar-that also needs to be handled.
Re: Thread for basic questions
Posted: June 10th, 2021, 10:16 am
by dvgrn
wwei47 wrote: ↑June 10th, 2021, 10:06 am
Yes, keep everything else, removing moving waves too. The soups would be the board on lifecompetes.com. And I sometimes intentionally create wicks and waves on the lifecompetes server so I know they need to be handled. Sometimes, the entire board consists of nothing but agar-that also needs to be handled.
Yikes, it seems like not at all an easy problem. There's just too wide a range of possible wick directions and periods. If there also may be a scattering of other non-wick stuff, and you want to keep that other stuff, then it's hard for an automated analysis to find the period of the wick or wave as a first step toward identifying it.
Maybe somebody else can suggest a better algorithm, but mine would be "look at the pattern and delete any wicks and waves you see".
Re: Thread for basic questions
Posted: June 10th, 2021, 11:50 am
by hotdogPi
About higher symmetries in Catagolue: maybe discoveries should count if there's more than one of them (for some symmetries, you would need 4 or
, meaning the symmetry wasn't vital to its forming.
edit: silly emoji, that's supposed to be an 8
Re: Thread for basic questions
Posted: June 10th, 2021, 11:59 am
by ihatecorderships
Or maybe we can have a separate category, called "Discoveries in Higher Symmetries."
Re: Thread for basic questions
Posted: June 10th, 2021, 2:24 pm
by cgoler2
ihatecorderships wrote: ↑June 10th, 2021, 11:59 am
Or maybe we can have a separate category, called "Discoveries in Higher Symmetries."
I don't think so, since then it would probably include my discoveries in B2n3-ce4ky6in8/S234c, and my discoveries in several other rules.
Re: Thread for basic questions
Posted: June 10th, 2021, 2:40 pm
by bubblegum
cgoler2 wrote: ↑June 10th, 2021, 2:24 pm
it would probably include my discoveries in B2n3-ce4ky6in8/S234c, and my discoveries in several other rules.
The Discoveries section is only for B3/S23, so the Discoveries in Higher Symmetries category-to-be really doesn't have a reason to include other rules.
Re: Thread for basic questions
Posted: June 11th, 2021, 10:42 am
by wwei47
dvgrn wrote: ↑June 10th, 2021, 10:16 am
wwei47 wrote: ↑June 10th, 2021, 10:06 am
Yes, keep everything else, removing moving waves too. The soups would be the board on lifecompetes.com. And I sometimes intentionally create wicks and waves on the lifecompetes server so I know they need to be handled. Sometimes, the entire board consists of nothing but agar-that also needs to be handled.
Yikes, it seems like not at all an easy problem. There's just too wide a range of possible wick directions and periods. If there also may be a scattering of other non-wick stuff, and you want to keep that other stuff, then it's hard for an automated analysis to find the period of the wick or wave as a first step toward identifying it.
Maybe somebody else can suggest a better algorithm, but mine would be "look at the pattern and delete any wicks and waves you see".
I don't care about wick/wave identification. To be specific, as far as I'm concerned:
This is a wick.
Code: Select all
x = 5, y = 50, rule = B3/S23:T100,50
o3bo$4bo$bo2bo$3b2o$2b3o$2b3o$3b2o$bo2bo$4bo$o3bo$o3bo$4bo$bo2bo$3b2o
$2b3o$2b3o$3b2o$bo2bo$4bo$o3bo$o3bo$4bo$bo2bo$3b2o$2b3o$2b3o$3b2o$bo2b
o$4bo$o3bo$o3bo$4bo$bo2bo$3b2o$2b3o$2b3o$3b2o$bo2bo$4bo$o3bo$o3bo$4bo
$bo2bo$3b2o$2b3o$2b3o$3b2o$bo2bo$4bo$o3bo!
This is a wick.
Code: Select all
x = 3, y = 49, rule = B3/S23:T100,50
b2o$b2o2$o$o$o$o$o$o2$b2o$b2o2$o$o$o$o$o$o2$b2o$b2o2$o$o$o$o$o$o2$b2o
$b2o2$o$o$o$o$o$o2$b2o$b2o2$o$o$o$o$o$o!
