DB 13 Aug 1992 (using OWSS; MWSS support found by DB on 16 Aug 1992)
where "DB" is David Bell.
I'm not sure if the OWSS version of the Coe Ship has been posted before, but it's essentially a replacement of a MWSS on LWSS reaction with a MWSS on OWSS reaction:
HartmutHolzwart wrote: August 20th, 2020, 3:26 pm
If I'd know an appropriate short wide c/6 spaceship with a domino spark, then this couls be a c/6 grey ship component:
You‘re right, pasting error... I just wanted to get that thing out quickly, as I won‘t be able to spend much time on spaceship searching for the next two weeks. Maybe you could spend some time on c/6 grey ship components? You seem to succeed where I fail.
Also 2c/5 and 2c/6 against the grain are still vexing me. Not the slightest progress. No reason to complain, though! Quite a lot of new results and closed gaps in grey ships and agars recently.
HartmutHolzwart wrote: August 21st, 2020, 4:38 am
Also 2c/5 and 2c/6 against the grain are still vexing me. Not the slightest progress. No reason to complain, though! Quite a lot of new results and closed gaps in grey ships and agars recently.
I haven't looked at 2c/6 perpendicular yet and have looked at 2c/5 perpendicular only a little. Fundamentally my current techniques are limited by a need for a pre-specified 2 row slice to "start" at. For other searches I've generally cheated and grabbed a common slice from existing work.
My best guess for a two row slice can be seen here (made up of top two rows in each generation):
This is the longest partial result for that search, although I may yet push it further (increase allowed widths).
I also had an idea about automated guessing of a good 2 row slice but it hasn't really worked out yet. The idea was to try searching from all very "thin" slices (where "thin" means the rows are mostly made up of agar on the left and dead cells on the right) and then analyze the collected results (especially in terms of connectivity between the thin rows, especially if there are any cycles since then we'd have a repeating edge component). This is sort of where I came up with the above slice from (I think it is also similar to an analogous 2c/4 slice from existing 2c/4 perpendicular work).
I guess I'm curious how you do your searches and/or if you have any ideas about how to guess up these starting slices.
I noticed that Adam P. Goucher uploaded a large (2,1)c/7 search from ikpx2 between August 27 and 31, although he didn't find anything. The partials are filed at ikpx2_k7_stdin_test.
Are there any c/2 spaceships whose forward propulsion mechanism has an even number of cells at the leading edge instead of an odd number of cells? I have mentioned a partial that I found here.
MathAndCode wrote: September 12th, 2020, 8:14 pm
Are there any c/2 spaceships whose forward propulsion mechanism has an even number of cells at the leading edge instead of an odd number of cells? I have mentioned a partial that I found here.
There's a way of supporting a similar partial from behind, but it still requires odd leading edge components so I'm not sure if it is within the spirit of what you're looking for.
x = 60, y = 30
$8bo31boobo$7boo12boo15b4oboo$6bob3o7b5obo12boo4boo$5boo3b3o4bo4bobboboo5boo4bo
bbo$6b4oboobb3o5b3o9b3o3boboo$7bobboobobbo4bobbo3bobb3obbobboobboo$9boobobbo5bo
3bobbo4bobo5bo7bo$4boobbo9bo5bo6bo7b3o6bobo$3b3oboobbo6boobobb6obo6bobbo4b3o5b4o
$bboobb3o8bobo9bo7boobo3boobobb4o4bo$4bo3bo3bo3boobbo5bo11boo4boo6bo$bbobbo8boob
o3b6o12boboo6bo3bo4bo$bbob4o4bobobbob7o15boobo3bo6boo$bo5boboo10b3o18boo$bo5bob
oo10b3o18boo$bbob4o4bobobbob7o15boobo3bo6boo$bbobbo8boobo3b6o12boboo6bo3bo4bo$4b
o3bo3bo3boobbo5bo11boo4boo6bo$bboobb3o8bobo9bo7boobo3boobobb4o4bo$3b3oboobbo6boo
bobb6obo6bobbo4b3o5b4o$4boobbo9bo5bo6bo7b3o6bobo$9boobobbo5bo3bobbo4bobo5bo7bo$
7bobboobobbo4bobbo3bobb3obbobboobboo$6b4oboobb3o5b3o9b3o3boboo$5boo3b3o4bo4bobbo
boo5boo4bobbo$6bob3o7b5obo12boo4boo$7boo12boo15b4oboo$8bo31boobo!
