Let's find a G-to-X (Done!)

[Kazyan]

I was actually thinking of starting with a pi, as a glider can convert five different still lifes to it cleanly, thus allowing for five different bait still lifes.

Martin Grant brought up this on page 2 of the Herschel thread. I'm not sure who all is exploring ways to obsolete the Herschel transciever, but I think it's worth looking into to see if we can make a G-to-H conduit, or a color-changing G-to-G. I suggest in-depth ptbsearch and Bellman runs on a Pi's explosion, and exploring in directions where the activity clusters around the site of our indeterminate bait SL.

Here are some Pi catalysts I found to get us started. The next steps are to investigate Bellman interactions with the front or the back of a Pi instead of the sides, and go from there. EDIT: I'm Bellman-searching the front end with repair and activity set to 9 right now.

Code: Select all

x = 93, y = 19, rule = B3/S23
15bo$14bobo$14bobo$o7b2o3b2ob3o$3o5b2o9bo23b2o$3bo9b2ob3o25bo$2bo10b2o
bo27bobo$bo3b2o32b2o4b2o31bo$b2o2b2o32bo37bobo$41bo36bo$37b4obo$36bo4b
o34b5o$36b5o34bo4bo$73bo2b2o$36b3o34b2obo$35bo2bo37bo$8b3o24b2o11b3o
25b2o11b3o$7bo3bo35bo3bo36bo3bo$7b2ob2o35b2ob2o36b2ob2o!

[dvgrn]

Here are some Pi catalysts I found to get us started.

This is an obvious point, but it's surprising how many beautiful new catalyst discoveries I've made, only to find out that I didn't guard the input lane correctly, and my search program has placed a key catalyst right in the way of the incoming glider.

I believe for these catalysts the only available input lanes (no matter what the target object turns out to be) are these, from the northeast:

Code: Select all

x = 113, y = 27, rule = B3/S23
30bo39bo40bo$29bo39bo40bo$29b3o37b3o38b3o$15bo$14bobo$14bobo$o7b2o3b2o
b3o$3o5b2o9bo23b2o$3bo9b2ob3o25bo$2bo10b2obo27bobo$bo3b2o32b2o4b2o31bo
$b2o2b2o32bo37bobo$41bo36bo$37b4obo$36bo4bo34b5o$36b5o34bo4bo$73bo2b2o
$36b3o34b2obo$35bo2bo37bo$35b2o39b2o6$9b2o38b2o39b2o$9b2o38b2o39b2o!

[Kazyan]

Unfortunately, yes, and I'm trying to find a catalyst at the front that will allow both lanes. Note that catalyst systems 1 and 2 are compatible if you get rid of catalyst 2's eater, but the arrangement blocks any entry lanes. I hope that front-end and far-to-the-side catalysts will be more flexible than these diagonally oriented ones.

EDIT: There's not a lot of room for it, but I'm going to try to find a Bellman replacement for the lower block here.

Code: Select all

x = 32, y = 28, rule = B3/S23
8bo$7bobo$8bo2$6b5o$5bo4bo2bo$4bo2bo5b3o$bo2bob2o8bo$obobo5bo4b2o$bo2b
o4bobo16bobo$4b2o2bo2bo16b2o$9b2o18bo4$2o$bo28b2o$bobo26b2o$2b2o$6b2o$
5bobo$5b2o2$18b2o$18b2o2$14b2o$14b2o!

[Kazyan]

This is the most bizarre catalyst I've yet seen come out of Bellman.

Code: Select all

x = 16, y = 24, rule = B3/S23
9bo$8bobo$3b2o3bobo$3bo3b2obob2o$2obobo2bobob2o$obobob2o2bo$2bobobo3bo
$2b2o3b3o2$7b5o$7bo4bob2o$9b2obob2o$10bobo$10bobo$11bo7$4b3o$4bobo$4bo
bo!

It barely works because the R it generates is quick to poke it again, and it's quite hard to avoid destruction by perturbing another part of the Pi, but I'm keeping it here in case it ever becomes useful. Its eating action appears to be a bit flexible, anyway.

