Page 8 of 38

Posted: July 9th, 2016, 4:08 am
One problem with the Pi wave stabilization problem is it's hard to stabilize it from the back. If there was a way to do that with a stream of spaceships then it could be possible to make a wave ship that synthesizes its own supports. But pi waves have a really low period for this sort of thing, so I'm not entirely sure such a reaction exists; at least, one that can be stuck onto the end of a normal pi wave.

Posted: July 17th, 2016, 7:40 am
The concept of a decently sized elementary 3c/10 ship still seems like something we could get our hands on before 2020.

If you look close enough at the front of the 3c/7 spaghetti monster, you can kind of make out the 3c/7 wave it can support (except not p28).

Either way, it seems a lot more likely than that c/26 orthogonal spaceship I proposed.

Posted: July 18th, 2016, 4:27 pm
Sphenocorona wrote:One problem with the Pi wave stabilization problem is it's hard to stabilize it from the back. If there was a way to do that with a stream of spaceships then it could be possible to make a wave ship that synthesizes its own supports.
You can stabilize Pi waves with sideways gliders, so theoretically one could build a catepillar based on Pi waves, but so far no way was found to turn a Pi wave into a rake. I tried it, but found almost nothing interesting.

Posted: July 18th, 2016, 4:36 pm
Asking obvious questions like I normally do, is c/8 the fastest possible speed for a slope 3 spaceship, and c/10 the fastest for a slope 4? Since the fastest for a slope 0 (orthogonal) is c/2, a slope 1 (diagonal) c/4 and a slope 2 c/6, it should probably follow that pattern.

Also, has anyone tried putting almost knightship through a further search yet, to see if anything useful appears at the back end?

Posted: July 18th, 2016, 7:19 pm
muzik wrote:Asking obvious questions like I normally do, is c/8 the fastest possible speed for a slope 3 spaceship, and c/10 the fastest for a slope 4? Since the fastest for a slope 0 (orthogonal) is c/2, a slope 1 (diagonal) c/4 and a slope 2 c/6, it should probably follow that pattern.

Also, has anyone tried putting almost knightship through a further search yet, to see if anything useful appears at the back end?
I couldn't find the Two Forbidden Directions thread when I tried to look for it, but that line of reasoning allows one to prove that in Life the shortest possible time in which a pattern can move (x,y) [with x >= y] is 2x+y generations.
Such a proof can be obtained by "forbidding" the two directions closest to the pattern's direction of travel.

Posted: July 18th, 2016, 8:37 pm
I personally think that specifying speed as (x,y)c/p rather than dc/p for ships that move in oblique directions is better practice, for the reason that the latter gets confusing, ambiguous, and misleading if you aren't VERY careful. You're either going to imply the ship is faster than it is, or you're going to imply that it's slower than it really is.
BlinkerSpawn wrote:I couldn't find the Two Forbidden Directions thread when I tried to look for it, but that line of reasoning allows one to prove that in Life the shortest possible time in which a pattern can move (x,y) [with x >= y] is 2x+y generations.
I'm pretty sure that it's been actually been proven to be 2(x+y) generations, though. Otherwise slope 1 could move at 2*1 + 1 = 3 -> c/3 diagonal. A spaceship can never out run a glider or a LWSS, and using those as a baseline gives 2(x+y) for the minimum possible period for a spaceship that moves by (x,y). This isn't rigorous but I'm fairly sure I've seen a rigorous proof of this.
Though 2x+y might be the fastest for any arbitrary B3 rule; it seems more possible in that case.

Posted: July 19th, 2016, 12:04 am
muzik wrote:Also, has anyone tried putting almost knightship through a further search yet, to see if anything useful appears at the back end?
In keeping with your theme, the obvious answer to this is "Yes". But nothing came of it, otherwise they would have posted it. I have myself run several searches with JLS trying to extend a (2, 1)c/7 partial in a rather ad hoc way, but haven't bothered to post any of the longer partials as it doesn't seem particularly promising or feel like it would be helpful to others. Unlike with gfind and zfind it is difficult to give concise negative results from WLS or JLS when incrementally expanding the search area in an attempt to continue the search further. I have on numerous occasions had wishful thoughts about a search utility which could automatically widen the search area of unsuccessful searches in a manner that is more fine-grained than increasing the total width of an orthogonal search by 1 and restarting the search (easily done with a driver script for gfind or zfind). But that is fodder for a different thread.