There's almost certainly something smaller, first I found
The 5S project (Smallest Spaceships Supporting Specific Speeds) is now maintained by AforAmpere. The latest collection is hosted on GitHub and contains well over 1,000,000 spaceships.
Semi-active here - recovering from a severe case of LWTDS.
MathAndCode wrote: September 12th, 2020, 8:14 pm
Are there any c/2 spaceships whose forward propulsion mechanism has an even number of cells at the leading edge instead of an odd number of cells? I have mentioned a partial that I found here.
There's a way of supporting a similar partial from behind, but it still requires odd leading edge components so I'm not sure if it is within the spirit of what you're looking for.
x = 32, y = 8, rule = B3/S23
6bo18bo$5b3o16b3o$3b2ob3o14b3ob2o$4bo2bob2o4b2o4b2obo2bo$b2obo4bobob2o2b2obobo4bob2o$b2obobo2bobo8bobo2bobob2o$bo8b3obo2bob3o8bo$2o7b2o10b2o7b2o!
However, yours is slightly closer because the even leading-edge component is in front of the odd leading-edge component, so thank you.
JP21 wrote: September 13th, 2020, 9:15 amThat's the first I found, but it uses too much connectors. (like gluing 2 spaceships into one)
I want a more direct example.
Here's a quick way of getting rid of a block that can be done from only one side.
I had an interesting idea for a true-period geminoid-like spaceship.
Each side consists of a main glider duplicator, a merge circuit, and some reflectors, as well as a faraway pair of Snarks, or a boojum reflector, or a rectifier, or whatever 180-degree reflector you want, for that matter. (Hooray, a 180-degree reflector being a key component to circuitry!) One side is 1hd forwards in the direction of travel with respect to the other side.
The operation works as follows: The front side receives the instruction tape, and it copies it. One copy gets sent to the faraway reflector through the merge circuit, and the other is sent at the back side, shooting it down and rebuilding it (including the reflector), with an offset of (1,1). This causes the positions to swap, the front is now the back, and vice versa.
Now, the other copy of the instruction tape finally arrives from the reflector, and is sent to the back (now the front) side . Before, this shot down the back side. However, the (1,1) move causes it to line up with the now-front side perfectly, allowing it to receive the instruction tape. Everything now repeats, reflected, and after that, the cycle repeats, with the entire spaceship having moved one cell diagonally in the cycle.
I think it's worth building one of these, since it would be possible to show that there were infinitely many spaceships that moved at a true-period c/n. This also operates differently than a usual geminoid, since usually each side constructs its next copy and removes its previous copy, as opposed to being removed and rebuilt by the other side. This allows for lower displacement values because the bounding box of each side is not a worry anymore, there's never more than one copy of each side. This would also allow for different displacements than the (1,1) mentioned, and might make a true geminoid knightship possible, since as far as I know, it never actually got completed.
(I have no idea how to make an explicit example of a self-constructing spaceship though lol)
wwei23 wrote: September 19th, 2020, 6:20 pm
I had an interesting idea for a true-period geminoid-like spaceship...
(I have no idea how to make an explicit example of a self-constructing spaceship though lol)
It's a little difficult to follow the description without a diagram. How is this idea different from chris_c's camelship? Ships with a design similar to the camelship can be constructed for any small step size you might want.
dvgrn wrote: September 19th, 2020, 10:06 pm
It's a little difficult to follow the description without a diagram. How is this idea different from chris_c's camelship?
I'm using glide-reflectivity so that I can use only one recipe, and sending one copy of the recipe to faraway reflectors to bounce back to the reconstructed other side. Here's a rough diagram I made.
5938D0DF-EE50-465E-84B8-93A3D6BD7BDC.png (22.42 KiB) Viewed 5484 times
EDIT: I would also like to mention that I was unaware that there was that camelship in the first place.
wwei23 wrote: September 19th, 2020, 10:44 pmEDIT: I would also like to mention that I was unaware that there was that camelship in the first place.
