"Glider hasslers"

[codeholic]

Does the following pattern have a name? What is the name of the class of such patterns? Is there a summary of these patterns of different periods somewhere?

Code: Select all

x = 20, y = 33, rule = B3/S23
13bo$12bo$12b3o13$9b2o$8bobo$7bo$7bo2bo7b2o$7bo10b2o$8bobo$9b2o2$bo$bo
$obo$bo$bo$bo$bo$obo$bo$bo!

I'm interested in both of those that change the lane of a glider and those that advance or delay it for certain number of generations, and especially in those patterns that perform a reaction on their edge.

[Extrementhusiast]

I had made this previously:

Code: Select all

x = 107, y = 73, rule = B3/S23
48bo$47bobo$47bobo24bobo2b2o$46b2ob2o8b2o12bo3bob3o$46bo3bo8b2o13bo6b
2o$44b3o3b3o22b5obo2bo$43bo2bo3bo2bo24b3o3b3o$47bobo9bobo3bobo19bo$19b
o23b2o7b2o7bo3bo12b3o3b3o$17b3o23b2o7b2o3bo11bo5b5obo2bo$16bo29bo3bo5b
o3bo5bo3bo3bo6b2o$16b2o28b2ob2o6bo11bo3bo3bob3o$30bo13b3o3b3o8bo3bo8bo
bo2b2o$28b3o13b2o5b2o6bobo3bobo$27bo11bo4bo7bo$27b2o8b3o$36bo$36b2o$
14bo$14b2o$13b4o$10b3o3b2o$10bo5bo$10bo$2b2o7bo4bo51bobo$bobo9bo2b2o
28b2ob2o18b2o$bo12bob2o29bobo19bo$2o13b2o30bobo$48bo3$17b2o$17b2o3$6b
2o$5bobo$5bo$4b2o$10b2o81bo3b3o$11bo80bobo2b5o$8b3o81bo3b2o$8bo84bo2bo
3b2obo$94b3o3b2ob3o$67b3o36bo$69bo24b3o3b2ob3o$68bo24bo2bo3b2obo$78b2o
12bo3b2o$70b2o6b2o12bobo2b5o$71b2o20bo3b3o$70bo3$70bo$69b2o$68b2o$69b
2o2b2o9bo$84b3o$87bo$84b3o$69b2o2b2o9bo$55bo12b2o$55b2o12b2o$56b2o12bo
$41bo9b2o2b2o$39b3o$38bo$39b3o$41bo9b2o2b2o$56b2o$21bo33b2o$21b2o32bo$
20bobo!

The input glider only reaches the bounding box again after about 4000 generations.

[codeholic]

A p46 glider hassler:

Code: Select all

x = 43, y = 25, rule = B3/S23
10b2o$2o7bobo15b2o$2o7bo17b2o$9b3o4$9b3o$9bo$9bobo$10b2o11$39b2o$18b2o
8b2o9bo$17bobo7bobo10b3o$19bo9bo12bo!

Though it is not an edge reaction, the distance between the twin bees shuttle and the eater is enough to fit 10hd construction lanes.

EDIT: In some other phases, it can eat gliders.

Code: Select all

x = 453, y = 29, rule = B3/S23
10b2o108b2o98b2o98b2o98b2o$2o7bobo15b2o81b2o7bobo15b2o71b2o7bobo15b2o
71b2o7bobo15b2o71b2o7bobo15b2o$2o7bo17b2o81b2o7bo17b2o71b2o7bo17b2o71b
2o7bo17b2o71b2o7bo17b2o$9b3o107b3o97b3o97b3o97b3o4$9b3o107b3o97b3o97b
3o97b3o$9bo109bo99bo99bo99bo$9bobo107bobo97bobo97bobo97bobo$10b2o108b
2o98b2o98b2o98b2o11$39b2o108b2o98b2o78b2o8b2o8b2o78b3o7b3o7b2o$39bo
109bo99bo80b2o8b2o7bo81bo9bo7bo$40b3o107b3o97b3o76bo9bo10b3o77bo9bo9b
3o$42bo83bo9bo15bo99bo99bo99bo$126b2o8b2o$14bo9bo100bobo7bobo86b3o7b3o
$14b2o8b2o200bo9bo$13bobo7bobo199bo9bo!