Others have certainly done the same as me with knightship searches and I'm sure you will even find a few posted partials if you dig deep enough in the archives, but the majority of negative results will have been consigned to the bit bucket or are languishing in personal archives of past search runs and results.

Posted: July 20th, 2016, 4:26 pm
Bringing back up the embarrasing and kind of impossible concept of a c/26 orthogonal spaceship, how likely is it that we could somehow make an engineered spaceship using that reaction as a front end?

This one:

Code: Select all

``````x = 13, y = 8, rule = B3/S23
ob2o\$3o\$bo3\$10bobo\$10b2o\$11bo!``````
Of course, we would need to find some other crawler-type reaction that moves at c/26, which we don't seem to have, so a caterloopillar-like approach could be used.

Also, I've had an idea for a search program that could potentially catch a few of the lower hanging fruits: it would kind of randomise what it searches for (the period being a random number from 6 to 12, the symmetries being randomly chosen as well). If anyone wants to devise a program that runs along those lines, the elusive c/8 orthogonal could be hanging on a branch painted with camouflage colours.

Posted: July 29th, 2016, 7:02 am
Has anyone considered programming that, or is there such a search program that exists?

anyway, decided to post a few c/8 diagonal partials from another thread:

Code: Select all

``````x = 24, y = 26, rule = B3/S23
bo\$bob2obo5b2ob2o\$obo3bo4b2o4bo3b3o\$b2o3bo2b2obobo3bob2obo\$2bo7bobo\$b
2ob2obobobob2o\$bo2b3o3bo2bo2bo\$bo4b3o4bo2bo\$bo3bo2b2o4bo\$2bo10bo\$3bo6b
4ob2o\$2obo2b2o2b2o3bo\$3b2ob2obo2b2o\$obo2bob3o2bo\$o3bo7b2o\$3bo4bobo\$2b
2o4bobobo\$2bob3o\$3b2o3bo\$3bo\$bob2o\$2bobo\$2bob2o\$3bobo\$o2bo\$o!``````

Code: Select all

``````x = 15, y = 21, rule = B3/S23
2b2o\$o3bo\$o5b2o\$bo3b3o\$2b2o4bo4bo\$2b2o8b2o\$2b2o5bo2b2o\$2bo2bo3bo2b3o\$
3bo3bo4bobo\$5b3o2b2o\$5b2o4bobo\$6b2o\$5b4o\$4b2o4bo\$4bo3bobo\$5b2o\$6b2obo\$
8b3o\$10bo\$10bo\$10bo!``````
Not entirely sure about them. They seem to be tangled up in some weird indescribable mess. Might be worth running those through a search with zfind or such to see if it makes any differences.

Looking though the forum, I haven't seen any partials for slope 3 knightships...

Posted: July 31st, 2016, 2:38 pm
If we already have successfully found 2c/8 and 4c/8 spaceships through searches, why do we not have c/8 and 3c/8 yet?

Similarly, we have the 3c/9 117P9H3V0, which I'm assuming was found with at least some help of a search program. So why no c/9, 2c/9 or 4c/9?

Posted: July 31st, 2016, 2:45 pm
muzik wrote:If we already have successfully found 2c/8 and 4c/8 spaceships through searches, why do we not have c/8 and 3c/8 yet?

Similarly, we have the 3c/9 117P9H3V0, which I'm assuming was found with at least some help of a search program. So why no c/9, 2c/9 or 4c/9?
Probably because a 3c/9 or 4c/8 ship usually still has a period 2, 3, or 4 front end, which aren't as hard to find.

Posted: July 31st, 2016, 11:02 pm
muzik wrote:Also, I've had an idea for a search program that could potentially catch a few of the lower hanging fruits: it would kind of randomise what it searches for (the period being a random number from 6 to 12, the symmetries being randomly chosen as well). If anyone wants to devise a program that runs along those lines, the elusive c/8 orthogonal could be hanging on a branch painted with camouflage colours.
muzik wrote:Has anyone considered programming that, or is there such a search program that exists?
What you describe is more of a way of running search programs than the design of a new search program. You would need to provide a lot more detail about how the search should operate for it to qualify as a design. Have you tried gsearch? It tries random patterns with all symmetries and can detect ships of any period (up to some large maximum). It is very unlikely to uncover something at a new speed in Life though.
muzik wrote:anyway, decided to post a few c/8 diagonal partials from another thread
The first one is a lost cause, as dvgrn explained at the time, but the second one is slightly more hopeful. Even so, the frontend only survives for one and a bit periods before exploding (not unusual for slow ship partials). I don't think anyone will pick up the torch on that one anytime soon.
muzik wrote:Looking though the forum, I haven't seen any partials for slope 3 knightships...
Here's my best attempt at a (3,1)c/8 knightship (from a width 11 search in gfind):

Code: Select all