[Scorbie]

This is the most bizarre catalyst I've yet seen come out of Bellman.

I would have believed you if you said that you found it with dr.

[Kazyan]

This one is much simpler and more usable.

Code: Select all

x = 9, y = 19, rule = B3/S23
5b2o$2o2bobo$o3bo$b3ob4o$3bobo2bo12$4b3o$4bobo$4bobo!

[Kazyan]

So close to a second Snark...

Code: Select all

x = 18, y = 24, rule = B3/S23
10bo$10b3o$13bo$12b2o13$3b2o10bobo$3bo2bo8b2o$4b3o9bo$2o$o2bob2o8b2o$
2b2obo9b2o$5bo$5b2o!

[Kazyan]

Code: Select all

#C Stable glider-to-Herschel converter--recovery time of 78 ticks.
#C Tanner Jacobi, March 19th, 2015
#C Special thanks to Mike Playle for Bellman, which made this converter possible.
#C Welded eater to suppress FNG is only partially optimized.
x = 55, y = 68, rule = B3/S23
o$b2o$2o18$19bobo$20b2o$20bo4$35bo$35b3o$38bo$37bobo$37bobo$38bo5$53b
2o$53b2o4$33b2o$32bobo$32bo$31b2o7b2o$40b2o2$48b2obo$48b2ob3o$54bo$48b
2ob3o$47bo2b2o$46bobo$45bobob2obo$46bo2bob2o$49bo$48b2o$45bobo2b2o$45b
2o2bo2bo$50b2o6$35b2o$34bo2bo$35b2o!

EDIT: Bookend-table-snake replaced with a fishhook. Don't know how I missed that.

EDIT2: Moved the block.

[Extrementhusiast]

Code: Select all

#C Stable glider-to-Herschel converter--recovery time of 78 ticks.
#C Tanner Jacobi, March 19th, 2015
#C Special thanks to Mike Playle for Bellman, which made this converter possible.
#C Welded eater to suppress FNG is only partially optimized.
(RLE)

...That just opened up a whole new realm of reflector possibilities. Somebody get me a list of Herschel-to-glider converters, stat!

[codeholic]

Congratulations!

[simsim314]

WOW!! Only 78 ticks, great news! This is the greatest discovery in herschel circuitry in the past 10 years! And probably the most important component for non spartan stable circuitry.

Huge congrats!

[Kazyan]

Thanks, guys. I booted it out onto the forums as soon as I could get it to work, but now I've given it touch-ups; the bounding box is 7 columns smaller than the version I posted an hour ago.

[simsim314]

Meanwhile 180 degree reflector:

Code: Select all

x = 36, y = 58, rule = LifeHistory
2.A$A.A$.2A4$16.A$16.3A$19.A$18.A.A$18.A.A$19.A5$34.2A$34.2A2$21.A$
22.A$14.2A4.3A$13.A.A$13.A$12.2A7.2A$21.2A2$29.2A.A$29.2A.3A$35.A$29.
2A.3A$28.A2.2A$27.A.A$9.A16.A.A.2A.A$9.3A15.A2.A.2A$12.A17.A$11.2A16.
2A$26.A.A2.2A$26.2A2.A2.A$31.2A7$31.2A$31.2A9$16.2A$16.2A!

I think that fast glider duplicator is a game changer.

[gameoflifeboy]

WOW!! Only 78 ticks, great news! This is the greatest discovery in herschel circuitry in the past 10 years! And probably the most important component for non spartan stable circuitry.

Huge congrats!

Why, how big was the last glider-to-herschel converter? And how long did it take to recover?

[biggiemac]

Wow! Well done.

[Scorbie]

Wow. Real congrats for the discovery and your dedicated search! Great to see it in time! Edit:ed the reflector thingy.

[codeholic]

Why, how big was the last glider-to-herschel converter? And how long did it take to recover?

I guess, the fastest was the 466-tick reflector, right?