[dvgrn]

In some other phases, it can eat gliders.

That could certainly be very useful! At either 9hd or 10hd -- but especially at 10hd, I think -- there are really significant improvements in efficiency available if singleton gliders can be used. Not to mention the extra elbow-op recipes you might get when you mix in the new 6hd shifted glider.

Just to extend the functionality of the p46 hassler a little more... I didn't do a thorough search, but I found a way to get a singleton glider on the other lane. It takes two extra glider pairs, so there might well be something better. There's also a way to suppress both gliders in a pair, though that's probably less necessary since NOP operations are a dime a dozen:

Code: Select all

x = 92, y = 100, rule = B3/S23
61bo3bo$49b2o9bo5bo9b2o$49b2o15bo9b2o$61bo3b2o$62b3o2$52b2obo6b3o$51bo
bobo5bo3b2o$51bo4bo9bo$51bobobo4bo5bo$52b2obo5bo3bo10$69b2o$70b2o16b2o
$69bo18bo$89b3o$91bo4$71b2o$72b2o$71bo9$50b2o$49bobo$51bo7$51b2o$52b2o
$51bo10$29b2o$30b2o$29bo6$31b2o$32b2o$31bo19$2o$b2o$o6$2bo$2b2o$bobo!

I don't know quite how to think about that 6hd hassler yet -- the timing is fixed, so you can only get one parity of glider out of it. But as long as there has been a stable intermediate elbow target since the last use of the 6hd glider, elbow recipes can be arranged to present either parity of target to a 6hd-hassled glider.

-- It's probably time to redesign and rewrite the DOpSearch elbow operation search utility to account for more of these different possibilities. DOpSearch is actually still set up to look for combinations of gliders on three different lanes, not two... and it's inefficient in cases like 10hd where many operations involve glider pairs with minimal timing constraints -- like a -3 elbow move using a block-pull on one side followed by a block-pull on the other; to the current DOpSearch code, that looks like thousands of different recipes with different glider timings.

[codeholic]

If you're thinking about using glider hasslers with stable circuitry, it may be actually more efficient to use the hassler I mentioned in the first post, since it can eat, hassle or pass in different phases as well, but there are already known slow salvo syntheses for the pentadecathlon and the queen bee shuttle.

Code: Select all

x = 32, y = 50, rule = B3/S23
2bo$b3o$obobo$obobo$b3o$2bo3$2bo$b3o$obobo$obobo$b3o$2bo5bobo$6bo3bo$
6bo$5bo4bo8b2o$6bo12b2o$6bo3bo$8bobo10$10b3o$10bo$11bo6$20bo$19b2o$19b
obo8$29b2o$29bobo$29bo!

This allows to use both 10hd and 9hd glider pairs on the same construction arm and also, if required, 11hd or 8hd at the cost of an additional hassler or two on the other lane, as well as singleton gliders.

[codeholic]

Does the following pattern have a name? What is the name of the class of such patterns?

Oh, I found it: Glider pusher

[Gpennder_____409]

Does the following pattern have a name? What is the name of the class of such patterns? Is there a summary of these patterns of different periods somewhere?

Code: Select all

x = 20, y = 33, rule = B3/S23
13bo$12bo$12b3o13$9b2o$8bobo$7bo$7bo2bo7b2o$7bo10b2o$8bobo$9b2o2$bo$bo
$obo$bo$bo$bo$bo$obo$bo$bo!

I'm interested in both of those that change the lane of a glider and those that advance or delay it for certain number of generations, and especially in those patterns that perform a reaction on their edge.

It rephases the glider 1 generation and translates it over 5 pixels
The term for that is "Rephaser" .