``````x = 9, y = 13, rule = B3/S23
7b2o\$4b2ob2o\$2o4b2o\$3o2bo\$2b6o\$2bo\$b4ob2o\$bo5bo\$5bo\$3b2o3bo\$4bo2b2o\$4b
3o\$5bo!``````

Posted: August 1st, 2016, 4:53 am
wildmyron wrote:
muzik wrote:Looking though the forum, I haven't seen any partials for slope 3 knightships...
Here's my best attempt at a (3,1)c/8 knightship (from a width 11 search in gfind):

Code: Select all

``````x = 9, y = 13, rule = B3/S23
7b2o\$4b2ob2o\$2o4b2o\$3o2bo\$2b6o\$2bo\$b4ob2o\$bo5bo\$5bo\$3b2o3bo\$4bo2b2o\$4b
3o\$5bo!``````
Natural longhook on block!

On a scale of 1-10 for promisingness I'd give this a 6.
gmc_nxtman wrote:
muzik wrote:If we already have successfully found 2c/8 and 4c/8 spaceships through searches, why do we not have c/8 and 3c/8 yet?

Similarly, we have the 3c/9 117P9H3V0, which I'm assuming was found with at least some help of a search program. So why no c/9, 2c/9 or 4c/9?
Probably because a 3c/9 or 4c/8 ship usually still has a period 2, 3, or 4 front end, which aren't as hard to find.
So only the back end has a different period?

Still, there are a few front ends, like this c/8:

Code: Select all

``````x = 34, y = 22, rule = B3/S23
8\$10bo8bo\$9bobo6bobo\$8bo2bo6bo2bo\$9b2o8b2o\$14b2o\$12b2o2b2o\$12bo4bo\$12b
o4bo\$11b8o\$10b4o2b4o\$9bo2bo4bo2bo\$8bo3bo4bo3bo\$9bo2bo4bo2bo!
``````
And since the back ends of the other ships had the higher period, extending this should be no problem.

Posted: August 2nd, 2016, 3:48 am
muzik wrote:Still, there are a few front ends, like this c/8:
Here is plausible continuation:

Code: Select all

``````x = 25, y = 14, rule = B3/S23
2bo8bo8b2o\$bobo6bobo2bo3b2obobo\$o2bo5bo5bo2b2ob2o\$b2o5b2o5bo3b2o\$5b8o
7bo\$5bo2b2o10b2ob2o\$4bo3bo13bobo\$4bo3bo13bobo\$5bo2b2o10b2ob2o\$5b8o7bo\$
b2o5b2o5bo3b2o\$o2bo5bo5bo2b2ob2o\$bobo6bobo2bo3b2obobo\$2bo8bo8b2o!
``````

Posted: August 2nd, 2016, 4:56 am
simsim314 wrote:
muzik wrote:Still, there are a few front ends, like this c/8:
Here is plausible continuation:

Code: Select all

``````x = 25, y = 14, rule = B3/S23
2bo8bo8b2o\$bobo6bobo2bo3b2obobo\$o2bo5bo5bo2b2ob2o\$b2o5b2o5bo3b2o\$5b8o
7bo\$5bo2b2o10b2ob2o\$4bo3bo13bobo\$4bo3bo13bobo\$5bo2b2o10b2ob2o\$5b8o7bo\$
b2o5b2o5bo3b2o\$o2bo5bo5bo2b2ob2o\$bobo6bobo2bo3b2obobo\$2bo8bo8b2o!
``````
The partial I actually cut the front end off of (props to josh):

Code: Select all