This is not a wick:
Code: Select all
x = 95, y = 4, rule = B3/S23:T100,50
b2o8b2o8b2o8b2o8b2o8b2o8b2o8b2o8b2o8b2o$4o6b4o6b4o6b4o6b4o6b4o6b4o6b4o
6b4o6b4o$2ob2o5b2ob2o5b2ob2o5b2ob2o5b2ob2o5b2ob2o5b2ob2o5b2ob2o5b2ob2o
5b2ob2o$2b2o8b2o8b2o8b2o8b2o8b2o8b2o8b2o8b2o8b2o!
This is a wick:
Code: Select all
x = 96, y = 29, rule = B3/S23:T100,50
43b2o48b2o$17bo25bo23bo25bo$17b3o21bobo23b3o21bobo$20bo20b2o27bo20b2o
$19b2o48b2o2$o49bo$3o47b3o$3bo49bo$2b2o11b2o11bo23b2o11b2o11bo$15b2o9b
3o36b2o9b3o$26bobo47bobo$26bo49bo4$bo49bo$bobo47bobo$b3o36b2o9b3o36b2o
$3bo23b2o11b2o11bo23b2o11b2o$28bo49bo$25b3o47b3o$25bo49bo2$44b2o48b2o
$16b2o27bo20b2o27bo$16bobo23b3o21bobo23b3o$18bo23bo25bo23bo$18b2o48b2o
!
This is not a wick:
Code: Select all
x = 14, y = 50, rule = B3/S23:T100,50
7bo$7b3o$10bo$7b2obo$8bob2o$8bo$5b2obo$b2obobob2o$2bobo$2bo2bo$b2obob
ob2ob2o$o3b2o3bobo$b3o5bobo$4b5obo$b2o$ob2o2b7o$o12bo$b3ob4ob3o$3b8o3$
6b2o$4bo4bo2$3bo6bo$b3obob2ob3o$o4bo3bo3bo$b5o2b2obobo$6bo2bob2o$3bob
5o$2bobo5b3o$2bobo3b2o3bo$b2ob2obobob2o$8bo2bo$9bobo$7bobob2o$7b2o$5b
2o$6bo$5bo$5b2o$6bo$5bo$5b2o$6bo$5bo$5b2o$6bo$5bo$5b2o!
This is a wick:
Code: Select all
x = 14, y = 50, rule = B3/S23:T100,50
7bo$7b3o$10bo$7b2obo$8bob2o$8bo$5b2obo$b2obobob2o$2bobo$2bo2bo$b2obob
ob2ob2o$o3b2o3bobo$b3o5bobo$4b5obo$b2obo2bo$obob2o2b5o$o3bo3bo4bo$b3o
b2obob3o$3bo6bo2$4bo4bo2$4bo4bo2$3bo6bo$b3obob2ob3o$o4bo3bo3bo$b5o2b2o
bobo$6bo2bob2o$3bob5o$2bobo5b3o$2bobo3b2o3bo$b2ob2obobob2o$8bo2bo$9bo
bo$7bobob2o$7b2o$5b2o$6bo$5bo$5b2o$6bo$5bo$5b2o$6bo$5bo$5b2o$6bo$5bo$
5b2o!
Re: Thread for basic questions
Posted: June 11th, 2021, 10:58 am
by dvgrn
wwei47 wrote: ↑June 11th, 2021, 10:42 am
I don't care about wick/wave identification. To be specific, as far as I'm concerned:
This is a wick...
This is a wick...
This is not a wick...
This is a wick..
This is not a wick...
This is a wick...
So far this is doing a wonderful job of proving my contention that the correct algorithm to run here is
"look at the pattern and delete any wicks and waves you see". There's no easy automated analysis algorithm that will reliably distinguish some of these things from each other.
Your last two examples seem like particularly good test cases. Your definition seems to be that the one that counts as a wick is actually changing the adjacent part of the pattern (and vice versa) -- it's not just making sparks that share the same space but don't actually affect each other.
I guess it's not impossible to write something that does this level of analysis, but it's pretty subtle -- which is presumably why non-interacting pieces with overlapping reaction envelopes are still one of the weaknesses of Catagolue's censusing code.
... Similar to the problem of designing a self-replicator that can defend itself against attacks, if you gave me the choice of either writing the algorithm, or building test patterns that would break somebody else's algorithm, I would definitely choose the second option. Without a clear way of algorithmically describing all possible classes of wicks and waves that are supposed to be found and removed, this seems like a really painfully difficult task -- the kind of thing someone could work on for months, and it still wouldn't give really accurate results in all cases.
