## The Hunting of the New Herschel Conduits

For discussion of specific patterns or specific families of patterns, both newly-discovered and well-known.
Sokwe
Moderator
Posts: 1664
Joined: July 9th, 2009, 2:44 pm

### Re: The Hunting of the New Herschel Conduits

dvgrn wrote:Does anyone see a way to let the F171's extra glider escape?
Yes, and you're going to hate yourself for not seeing it:

Code: Select all

x = 52, y = 82, rule = B3/S23
14bo$13bobo$13bobo$14bo$25bo$25bo$23b3o9b2o3b2o$23bo11b2o3b2o8$43b2o$28b2o13b2o$28b2o9$44b2o$44b2o8$8b2o$7bobo$8bo3$43b2o$43bo$33b2o3b2o4b
3o$2bo28bo2bo4bo6bo$bobo27b3o4bo$2bo35b2o$22b2o7b2obo$22bo8bob2o6bo$
20bobo16b3o$20b2o16bo$38b2o5$2o$2o$46b2o$46bo$44bobo$3b2o39b2o$3b2o2$
51bo$49b3o$48bo$48b2o$12b2o$12bo$13b3o$15bo3$50bo$48bobo$48b3o$48bo2$
40b2o$40bo$41b3o$43bo! -Matthias Merzenich dvgrn Moderator Posts: 6977 Joined: May 17th, 2009, 11:00 pm Location: Madison, WI Contact: ### Re: The Hunting of the New Herschel Conduits Sokwe wrote: dvgrn wrote:Does anyone see a way to let the F171's extra glider escape? Yes, and you're going to hate yourself for not seeing it. Nah, I'm okay, really... So are there any other decent Edge Shooters In Passing that are Spartan or near-Spartan? I haven't done a full review, but it seems to me I would have noticed if any such thing had showed up. A for awesome Posts: 2076 Joined: September 13th, 2014, 5:36 pm Location: Pembina University, Home of the Gliders Contact: ### Re: The Hunting of the New Herschel Conduits Why have I not seen this R301 before: Code: Select all x = 69, y = 54, rule = LifeHistory 11.B2.2B2.2B$10.11BD$6.B2.12BDBD$5.2CB.B2C9B3DB$5.2C3B2C11BD$6.B.14B$8.7B.B$7.7B$6.8B$6.8B$3.3B.8B$3.12B$2.2C11B.2B$2.2C13B2C$3.B.10B.B2C$
5.10B2.B$6.9B$6.10B$6.11B$7.11B$6.13B$5.15B$2.19B$.2C19B$.2C15B.4B$2.
17B.4B$6.19B$7.19B$7.20B26.2C$8.20B23.2B2CB$5.24B22.4B$5.20B.4B13.4B.
6B4.2C$4.20B3.4B10.14B2.B2CB$3.22B3.4B8.16B.3B$3.24B2.4B5.18B2.B.2B$
4.30B2.27B$2.10B.21BD29B$2.2C4.2B2.7B.13BD31B$3.C4.B4.B.3B2.12B2D31B$
3C4.2C3.2C2B4.13B2D30B$C6.C4.C2.CB4.13BD23B.B4.B2C$8.C4.3C12.28B8.BC.
C$7.2C20.4B4.B3.2B5.7B12.C$12.C.2C12.5B7.BC5.11B9.2C$12.2C.C11.8B5.C. C4.12B$33.2C6.2C4.12B$33.C14.11B$34.3C9.4B.4B3CB$36.C9.2C4.4BC2B$47.C
4.2B3C2B$44.3C6.6B$44.C8.7B$55.B.4B$58.2B!

praosylen#5847 (Discord)

x₁=ηx
V*_η=c²√(Λη)
K=(Λu²)/2
Pₐ=1−1/(∫^∞_t₀(p(t)ˡ⁽ᵗ⁾)dt)

$$x_1=\eta x$$
$$V^*_\eta=c^2\sqrt{\Lambda\eta}$$
$$K=\frac{\Lambda u^2}2$$
$$P_a=1-\frac1{\int^\infty_{t_0}p(t)^{l(t)}dt}$$

simsim314
Posts: 1766
Joined: February 10th, 2014, 1:27 pm

### Re: The Hunting of the New Herschel Conduits

A for awesome wrote:Why have I not seen this R301 before
Haha I just wanted to publish almost the same finding from my collection when I saw your post.