Code: Select all

#C Calcyman, 24 January 2009
x = 100, y = 80, rule = B3/S23, color=(64,0,0)
80boo$80boo5boo$87boo$51boo$52bo$39bo11bo14boo17boo$39b3o9boo14bo17boo
$42bo24bobo21boo$29boo10boo25boo21boo$4bo9bo15bo$4b3o5b3o15bobo$7bo3bo
19boo$bbo3boo3boo41boo$obo51boo$boo6$20boo$20boo23boo$8boo35bo$7bobbo
35b3o$bboo4boo38bo$bobo$bo$oo$10boo$10bo$11b3o84boo$13bo84bo$96bobo$
96boo$72boo$36bo35boo$36b3o$39bo$38boo$19bo$19b3o$22bo51boo$21bo15boo
36bo$21boo14boo33b3o$17boo53bo$18bo29boo23boo$18bobo27bobo23bo$19boo
29bo23bobo$50boo23boo$46boo43bo$47bo41b3o$46bo41bo$13boo32b3o38boo$13b
oo34bo$29boo$29bo$30b3o$bboo28bo8boo14bo9bo$3bo23boo12bobo13b3o5b3o$3b
obo21bo15bo16bo3bo$4boo19bobo15b3o13boo3boo$25boo19bo$45boo$95boo$95b
oo4$42boo46boo$42boo46boo$94boo$94boo$$13boo$9boobboo73boo$8bobo62boo
13boo$8bo23boo19bo6boo11bobo$7boo23bo18b3o6boo13bo$33b3o14bo24boo$35bo
14boo!

And the smallest was probably Silver's reflector:

Code: Select all

#C 497-tick stable reflector
#O Stephen Silver
#C Discovered in November 1998
#C The smallest and fastest spartan 180- and 90-degree stable reflector,
#C the smallest stable glider duplicator altogether, as of February 2015.
x = 194, y = 219, rule = B3/S23
182b2o$182b2o$162bo$150bo11b3o$148b3o14bo$132bo14bo16b2o$132b3o12b2o$
135bo$134b2o3$135b2o$135b2o17b2o$154b2o2$192b2o$192b2o3$151b2o$151bo
19b2o$152b3o15bobo$154bo15bo$148b2o19b2o$148bo$149b3o$151bo4$132b2o$
131bobo$131bo25b2o$130b2o25bo$138b2o15bobo$138b2o15b2o12$137b2o$136bob
o$136bo$135b2o9$147b2o$147b2o6$136b2o$137bo19b2o$137bobo17bo$138b2o15b
obo$150bo4b2o$149bobo$149bobo$138b2o10bo$137bobo$137bo$136b2o$151b2o$
151bo$152b3o$154bo12$125b2o$124bobo$126bo122$bo$b2o$obo!

[simsim314]

Why, how big was the last glider-to-herschel converter? And how long did it take to recover?

I'm not sure, but as far as I know this one is the fastest (386 ticks):

Code: Select all

x = 162, y = 138, rule = B3/S23
o$b2o$2o45$134b2o$133bo2bo$134b2o26$137b2o$137b2o2$128b2obo$128bob2o$
146b2o$146bo$144bobo$144b2o2$129bo$129b3o$120bo11bo$94bo9bo15b3o8b2o
14bo$94b3o5b3o18bo22bobo$97bo3bo20b2o23bo$96b2o3b2o$152bo$150b3o$149bo
$149b2o2$96bobo$97b2o58b2o$97bo12b2o46bo$110b2o23b2o21bob2o$98b2o35bo
14b2o4b3o2bo$97bo2bo35b3o11b2o3bo3b2o$92b2o4b2o38bo16b4o$91bobo47b2o
15bo$91bo48bobo12b3o$90b2o48bo13bo$100b2o37b2o14b5o$100bo58bo$101b3o
53bo$103bo53b2o5$127bo$127b3o$130bo$129b2o7$139b2o$132b2o5bobo$132b2o
7bo$141b2o2$128bo$127bobob2o$127bobobobo$124b2obobobobo2bo$124bo2bo2b
2ob4o$126b2o4bo$132bobo$133b2o!