Re: Thread for basic questions
Posted: June 11th, 2021, 11:16 am
by wwei47
dvgrn wrote: ↑June 11th, 2021, 10:58 am
... Similar to the problem of designing a self-replicator that can defend itself against attacks, if you gave me the choice of either writing the algorithm, or building test patterns that would break somebody else's algorithm, I would definitely choose the second option. Without a clear way of algorithmically describing all possible classes of wicks and waves that are supposed to be found and removed, this seems like a really painfully difficult task -- the kind of thing someone could work on for months, and it still wouldn't give really accurate results in all cases.
I realized something: If we double the "wicks" on themselves, the left one is still a single object but the right one is now two objects!
Code: Select all
x = 114, y = 100, rule = B3/S23:T200,100
7bo99bo$7b3o97b3o$10bo99bo$7b2obo96b2obo$8bob2o96bob2o$8bo99bo$5b2obo
96b2obo$b2obobob2o91b2obobob2o$2bobo97bobo$2bo2bo96bo2bo$b2obobob2ob2o
88b2obobob2ob2o$o3b2o3bobo88bo3b2o3bobo$b3o5bobo89b3o5bobo$4b5obo93b5o
bo$b2obo2bo93b2o$obob2o2b5o87bob2o2b7o$o3bo3bo4bo86bo12bo$b3ob2obob3o
88b3ob4ob3o$3bo6bo92b8o2$4bo4bo$106b2o$4bo4bo94bo4bo2$3bo6bo92bo6bo$b
3obob2ob3o88b3obob2ob3o$o4bo3bo3bo86bo4bo3bo3bo$b5o2b2obobo87b5o2b2ob
obo$6bo2bob2o93bo2bob2o$3bob5o93bob5o$2bobo5b3o89bobo5b3o$2bobo3b2o3b
o88bobo3b2o3bo$b2ob2obobob2o88b2ob2obobob2o$8bo2bo96bo2bo$9bobo97bobo
$7bobob2o94bobob2o$7b2o98b2o$5b2o98b2o$6bo99bo$5bo99bo$5b2o98b2o$6bo99b
o$5bo99bo$5b2o98b2o$6bo99bo$5bo99bo$5b2o98b2o$6bo99bo$5bo99bo$5b2o98b
2o$7bo99bo$7b3o97b3o$10bo99bo$7b2obo96b2obo$8bob2o96bob2o$8bo99bo$5b2o
bo96b2obo$b2obobob2o91b2obobob2o$2bobo97bobo$2bo2bo96bo2bo$b2obobob2o
b2o88b2obobob2ob2o$o3b2o3bobo88bo3b2o3bobo$b3o5bobo89b3o5bobo$4b5obo93b
5obo$b2obo2bo93b2o$obob2o2b5o87bob2o2b7o$o3bo3bo4bo86bo12bo$b3ob2obob
3o88b3ob4ob3o$3bo6bo92b8o2$4bo4bo$106b2o$4bo4bo94bo4bo2$3bo6bo92bo6bo
$b3obob2ob3o88b3obob2ob3o$o4bo3bo3bo86bo4bo3bo3bo$b5o2b2obobo87b5o2b2o
bobo$6bo2bob2o93bo2bob2o$3bob5o93bob5o$2bobo5b3o89bobo5b3o$2bobo3b2o3b
o88bobo3b2o3bo$b2ob2obobob2o88b2ob2obobob2o$8bo2bo96bo2bo$9bobo97bobo
$7bobob2o94bobob2o$7b2o98b2o$5b2o98b2o$6bo99bo$5bo99bo$5b2o98b2o$6bo99b
o$5bo99bo$5b2o98b2o$6bo99bo$5bo99bo$5b2o98b2o$6bo99bo$5bo99bo$5b2o98b
2o!
EDIT: For each potential wick/agar, put four copies of it into a 200-by-100 torus. Non-wicks will separate into four copies, while wicks will become two copies, and agars will still be one copy.
Re: Thread for basic questions
Posted: June 11th, 2021, 11:26 am
by dvgrn
wwei47 wrote: ↑June 11th, 2021, 11:16 am
I realized something: If we double the "wicks" on themselves, the left one is still a single object but the right one is now two objects!
To clarify:
you know that the right one is two objects, but there isn't necessarily an existing object census algorithm that can reliably separate a "pseudo-wick" like that into two objects -- correct?