There's also a half-finished giraffeship design that avoids the need for the tricky recipe-crossing glider used in the camelship.
If I remember right, that recipe-crossing camelship glider does the switching between "ignore the first half-recipe" and "rebuild things with the second half-recipe", for the back half of the giraffeship. At the front end, the first half-recipe ends with an "ignore the second half-recipe" instruction, building an eater on its own single-channel lane, so that part is easy -- it's only unsuppressing a suppressed single-channel stream that's a little bit difficult.
Anyway, there's currently not really a shortage of workable self-constructing spaceship designs -- and that link is an old list, in serious need of an update. The big shortage seems to be of people who are willing to learn the construction tools and techniques to implement any of these designs.
dvgrn wrote: September 20th, 2020, 8:08 am The big shortage seems to be of people who are willing to learn the construction tools and techniques to implement any of these designs.
Well, I just ask slmake to build the slow salvo recipe and put its output between the sides of the spaceship, right?
EDIT: Or is there something I'm missing here?
wwei23 wrote: September 20th, 2020, 8:45 amWell, I just ask slmake to build the slow salvo recipe and put its output between the sides of the spaceship, right?
EDIT: Or is there something I'm missing here?
As a high-level overview, that's pretty much correct. But there's a good bit of work to do before you can "just ask slmake" anything.
The first step is probably to build a working piece of circuitry that can successfully accept just one glider, standing in for a full recipe, and test it out by moving your 180-degree reflectors (1,1) by hand.
The "shooting down" part of the design isn't something that slsparse handles automatically, so you'll have to come up with some kind of "meteor shower" of gliders, probably leaving behind one target block to rebuild from. Then you figure out whether you can do the shooting-down and reconstruction with 0-degree gliders, or if you'll use a 90-degree construction elbow somehow, and what that means for the design of the main part of the circuitry.
A pair of Snarks might be a good choice for the 180-degree reflector, or else a syringe plus some spacer Herschel conduits plus an H-to-G. That lets you separate the output glider stream from the input stream by a reasonable distance. You can't really use rectifiers or boojum reflectors in a design compiled by slsparse, without completely rebuilding the single-channel database and throwing away most of the recipes. Single-channel recipes mostly include gliders with separation of ninety-some ticks, and neither rectifiers nor boojum reflectors can support that.
Then... it's not quite clear to me how the two halves of your main circuitry will stay out of each other's way while they're rebuilding each other. But don't worry about explaining it in words. Words at this stage have a way of hiding serious design flaws. Just build a working sample of the main circuitry and the reflectors, and then it will be obvious how it works by inspection.
dvgrn wrote: September 20th, 2020, 11:21 am
A pair of Snarks might be a good choice for the 180-degree reflector, or else a syringe plus some spacer Herschel conduits plus an H-to-G. That lets you separate the output glider stream from the input stream by a reasonable distance.
It's separated by 1hd because glide-reflectivity.
Edit: How do I sryinge when slow-salvo-ing?
Edit 2: This means I need two merge circuits, because I'll have to bounce both copies of the recipe, the entire thing needs to be received before it's retransmitted.
Edit 3: This is going to be harder than I thought.
EDIT 4: Before the circuitry has been moved (right). the beehive and block when hit, become a construction arm. After the circuitry has moved (left), it is finally ready to receive the instruction tape, and sends it back twice. The reason why it's sent back twice is because one copy moves the other side, and one copy is received by the other side.
wwei23 wrote: September 20th, 2020, 11:54 am
Edit: How do I sryinge when slow-salvo-ing?
Not sure what that question means. Syringes are an integral part of self-constructing circuitry, but you have to use the eater2-only version followed by an Lx200, because that's more or less the only glider-to-Herschel conversion that works with a repeat time of 90 ticks, that slsparse knows how to build.
Your blueprint above has highway robbers in it, but I"m not sure what they would be good for in your design, because they can't do anything with a single-channel recipe with gliders 90-some ticks apart.