Fast G->H considered the holy grail of herschel circuitry. This means we can cut the time of anything using herschels (reflectors, duplicators, guns, calculators etc) by factor of 4.

The only thing we left with, is fast spartan G->H (or at least built from SLs with reasonable construction time), this is needed for self replication circuitry i.e. Gemini like construction (but the technology used to discover such conduits is very well explored and old - this is new discovery using Bellman search utility by Mike Playle from 2013, that searches in space of arbitrary complex SLs).

[codeholic]

Welded eater to suppress FNG is only partially optimized.

In fact it is optimized quite well. All I could find is just another variant with a slightly smaller bounding box:

Code: Select all

x = 36, y = 39, rule = B3/S23
obo$b2o$bo4$16bo$16b3o$19bo$18bobo$18bobo$19bo5$34b2o$34b2o4$14b2o$13b
obo$13bo$12b2o7b2o$21b2o2$29b2obo$29b2ob3o$35bo$29b2ob3o$28bo2b2o$27bo
bo$26bobob3o$27bo2bo2bo$30bobobo$29b2o2bobo$26bobo2bobobo$26b2o2b2ob2o
!

[Kazyan]

In fact it is optimized quite well. All I could find is just another variant with a slightly smaller bounding box:

Code: Select all

x = 36, y = 39, rule = B3/S23
obo$b2o$bo4$16bo$16b3o$19bo$18bobo$18bobo$19bo5$34b2o$34b2o4$14b2o$13b
obo$13bo$12b2o7b2o$21b2o2$29b2obo$29b2ob3o$35bo$29b2ob3o$28bo2b2o$27bo
bo$26bobob3o$27bo2bo2bo$30bobobo$29b2o2bobo$26bobo2bobobo$26b2o2b2ob2o
!

Ah, alright. I guess the current form should be the canonical one, then; the bounding box doesn't need to be optimized in that direction because Herschel circuitry is going to be put over there anyway.

[Scorbie]

The only thing we left with, is fast spartan G->H (or at least built from SLs with reasonable construction time), this is needed for self replication circuitry i.e. Gemini like construction (but the technology used to discover such conduits is very well explored and old - this is new discovery using Bellman search utility by Mike Playle from 2013, that searches in space of arbitrary complex SLs).

Is it possible to remove the FNG? This would make it much cleaner. I was thinking something like this:

Code: Select all

x = 36, y = 41, rule = B3/S23
obo$b2o$bo4$16bo$16b3o$19bo$18bobo$18bobo$19bo5$34b2o$34b2o4$14b2o$13b
obo$13bo$12b2o7b2o$21b2o2$29b2obo$29b2ob3o$35bo$29b2ob3o$30bobo$30bobo
$31bo6$21b2o$21b2o!

Which uses this transparent block. (I saw a H2G and mistakenly thought that it worked with this g2h.)

Code: Select all

x = 5, y = 14, rule = LifeHistory
.2A$.2A2$.3D$2.D$2.3D6$2.E$.3E$2E.E!

[simsim314]

Notice that if we leave a block, it gets destroyed but we return the G->H

Code: Select all

#C [[ THUMBNAIL AUTOSTART LOOP 250 ]]
x = 36, y = 29, rule = B3/S23
obo$b2o$bo4$16bo$16b3o$19bo$18bobo$18bobo$19bo5$34b2o$34b2o4$14b2o$13b
obo$13bo$12b2o7b2o$21b2o2$29b2o$29b2o!

Now if we could convert glider back into a block in similar way we can convert glider into boat:

Code: Select all

#C [[ THUMBNAIL AUTOSTART LOOP 12 GPS 5 ]]
x = 10, y = 6, rule = B3/S23
6b2obo$6bob2o2$3o$2bo$bo!

We could get something more "self construction friendly".