It's hard to see how to algorithmically distinguish your case with the left-to-right traveling LWSSes (which you said was "not a wick") from your case with the left-to-right traveling Herschels (which you said
was a wick). In both cases, there are moving objects that traverse the entire length of the apparent wick. In both cases you can take one of the moving objects out without changing the behavior of the others. And in both cases, if you take _all_ of the moving objects out you definitely get something that you'll say is "not a wick".
So... why exactly
is the Herschel Fx77 chain a wick, but the LWSS stream is not?
Re: Thread for basic questions
Posted: June 11th, 2021, 11:35 am
by pcallahan
dvgrn wrote: ↑June 11th, 2021, 10:58 am
"look at the pattern and delete any wicks and waves you see"
That's not very scalable, though, unless you have a way to crowdsource it. And the "crowd" that could provide more signal than noise is going to be extremely small compared to the usual case.
I think I'm paraphrasing the same point you made, but the non-wicks had discontinuities in which the previous part of the pattern was not needed to stabilize its successor.
Would it be accurate to say all non-wicks can be made semi-infinite, but wicks cannot?
The Herschel conduit doesn't look like a wick to me. You can remove individual Herschels. I concede that you can't remove any of the conduit parts though. So it cannot be made semi-infinite.
It doesn't seem that hard to develop a wick identification algorithm that at least separates the cases above and a large class of generalizations. It may not be worth the effort, depending on how many of them you will need to look at.
Finally, the algorithm of "just look at it" is still potentially useful. Last weekend, I found myself using a technique I applied in the early 1990s, which is to set up thousands of collision experiments spaced so they don't interact, run a few hundred generations and observe the products, zoomed out far enough to see them all at once. In this case, I was looking for stabilized interactions of queenbee shuttles and catalysts, but didn't have the tools set up to do what I needed. It's quite easy to see (particularly with LifeHistory) whether the interaction is stable or if it explodes into a blob. You can also see if it releases any orthogonal spaceships because of their distinct trail.
It's a painful and slow process compared to automatic filters. However, I wonder if there are things we miss by not doing it. Is there anything to be discovered still just by setting up a lot of interactions and visually picking out a few that just don't look like the others? (And for that matter, is it possible a machine learning algorithm could also cluster them using such vague criteria?)
Update: It's hard to see how to algorithmically distinguish your case with the left-to-right traveling LWSSes (which you said was "not a wick") from your case with the left-to-right traveling Herschels (which you said was a wick).
Doesn't it work (as above) to observe that the spaceship train can be made semi-infinite but the Herschel conduit cannot?
Re: Thread for basic questions
Posted: June 11th, 2021, 11:47 am
by wwei47
dvgrn wrote: ↑June 11th, 2021, 11:26 am
To clarify:
you know that the right one is two objects, but there isn't necessarily an existing object census algorithm that can reliably separate a "pseudo-wick" like that into two objects -- correct?
Correct. The hard part is separating things like the quad pseudo still life.
At least one new post has been made to this topic. You may wish to review your post in light of this.
pcallahan wrote: ↑June 11th, 2021, 11:35 am
The Herschel conduit doesn't look like a wick to me. You can remove individual Herschels. I concede that you can't remove any of the conduit parts though. So it cannot be made semi-infinite.
If I remove a Herschel from the Fx77 chain, then the Herschel no longer behaves the same way in which it behaved before.
pcallahan wrote: ↑June 11th, 2021, 11:35 am
Update: It's hard to see how to algorithmically distinguish your case with the left-to-right traveling LWSSes (which you said was "not a wick") from your case with the left-to-right traveling Herschels (which you said was a wick).
Doesn't it work (as above) to observe that the spaceship train can be made semi-infinite but the Herschel conduit cannot?
If I remove an LWSS from the LWSS row, it still behaves the same way.
EDIT: The worrisome thing about object decomposition now is stuff like
https://catagolue.hatsya.com/object?apg ... rule=b3s23, where there's more than one way to do it, since I seem to have come up with an algorithm capable of decomposing the triple and quad pseudo still lives.
Re: Thread for basic questions
Posted: June 13th, 2021, 3:36 am
by bibunsekibun
Is there a program that generates random INT rules?
Re: Thread for basic questions
Posted: June 13th, 2021, 8:51 am
by dvgrn
bibunsekibun wrote: ↑June 13th, 2021, 3:36 am
Is there a program that generates random INT rules?
For some definitions of "random",
yes.
You may not find that a 50% chance of each isotropic bit being ON tends to give very interesting rules, unless you really like chaotic explosions. It's easy to change the random weightings in the script if you want to try to improve its ability to find good rules.