``````x = 14, y = 35, rule = B3/S23
2bo8bo\$bobo6bobo\$o2bo6bo2bo\$b2o8b2o\$6b2o\$4b2o2b2o\$4bo4bo\$4bo4bo\$3b8o\$
2b4o2b4o\$bo2bo4bo2bo\$o3bo4bo3bo\$bo2bo4bo2bo3\$b3o6b3o3\$2bo8bo\$b3o6b3o\$
2ob3o2b3ob2o\$obo2bo2bo2bobo\$b2o3b2o3b2o\$5bo2bo\$bo3b4o3bo2\$bo2bo4bo2bo\$
2b3o4b3o\$2bo8bo\$b2o2b4o2b2o\$14o\$3o8b3o\$6b2o\$2o3b4o3b2o\$o12bo!
``````

Posted: August 5th, 2016, 9:07 pm
c/8 partials,not promising

Code: Select all

``````x = 60, y = 61, rule = B3/S23
3bo\$b2obo4b2o\$b2o5bobo\$2obobobo\$3o3bob2o\$b2obob4o\$b2o2bobo\$7b2o\$5bobo\$
6bo\$58b2o\$57b3o\$50b2o3b3o\$50b2o2bo\$50b2o3bob2o\$51b3obo\$6b2o2bo3bo29b3o
b2ob3o\$6b2o2bobobo34b2o2b2o3bo\$6b2o3bo41bo2b2o\$7b3ob2o42b3obo\$3ob2ob3o
bo2bo38b2o\$5b2o2b2ob2o42b4o\$9bo3b2o38b2o\$13bo41b3obo\$9b2ob2o39bo2b2o\$
49b2o2b2o3bo\$44b3ob2ob3o\$51b3obo\$50b2o3bob2o\$50b2o2bo\$50b2o3b3o\$57b3o\$
6b2o2b2o46b2o\$4b4o\$4bo3bob2o\$5bo4b2o\$3ob2obo2b2o2bo\$2bo7bo\$2b2ob2o2b2o
\$3bobob2obo13\$6bo3bo3b2o\$6bo3b2obo\$5b2o3b2ob2obo\$5bobobo4bo\$3ob2ob2ob
2ob4o\$5bo2bob3ob2o\$6b4o\$12bo3bo\$15b2o!
``````

Posted: August 5th, 2016, 9:47 pm
GUYTU6J wrote:c/8 partials,not promising

Code: Select all

``````x = 60, y = 61, rule = B3/S23
3bo\$b2obo4b2o\$b2o5bobo\$2obobobo\$3o3bob2o\$b2obob4o\$b2o2bobo\$7b2o\$5bobo\$
6bo\$58b2o\$57b3o\$50b2o3b3o\$50b2o2bo\$50b2o3bob2o\$51b3obo\$6b2o2bo3bo29b3o
b2ob3o\$6b2o2bobobo34b2o2b2o3bo\$6b2o3bo41bo2b2o\$7b3ob2o42b3obo\$3ob2ob3o
bo2bo38b2o\$5b2o2b2ob2o42b4o\$9bo3b2o38b2o\$13bo41b3obo\$9b2ob2o39bo2b2o\$
49b2o2b2o3bo\$44b3ob2ob3o\$51b3obo\$50b2o3bob2o\$50b2o2bo\$50b2o3b3o\$57b3o\$
6b2o2b2o46b2o\$4b4o\$4bo3bob2o\$5bo4b2o\$3ob2obo2b2o2bo\$2bo7bo\$2b2ob2o2b2o
\$3bobob2obo13\$6bo3bo3b2o\$6bo3b2obo\$5b2o3b2ob2obo\$5bobobo4bo\$3ob2ob2ob
2ob4o\$5bo2bob3ob2o\$6b4o\$12bo3bo\$15b2o!
``````
That may be so but I, for one, can't say I've seen any partials quite like these before.
How did you find them?

Posted: August 5th, 2016, 10:23 pm
GUYTU6J wrote:c/8 partials,not promising

Code: Select all

``````rle
``````
That may be so but I, for one, can't say I've seen any partials quite like these before.
How did you find them?
By using wls for several months.I don't really understand what the options mean.
I said "not promising"because the front ends can't go very long.

Posted: August 5th, 2016, 11:04 pm
At the very least the front ends are there.
The bottommost one looks the best at first glance.
Now that these have been found, maybe you could try extending these partials as an exercise for other search programs. I know gfind and zfind work, but there's compiling stuff that has to go on first that I haven't even really bothered to go through.

Posted: August 5th, 2016, 11:40 pm
BlinkerSpawn wrote:At the very least the front ends are there.
The bottommost one looks the best at first glance.
Now that these have been found, maybe you could try extending these partials as an exercise for other search programs. I know gfind and zfind work, but there's compiling stuff that has to go on first that I haven't even really bothered to go through.
When I enter "zfind b3s23 p8 k1" ,it immediately output:

Code: Select all