[Kazyan]

Is it possible to move the herschel a little bit to remove the glider eater? This would make it much cleaner. I was thinking something like this:

Code: Select all

x = 36, y = 41, rule = B3/S23
obo$b2o$bo4$16bo$16b3o$19bo$18bobo$18bobo$19bo5$34b2o$34b2o4$14b2o$13b
obo$13bo$12b2o7b2o$21b2o2$29b2obo$29b2ob3o$35bo$29b2ob3o$30bobo$30bobo
$31bo6$21b2o$21b2o!

Which uses this transparent block. (I saw a H2G and mistakenly thought that it worked with this g2h. Is there a h2h with this component?)

Code: Select all

x = 5, y = 14, rule = LifeHistory
.2A$.2A2$.3D$2.D$2.3D6$2.E$.3E$2E.E!

Yes, there's a component that uses that. It's a bit of a tight fit, but workable, I think.

Code: Select all

x = 44, y = 79, rule = B3/S23
obo$b2o$bo4$16bo$16b3o$19bo$18bobo$18bobo$19bo5$34b2o$34b2o4$14b2o$13b
obo$13bo$12b2o7b2o$21b2o2$29b2obo$29b2ob3o$35bo$29b2ob3o$30bobo$30bobo
$31bo2$32bo$30b3o$29bo$29b2o$21b2o$21b2o6$38b2o$38bo$36bobo$36b2o9$40b
2o$40bobo$42bo$42b2o$32b2o$32b2o$13bob2o$13b2obo2$22b2o$22b2o7$32b2o$
32bo$30bobo$30b2o!

Now, I think I had a blockic seed for a beehive with tail around here somewhere...

EDIT: Yep, a little artifact from my "interesting reactions" file before apgsearch:

Code: Select all

x = 13, y = 22, rule = B3/S23
2o$2o$7b2o$7b2o3$2b2o2b2o$2b2o2b2o9$2o$2o2$11bo$10b2o$10bobo!

[simsim314]

It's a bit of a tight fit, but workable, I think.

Well eater2 is also kind of a problem... at least we don't have for now any way to reasonably construct eater2 with slow salvo, even not two construction arm recipe.

Also the timing of this conduit increases as the 78 ticks doesn't work for the second part:

Code: Select all

#C [[ AUTOSTART LOOP 500 ]]
x = 165, y = 115, rule = B3/S23
164bo2$164bo14$o$b2o$2o18$19bobo$20b2o$20bo4$35bo$35b3o$38bo$37bobo$
37bobo$38bo5$53b2o$53b2o4$33b2o$32bobo$32bo$31b2o7b2o$40b2o2$48b2obo$
48b2ob3o$54bo$48b2ob3o$49bobo$49bobo$50bo2$51bo$49b3o$48bo$48b2o$40b2o
$40b2o6$57b2o$57bo$55bobo$55b2o9$59b2o$59bobo$61bo$61b2o$51b2o$51b2o$
32bob2o$32b2obo2$41b2o$41b2o7$51b2o$51bo$49bobo$49b2o!

EDIT Great news for beehive with tail.

[codeholic]

This makes an option for a different catalyst, e. g.:

Code: Select all

x = 44, y = 79, rule = B3/S23
obo$b2o$bo4$16bo$16b3o$19bo$18bobo$18bobo$19bo5$34b2o$34b2o4$14b2o$13b
obo$13bo$12b2o7b2o$21b2o2$29b2o$29b2o2$28bob3o$28b2o3bo$31b3o$28b3o$
28bo3b2o$33bo$30b3o$29bo$29b2o$21b2o$21b2o6$38b2o$38bo$36bobo$36b2o9$
40b2o$40bobo$42bo$42b2o$32b2o$32b2o$13bob2o$13b2obo2$22b2o$22b2o7$32b
2o$32bo$30bobo$30b2o!

That's the catalyst:

Code: Select all

#P 0 0
..........**..????..............................................
..........**..????..............................................
..............????..............................................
.........*.***????..............................................
........??*...????..............................................
.......???????????..............................................
.......???????????..............................................
.......???????????..............................................
.......???????????..............................................