There are
other scripts that do a rule-by-rule random walk starting from Golly's current rule, which might be more entertaining to experiment with.
Re: Thread for basic questions
Posted: June 13th, 2021, 11:58 am
by MathAndCode
What is the rarest object that is created in a three-glider collision?
Re: Thread for basic questions
Posted: June 13th, 2021, 11:59 am
by hotdogPi
MathAndCode wrote: ↑June 13th, 2021, 11:58 am
What is the rarest object that is created in a three-glider collision?
pentadecathlon
Re: Thread for basic questions
Posted: June 13th, 2021, 12:03 pm
by MathAndCode
hotdogPi wrote: ↑June 13th, 2021, 11:59 am
pentadecathlon
They don't have to be created cleanly. For example, this counts:
Code: Select all
x = 35, y = 19, rule = B3/S23
2bo$2o$b2o3$4b3o$4bo$5bo9$33bo$32b2o$32bobo!
Re: Thread for basic questions
Posted: June 14th, 2021, 9:05 am
by hotdogPi
Frequency of evolutionary sequences can be compared to still lifes and oscillators by using the pulsar as reference. On the
list of common evolutionary sequences page, the honey farm is 49 times as common as the pulsar (taking into account the ones farther down the list), and we know how common the pulsar is in relation to other still lifes and oscillators. However, there are some short evolutionary sequences that we don't know their frequency.
What is the frequency of these ten short-sequence objects that don't qualify in the "list of evolutionary sequences" list due to never reaching 20 cells or not being above 20 for long enough, and am I missing any particularly common ones?
Code: Select all
x = 67, y = 72, rule = B3/S23
7$15bo31b2o$14b3o29bo2bo$46b4o10$46b3o$46bobo$14b4o28bo$14bo2bo29bo11$
12b2o2b3o29bo$12b2o2bobo28bobo$16b3o28b2o$47bo8$15bo$15bo30b3o$15bo30b
obo$14bo31bobo$15bo33bo$15bo$15bo8$17b2o26b2o$13b2obo2bo25bo2bo$12bo2b
ob2o26bob2o$13b2o!
EDIT: Forgot about the fleet and bakery
Re: Thread for basic questions
Posted: June 14th, 2021, 1:54 pm
by MathAndCode
hotdogPi wrote: ↑June 14th, 2021, 9:05 am
Am I missing any particularly common ones?
Should the following count?
Code: Select all
x = 5, y = 4, rule = B3/S23
2bo$o3bo$o3bo$o3bo!
Re: Thread for basic questions
Posted: June 15th, 2021, 12:00 am
by Citation needed
Hello. What is an "MTCC"? I did not find it on LifeWiki.
Re: Thread for basic questions
Posted: June 15th, 2021, 12:13 am
by ihatecorderships
Citation needed wrote: ↑June 15th, 2021, 12:00 am
Hello. What is an "MTCC"? I did not find it on LifeWiki.
Ah, it means my typical crummy contribution. You should check out
Saka's forum term dictionary.