Here is my version (did you peek into the collection I posted?):

Code: Select all

x = 36, y = 31, rule = LifeHistory
21.2C$19.2B2CB$19.4B$16.6B4.2C6.C$9.3B3.8B2.B2CB3.3C$8.16B.3B3.C$6.
18B2.B.B2.2C$4.25B.3B$3.10BDBD15B$2.11B2D17B$.12BD18B$4B3.23B.B2C$3B
7.16B.B3.BC.C$2B9.2B.10B10.C$B15.7B11.2C$15.11B$15.12B$15.12B$16.11B$14.4B.4B3CB$14.2C4.4BC2B$15.C4.2B3C2B$12.3C6.6B$12.C8.7B$23.B.4B$26. 4B$27.4B$28.4B$29.4B$30.4B$31.4B!

Anyway this is great H->R that can use boat for R->B (as A for awesome posted) and then pretty free to use any B->H except the one with block+snake.

For example (I think it's Bx260):

Code: Select all

x = 80, y = 48, rule = LifeHistory
17.B10.2C$15.5B8.C$15.6B4.BC.C$15.6B2.B.B2C$13.13B10.C16.2C$13.12B9. 3C14.2B2CB$13.14B6.C17.4B$12.16B5.2C8.4B.6B4.2C6.C$5.3B4.15B.B12.14B
2.B2CB3.3C$4.7B.19B9.16B.3B3.C$4.5B2CB.18B7.18B2.B.B2.2C$3.5BC2BC19B. 2B2.25B.3B$3.6B2C23BD9B2D17B$4.29BD9BD2BD17B$5.2B.6B.17B2D9BD2BD17B$5.B.8B2.2B.13B2D9BDBD15B.B2C$4.2CB.6B7.13BD23B.B3.BC.C$3.C.C3.B15.31B 10.C$3.2C19.9B4.B3.2B5.7B11.2C$24.9B7.BC5.11B$25.10B5.C.C4.12B$25.3D 2B3.2C6.2C4.12B$25.2BDB4.C14.11B$24.2B3D5.3C9.4B.4B3CB$23.5B8.C9.2C4.
4BC2B$22.4B21.C4.2B3C2B$21.4B19.3C6.6B$20.4B20.C8.7B$19.4B32.B.4B$18. 4B36.4B$17.4B38.4B$16.4B40.4B$15.4B42.4B$14.4B44.4B$13.4B46.4B$12.4B 48.4B$11.4B50.4B$10.4B52.4B$9.4B54.4B$8.4B56.4B$7.4B58.4B$6.4B60.4B$
5.4B62.4B$4.4B64.4B$3.4B66.4B$2.4B68.4B$.4B70.4B$4B72.4B!  simsim314 Posts: 1766 Joined: February 10th, 2014, 1:27 pm ### Re: The Hunting of the New Herschel Conduits New B->B in very useful place for all those who hated the old "ugly and slow" solution for badly located Bs (this is pure CatForce work with three catalysts): Code: Select all x = 25, y = 24, rule = LifeHistory 17.2B$16.4B$16.4B$15.6B2.B$15.2B2A2B.B2A$11.B.3BA2BA3B2A$5.2B3.7B2A2B .2B$2.D18B$.2D17B$2D18B$.2D17B$2.19B$5.B2.13B.B$8.12B.B2A$8.9BC4B2A$
9.7B3CB.2B$9.6B2C2BC$10.10B$14.6B$14.5B$14.6B$15.4B$15.4B$16.2B!