``````Building lookup tables.
Lookup tables built.
31
o.....
..o...
...o..
....o.
o...o.
.o....
.oo...
.o....
.oo...
......
.oooo.
.....o
..ooo.
oooo..
......
o.....
...ooo
oo..oo
.oooo.
....o.
160
28
33554432
2636
o.....
.o....
.o....
o.....
......
......
..o...
oo.o..
.oo...
.oo...
......
.ooo..
.o..o.
.o.oo.
oo....
.o.o..
.ooo..
.oo.o.
......
...o..
ooooo.
168
56
67108864
5269
...o..
...oo.
..o...
....oo
..o...
.o..oo
.oo.oo
..o...
......
.ooo..
..o...
..o...
.oo...
o.o...
..oo..
o.o...
......
o...o.
......
.ooo..
.oo...
......
..o...
.ooo..
...o..
....o.
o.o.o.
o...o.
.o..o.
.ooooo
.o....
247
Search complete: no spaceships found.
175377536
13694
``````
By the way ,how to put my partials into gfind?

Posted: August 6th, 2016, 3:59 am
GUYTU6J wrote: By the way ,how to put my partials into gfind?
Some information can be found here ,but I haven't succeeded on it.

Posted: August 6th, 2016, 8:47 am
The wiki mentions an 8-engine Cordership, but there isn't currently a page on one.

Does anyone have any examples?

Posted: August 6th, 2016, 11:20 am
muzik wrote:The wiki mentions an 8-engine Cordership, but there isn't currently a page on one.

Does anyone have any examples?

Code: Select all

``````x = 127, y = 127, rule = B3/S23
106bo\$92b2o7b2o2bobo\$92b3o6b4o3bo\$92b2obobo7bo\$95b2o10bob2o\$107bobo\$
108bo8b2o\$117b2o6\$117b2o\$115b3ob2o4b2o\$115bobob2o4b2o\$113b2obo\$106b2o
3bo\$97b2o5bo3bobo\$97b2o6b2obob3o\$105bo3b4o\$103bo\$102bobo\$101bo\$85bobo
6bobo4bo\$84bo9bobo4bobo\$85bo2bo6bo6bo\$87b3o5\$77bo\$76bobo2\$76bo2bo\$78b
2o\$79bo13\$59bobo\$58bo\$59bo2bo\$61b3o5\$51bo\$50bobo2\$50bo2bo\$52b2o\$53bo
13\$33bobo\$32bo\$33bo2bo\$35b3o5\$25bo\$24bobo2\$24bo2bo\$26b2o\$27bo3\$b3o\$b3o
\$2bo21b2o\$3b2o21bo\$4bo19b2o\$3bo14b2o\$18b2o3\$b2o20b3o\$b2o19bo3bo\$2bo18b
o3bo\$2bo15bo3bo\$bobo15b2o\$o16bobo\$bo2b2o11bo\$2bo3bo11b2o\$4b2o14bo\$4bo
13b3o\$17bob2o\$19b2o\$16bo\$16bo\$14b2o\$14bobo\$6b2o5b3o\$6b2o5bo\$14b2o\$14b
2o5\$14b2o\$14b2o!
``````
I found a stabilization for your switch engine wave here.

Posted: August 6th, 2016, 11:25 am
muzik wrote:The wiki mentions an 8-engine Cordership, but there isn't currently a page on one.

Does anyone have any examples?
There probably isn't a page because
• It's not the smallest, lowest population, or most interesting
• There are likely dozens (if not hundreds) of ways to combine 8 Corderengines into a spaceship of some sort. Listing them all would be tedious (although useful for building larger Corderconglomerates) and trying to only choose a few would limit the page's usefulness.

Posted: August 7th, 2016, 3:13 am
This search came out empty for p8 (height 16):

Code: Select all

``````#P 0 0
...............
...*........*..
..*.*......*.*.
.*..*.....*....
..**.....**....
......********.
......*..**....
.....*...*.....
.....*...*.....
......*..**....
......********.
..**.....**....
.*..*.....*....
..*.*......*.*.
...*........*..
...............
``````