Re: Thread for basic questions
Posted: June 15th, 2021, 9:46 pm
by wwei47
This question is a lot harder than it looks. Given a single true object on a 100-by-50 torus (I've already written the object separation algorithms), how would I unwrap it onto an infinite plane? This:
Code: Select all
x = 96, y = 50, rule = B3/S23:T100,50
2o4b2o7bo6bo7b2o7bo6bo7b2o7bo6bo7b2o7bo5bo$o5bo7b2o7bo6bo7b2o7bo6bo7b
2o7bo6bo7b2o4b2o$2bo4bo6bo7b2o7bo6bo7b2o7bo6bo7b2o7bo6bo7b2o$b2o3b2o7b
o6bo7b2o7bo6bo7b2o7bo6bo7b2o7bo6bo$3b2o3b2o4b2o7bo6bo7b2o7bo6bo7b2o7bo
6bo7b2o7bo$3bo4bo5bo7b2o7bo6bo7b2o7bo6bo7b2o7bo6bo7b2o$5bo4bo4bo6bo7b
2o7bo6bo7b2o7bo6bo7b2o7bo6bo$4b2o3b2o3b2o7bo6bo7b2o7bo6bo7b2o7bo6bo7b
2o7bo$6b2o3b2o3b2o4b2o7bo6bo7b2o7bo6bo7b2o7bo6bo7b2o$6bo4bo4bo5bo7b2o
7bo6bo7b2o7bo6bo7b2o7bo6bo$7bo5bo4bo4bo6bo7b2o7bo6bo7b2o7bo6bo7b2o7bo$
6b2o4b2o3b2o3b2o7bo6bo7b2o7bo6bo7b2o7bo6bo7b2o$6bo7b2o3b2o3b2o4b2o7bo
6bo7b2o7bo6bo7b2o7bo6bo$7bo6bo4bo4bo5bo7b2o7bo6bo7b2o7bo6bo7b2o7bo$6b
2o7bo5bo4bo4bo6bo7b2o7bo6bo7b2o7bo6bo7b2o$6bo7b2o4b2o3b2o3b2o7bo6bo7b
2o7bo6bo7b2o7bo6bo$7bo6bo7b2o3b2o3b2o4b2o7bo6bo7b2o7bo6bo7b2o7bo$6b2o
7bo6bo4bo4bo5bo7b2o7bo6bo7b2o7bo6bo7b2o$6bo7b2o7bo5bo4bo4bo6bo7b2o7bo
6bo7b2o7bo6bo$7bo6bo7b2o4b2o3b2o3b2o7bo6bo7b2o7bo6bo7b2o7bo$6b2o7bo6bo
7b2o3b2o3b2o4b2o7bo6bo7b2o7bo6bo7b2o$6bo7b2o7bo6bo4bo4bo5bo7b2o7bo6bo
7b2o7bo6bo$7bo6bo7b2o7bo5bo4bo4bo6bo7b2o7bo6bo7b2o7bo$6b2o7bo6bo7b2o4b
2o3b2o3b2o7bo6bo7b2o7bo6bo7b2o$6bo7b2o7bo6bo7b2o3b2o3b2o4b2o7bo6bo7b2o
7bo6bo$7bo6bo7b2o7bo6bo4bo4bo5bo7b2o7bo6bo7b2o7bo$6b2o7bo6bo7b2o7bo5bo
4bo4bo6bo7b2o7bo6bo7b2o$6bo7b2o7bo6bo7b2o4b2o3b2o3b2o7bo6bo7b2o7bo6bo$
7bo6bo7b2o7bo6bo7b2o3b2o3b2o4b2o7bo6bo7b2o7bo$6b2o7bo6bo7b2o7bo6bo4bo
4bo5bo7b2o7bo6bo7b2o$6bo7b2o7bo6bo7b2o7bo5bo4bo4bo6bo7b2o7bo6bo$7bo6bo
7b2o7bo6bo7b2o4b2o3b2o3b2o7bo6bo7b2o7bo$6b2o7bo6bo7b2o7bo6bo7b2o3b2o3b
2o4b2o7bo6bo7b2o$6bo7b2o7bo6bo7b2o7bo6bo4bo4bo5bo7b2o7bo6bo$7bo6bo7b2o
7bo6bo7b2o7bo5bo4bo4bo6bo7b2o7bo$6b2o7bo6bo7b2o7bo6bo7b2o4b2o3b2o3b2o
7bo6bo7b2o$6bo7b2o7bo6bo7b2o7bo6bo7b2o3b2o3b2o4b2o7bo6bo$7bo6bo7b2o7bo
6bo7b2o7bo6bo4bo4bo5bo7b2o7bo$6b2o7bo6bo7b2o7bo6bo7b2o7bo5bo4bo4bo6bo
7b2o$6bo7b2o7bo6bo7b2o7bo6bo7b2o4b2o3b2o3b2o7bo6bo$7bo6bo7b2o7bo6bo7b
2o7bo6bo7b2o3b2o3b2o4b2o7bo$6b2o7bo6bo7b2o7bo6bo7b2o7bo6bo4bo4bo5bo7b
2o$6bo7b2o7bo6bo7b2o7bo6bo7b2o7bo5bo4bo4bo6bo$7bo6bo7b2o7bo6bo7b2o7bo
6bo7b2o4b2o3b2o3b2o7bo$6b2o7bo6bo7b2o7bo6bo7b2o7bo6bo7b2o3b2o3b2o4b2o$
6bo7b2o7bo6bo7b2o7bo6bo7b2o7bo6bo4bo4bo5bo$7bo6bo7b2o7bo6bo7b2o7bo6bo
7b2o7bo5bo4bo4bo$6b2o7bo6bo7b2o7bo6bo7b2o7bo6bo7b2o4b2o3b2o3b2o$6bo7b
2o7bo6bo7b2o7bo6bo7b2o7bo6bo7b2o3b2o$7bo6bo7b2o7bo6bo7b2o7bo6bo7b2o7bo
6bo4bo!