I think it's a great finding so I post it in separate post.

calcyman
Posts: 2243
Joined: June 1st, 2009, 4:32 pm

### Re: The Hunting of the New Herschel Conduits

New B->B in very useful place for all those who hated the old "ugly and slow" solution for badly located Bs (this is pure CatForce work with three catalysts):
Wow! It's amazing how many surprisingly simple things are only just being discovered.

I suspect we now need a variant of Hersrch which allows multiple different transients (especially now that the number of indecomposable Herschel conduits is now infinite).
What do you do with ill crystallographers? Take them to the mono-clinic!

simsim314
Posts: 1766
Joined: February 10th, 2014, 1:27 pm

### Re: The Hunting of the New Herschel Conduits

calcyman wrote:Wow! It's amazing how many surprisingly simple things are only just being discovered.
My prediction is that we're missing order of magnitude of conduits (most of the current conduits weren't using bellman, and now CatForce provides new alternative approach for search). But actually it's harder to crack than what I thought, most of the useful stuff is still the old good ones, considering that the search space is pretty limited and was explored by many talented people for two decades now.

EDIT I would say the non spartan part is now pretty well developed having syringe and snark, so we need something of this kind in spartan circuitry as well. This is why I'm working on CatForce which provides new search capabilities for the spartan circuitry.

----

Anyway here is a nice trick I've found (I like it because the emitted glider is actually part of the conduit):

Code: Select all

x = 76, y = 44, rule = LifeHistory
52.2A$50.2B2AB$50.4B$21.2A19.4B.6B4.2A6.A$20.B2AB6.2A8.14B2.B2AB3.3A$21.3B5.B2AB6.16B.3B3.A$22.B7.3B4.18B2.B.B2.2A$20.5B3.2B.B3.25B.3B$18.
4BA39B$17.4BABA39B$17.4BABA39B$18.4BA38B.B2A$20.37B.B3.BA.A$22.33B10. A$22.13B.B3.2B5.7B11.2A$23.12B4.BA5.11B$22.13B4.A.A4.12B$23.12B5.2A4. 12B$24.10B13.11B$23.9B13.4B.4B3CB$22.7B16.2A4.4BC2B$21.8B17.A4.2B3C2B$20.10B13.3A6.6B$19.10B14.A8.7B$14.2B2.11B25.B.4B$13.16B28.4B$13.16B.
2B26.4B$11.20B2A26.4B$10.19B.B2A27.4B$5.2B2.21B.B29.4B$3.D26B32.4B$.D BD25B34.4B$B3D9B2A14B35.4B$.D11B2A14B36.4B$3.27B36.4B$8.B.3B.2B3.7B2A 2B.2B34.4B$20.B.3BA2BA3B2A34.4B$24.2B2A2B.B2A35.4B$24.6B2.B37.4B$25. 4B42.4B$25.4B43.4B$26.2B45.3B$74.2B$75.B!  Last edited by simsim314 on April 25th, 2015, 7:36 pm, edited 1 time in total. calcyman Posts: 2243 Joined: June 1st, 2009, 4:32 pm ### Re: The Hunting of the New Herschel Conduits Anyway here is a nice trick I've found (I like it because the emitted glider is actually part of the conduit): That's arbitrarily extensible, and has a nice visual effect since the glider is only very slightly faster than the Herschel (and consequently I think the recovery time increases very slightly as a function of length). But at 11c/46, this is probably the fastest way to send Herschels diagonally without any of that tedious mucking about in gliderspace: Code: Select all x = 186, y = 170, rule = B3/S23 171b2o$171b2o2$140b2o34b2o6bo$140b2o7b2o25b2o4b3o$149b2o30bo$181b2o2$141bo$140bobo$140bobo$141bo40b2o$182bobo$184bo$184b2o$159bo$158bobo$
159b2o2$173b3o$164b2o8bo$165bo6b3o$162b3o$162bo2$118b2o$118b2o7b2o$
127b2o21b2o$150b2o2$119bo$118bobo$118bobo$119bo2$145b2o$144bo2bo3b2o$
145b2o4b2o10$96b2o$96b2o7b2o$105b2o21b2o$128b2o2$97bo$96bobo$96bobo$
97bo2$123b2o$122bo2bo3b2o$123b2o4b2o10$74b2o$74b2o7b2o$83b2o21b2o$106b 2o2$75bo$74bobo$74bobo$75bo2$101b2o$100bo2bo3b2o$101b2o4b2o10$52b2o$
52b2o7b2o$61b2o21b2o$84b2o2$53bo$52bobo$52bobo$53bo2$79b2o$78bo2bo3b2o
$79b2o4b2o10$30b2o$30b2o7b2o$39b2o21b2o$62b2o2$31bo$30bobo$30bobo$31bo 2$57b2o$56bo2bo3b2o$57b2o4b2o10$8b2o$8b2o7b2o$17b2o21b2o$40b2o2$9bo$8b
obo$8bobo$9bo2$35b2o$34bo2bo3b2o$35b2o4b2o12$18b2o$18b2o4$2o$2o2$13b2o
$12bo2bo3b2o$13b2o4b2o!
What do you do with ill crystallographers? Take them to the mono-clinic!

dvgrn
Moderator
Posts: 6977
Joined: May 17th, 2009, 11:00 pm
Contact:

### Re: The Hunting of the New Herschel Conduits

calcyman wrote:
Anyway here is a nice trick I've found (I like it because the emitted glider is actually part of the conduit):
That's arbitrarily extensible, and has a nice visual effect since the glider is only very slightly faster than the Herschel (and consequently I think the recovery time increases very slightly as a function of length). But at 11c/46, this is probably the fastest way to send Herschels diagonally without any of that tedious mucking about in gliderspace...
Beautiful! Of course we get the diagonal travel just as well without the cute leading-glider trick.

That B-to-B turn is a really impressively big step forward for self-constructing circuitry. There are so many new combinations of components that it's making Spartan Herschel circuits look downright unfamiliar:

Code: Select all

x = 245, y = 274, rule = LifeHistory
71.A$69.3A11.A$68.A14.3A$68.2A16.A14.A$33.A51.2A12.3A$31.3A11.A52.A$
30.A14.3A50.2A$30.2A16.A34.D$3.2A42.2A11.D21.D$4.A55.3D18.2D14.2A$4.A
.A53.D.D15.2A2.2D13.2A$5.2A2.2A51.D15.2A3.D$9.2A$43.D$20.D21.D$20.3D 18.2D$20.D.D15.2A2.2D$22.D15.2A3.D37.2A$61.2A19.A$61.A.A15.3A$63.A15.
A13.3D$51.2A10.2A19.2A8.D$51.A33.A6.3D$3.2A36.2A9.3A27.3A$2.A.A16.2A
19.A11.A27.A$2.A18.A.A17.A$.2A4.D2.2D11.A17.2A$8.3D6.2A4.2A$9.D7.A.A$19.A$10.2A7.2A81.2A$10.2A90.A.A$104.A$104.2A5$2A90.2A$.A90.2A$.A.A$2. 2A79.2A.A7.D$83.A.2A6.3D$92.2D2.D2$21.A$19.3A$9.3D6.A$9.D8.2A$8.3D4$116.2A4.2A$115.A2.A3.2A$23.2A91.2A$23.A77.D$21.A.A66.A9.2D$21.2A66.A.
A7.2D$89.A.A8.D$90.A10.D$7.2A$6.A.A112.2A$6.A91.2A16.D4.2A$5.2A82.2A
7.2A15.3D$89.2A23.2D.D$13.D2.2D23.2A$14.3D15.2A7.2A$9.2A4.D16.2A$9.2A 2$30.D10.A$31.D8.A.A$31.2D7.A.A$30.2D9.A$30.D107.2A4.2A$14.2A121.A2.A 3.2A$8.2A3.A2.A121.2A$8.2A4.2A107.D$112.A9.2D$111.A.A7.2D$111.A.A8.D$112.A10.D2$143.2A$120.2A16.D4.2A$111.2A7.2A15.3D$111.2A23.2D.D$36.D.