Would unwrap to this.
Code: Select all
x = 96, y = 648, rule = B3/S23
2o$o$2bo$b2o$3b2o$3bo$5bo$4b2o$6b2o$6bo$7bo$6b2o$6bo$7bo$6b2o$6bo$7bo$
6b2o$6bo$7bo$6b2o$6bo$7bo$6b2o$6bo$7bo$6b2o$6bo$7bo$6b2o$6bo$7bo$6b2o$
6bo$7bo$6b2o$6bo$7bo$6b2o$6bo$7bo$6b2o$6bo$7bo$6b2o$6bo$7bo$6b2o$6bo$
7bo$6b2o$6bo$7bo$6b2o$8b2o$8bo$10bo$9b2o$11b2o$11bo$13bo$12b2o$14b2o$
14bo$15bo$14b2o$14bo$15bo$14b2o$14bo$15bo$14b2o$14bo$15bo$14b2o$14bo$
15bo$14b2o$14bo$15bo$14b2o$14bo$15bo$14b2o$14bo$15bo$14b2o$14bo$15bo$
14b2o$14bo$15bo$14b2o$14bo$15bo$14b2o$14bo$15bo$14b2o$14bo$15bo$14b2o$
14bo$15bo$14b2o$14bo$15bo$14b2o$16b2o$16bo$18bo$17b2o$19b2o$19bo$21bo$
20b2o$22b2o$22bo$23bo$22b2o$22bo$23bo$22b2o$22bo$23bo$22b2o$22bo$23bo$
22b2o$22bo$23bo$22b2o$22bo$23bo$22b2o$22bo$23bo$22b2o$22bo$23bo$22b2o$
22bo$23bo$22b2o$22bo$23bo$22b2o$22bo$23bo$22b2o$22bo$23bo$22b2o$22bo$
23bo$22b2o$22bo$23bo$22b2o$22bo$23bo$22b2o$24b2o$24bo$26bo$25b2o$27b2o
$27bo$29bo$28b2o$30b2o$30bo$31bo$30b2o$30bo$31bo$30b2o$30bo$31bo$30b2o
$30bo$31bo$30b2o$30bo$31bo$30b2o$30bo$31bo$30b2o$30bo$31bo$30b2o$30bo$
31bo$30b2o$30bo$31bo$30b2o$30bo$31bo$30b2o$30bo$31bo$30b2o$30bo$31bo$
30b2o$30bo$31bo$30b2o$30bo$31bo$30b2o$30bo$31bo$30b2o$32b2o$32bo$34bo$
33b2o$35b2o$35bo$37bo$36b2o$38b2o$38bo$39bo$38b2o$38bo$39bo$38b2o$38bo
$39bo$38b2o$38bo$39bo$38b2o$38bo$39bo$38b2o$38bo$39bo$38b2o$38bo$39bo$
38b2o$38bo$39bo$38b2o$38bo$39bo$38b2o$38bo$39bo$38b2o$38bo$39bo$38b2o$
38bo$39bo$38b2o$38bo$39bo$38b2o$38bo$39bo$38b2o$38bo$39bo$38b2o$40b2o$
40bo$42bo$41b2o$43b2o$43bo$45bo$44b2o$46b2o$46bo$47bo$46b2o$46bo$47bo$
46b2o$46bo$47bo$46b2o$46bo$47bo$46b2o$46bo$47bo$46b2o$46bo$47bo$46b2o$
46bo$47bo$46b2o$46bo$47bo$46b2o$46bo$47bo$46b2o$46bo$47bo$46b2o$46bo$
47bo$46b2o$46bo$47bo$46b2o$46bo$47bo$46b2o$46bo$47bo$46b2o$46bo$47bo$
46b2o$48b2o$48bo$50bo$49b2o$51b2o$51bo$53bo$52b2o$54b2o$54bo$55bo$54b
2o$54bo$55bo$54b2o$54bo$55bo$54b2o$54bo$55bo$54b2o$54bo$55bo$54b2o$54b
o$55bo$54b2o$54bo$55bo$54b2o$54bo$55bo$54b2o$54bo$55bo$54b2o$54bo$55bo
$54b2o$54bo$55bo$54b2o$54bo$55bo$54b2o$54bo$55bo$54b2o$54bo$55bo$54b2o
$54bo$55bo$54b2o$56b2o$56bo$58bo$57b2o$59b2o$59bo$61bo$60b2o$62b2o$62b
o$63bo$62b2o$62bo$63bo$62b2o$62bo$63bo$62b2o$62bo$63bo$62b2o$62bo$63bo
$62b2o$62bo$63bo$62b2o$62bo$63bo$62b2o$62bo$63bo$62b2o$62bo$63bo$62b2o
$62bo$63bo$62b2o$62bo$63bo$62b2o$62bo$63bo$62b2o$62bo$63bo$62b2o$62bo$
63bo$62b2o$62bo$63bo$62b2o$64b2o$64bo$66bo$65b2o$67b2o$67bo$69bo$68b2o
$70b2o$70bo$71bo$70b2o$70bo$71bo$70b2o$70bo$71bo$70b2o$70bo$71bo$70b2o
$70bo$71bo$70b2o$70bo$71bo$70b2o$70bo$71bo$70b2o$70bo$71bo$70b2o$70bo$
71bo$70b2o$70bo$71bo$70b2o$70bo$71bo$70b2o$70bo$71bo$70b2o$70bo$71bo$
70b2o$70bo$71bo$70b2o$70bo$71bo$70b2o$72b2o$72bo$74bo$73b2o$75b2o$75bo
$77bo$76b2o$78b2o$78bo$79bo$78b2o$78bo$79bo$78b2o$78bo$79bo$78b2o$78bo
$79bo$78b2o$78bo$79bo$78b2o$78bo$79bo$78b2o$78bo$79bo$78b2o$78bo$79bo$
78b2o$78bo$79bo$78b2o$78bo$79bo$78b2o$78bo$79bo$78b2o$78bo$79bo$78b2o$
78bo$79bo$78b2o$78bo$79bo$78b2o$78bo$79bo$78b2o$80b2o$80bo$82bo$81b2o$
83b2o$83bo$85bo$84b2o$86b2o$86bo$87bo$86b2o$86bo$87bo$86b2o$86bo$87bo$
86b2o$86bo$87bo$86b2o$86bo$87bo$86b2o$86bo$87bo$86b2o$86bo$87bo$86b2o$
86bo$87bo$86b2o$86bo$87bo$86b2o$86bo$87bo$86b2o$86bo$87bo$86b2o$86bo$
87bo$86b2o$86bo$87bo$86b2o$86bo$87bo$86b2o$86bo$87bo$86b2o$88b2o$88bo$
90bo$89b2o$91b2o$91bo$93bo$92b2o$94b2o$94bo$95bo$94b2o$94bo$95bo$94b2o
$94bo$95bo$94b2o$94bo$95bo$94b2o$94bo$95bo$94b2o$94bo$95bo$94b2o$94bo$
95bo$94b2o$94bo$95bo$94b2o$94bo$95bo$94b2o$94bo$95bo$94b2o$94bo$95bo$
94b2o$94bo$95bo$94b2o$94bo$95bo$94b2o$94bo$95bo$94b2o$94bo$95bo$94b2o!
Re: Thread for basic questions
Posted: June 15th, 2021, 10:49 pm
by mniemiec
wwei47 wrote: ↑June 15th, 2021, 9:46 pm
This question is a lot harder than it looks. Given a single true object on a 100-by-50 torus (I've already written the object separation algorithms), how would I unwrap it onto an infinite plane? This: ... Would unwrap to this. ...
This isn't really a well-defined function. E.g. if there's a bridge connecting any of the stripes, this may not be possible, if the connections form cycles that wrap entirely around the torus. One can find a solution that works for the limited number of still-lifes that don't include loops like this.
One way to do it is similar to how one might do object separation - i.e. mark one cell as part of an object, and then look at any unmarked cells adjacent to marked ones as part of the same object.
In the same way, mark one arbitrary living cell as being in the corresponding place in the infinite plane (i.e. by tatting it as having a known offset, based on how often one's viewpoint wraps around the torus, and being ininitially 0 for the first chosen tell). Then, expand from marked cells to unmarked cells, adjusting the offset every time you wrap around the torus boundaries. If you ever want to mark a cell with a different offset than it is already marked as, you have found an object that cannot be mapped on to the plane. Once all cells have been mapped without encountering such a problem, you have successfully unwrapped your still life.