2D23.2A$36.3D15.2A7.2A$31.2A4.D16.2A$31.2A136.2A3.2A$169.2A3.2A$52.D 10.A$53.D8.A.A$53.2D7.A.A116.2A$52.2D9.A117.A$52.D126.A.A$36.2A141.2A
$30.2A3.A2.A$30.2A4.2A107.D$134.A9.2D$133.A.A7.2D$133.A.A8.D$134.A10.
D$176.2A$176.A$142.2A3.2A28.3A$133.2A7.2A3.A31.A$133.2A13.A$58.D.2D
23.2A60.2A$58.3D15.2A7.2A$53.2A4.D16.2A$53.2A2$74.D10.A77.2A$75.D8.A. A76.2A$75.2D7.A.A$74.2D9.A76.3D$74.D88.D$58.2A101.3D$52.2A3.A2.A$52. 2A4.2A10$79.D2.2D$80.3D6.2A.A$81.D7.A.2A2$82.2A77.2A$82.2A77.2A2$152. 2A.A7.D$152.A.2A6.3D$161.2D2.D10$185.2A4.2A$184.A2.A3.2A$70.A10.3D
101.2A$70.3A8.D88.D$73.A6.3D76.A9.2D$72.2A84.A.A7.2D$80.2A76.A.A8.D$80.2A77.A10.D2$190.2A$167.2A16.D4.2A$158.2A7.2A15.3D$96.2A60.2A23.2D. D$96.A13.2A$65.A31.A3.2A7.2A$65.3A28.2A3.2A$68.A$67.2A$99.D10.A$100.D
8.A.A$100.2D7.A.A$99.2D9.A$99.D107.2A4.2A$206.A2.A3.2A$64.2A141.2A$
63.A.A126.D$63.A117.A9.2D$62.2A116.A.A7.2D$180.A.A8.D$181.A10.D$69.2A 3.2A8.2A$69.2A3.2A9.A126.2A$82.3A104.2A21.2A$82.A97.2A7.2A$180.2A$
105.D.2D23.2A$105.3D15.2A7.2A$100.2A4.D16.2A$100.2A2$121.D10.A$122.D 8.A.A$122.2D7.A.A$121.2D9.A$121.D107.2A4.2A$105.2A121.A2.A3.2A$99.2A
3.A2.A121.2A$99.2A4.2A107.D$203.A9.2D$202.A.A7.2D$202.A.A8.D$203.A10. D2$234.2A$211.2A16.D4.2A$202.2A7.2A15.3D$202.2A23.2D2.D$127.D.2D23.2A
$127.3D15.2A7.2A82.2A$122.2A4.D16.2A91.A$122.2A112.A.A$236.2A$143.D 10.A$144.D8.A.A$144.2D7.A.A66.2A$143.2D9.A66.A.A$143.D77.A$127.2A91.
2A$121.2A3.A2.A$121.2A4.2A4$234.3D$225.2A8.D$226.A6.3D$223.3A$223.A2$
148.D2.2D$149.3D6.2A.A$150.D7.A.2A79.2A$241.A.A$151.2A90.A$151.2A90. 2A5$139.2A$140.A$140.A.A90.2A$141.2A81.2A7.2A$225.A$225.A.A7.D$220.2A
4.2A6.3D$202.2A17.A11.2D2.D4.2A$203.A17.A.A18.A$162.A27.A11.A19.2A16. A.A$160.3A27.3A9.2A36.2A$150.3D6.A33.A$150.D8.2A19.2A10.2A$149.3D13.A 15.A$163.3A15.A.A$162.A19.2A$162.2A37.D3.2A15.C$201.2D2.2A15.C.C$202.
2D18.3C$202.D21.C$201.D$234.2A$161.D3.2A15.D51.2A2.2A$146.2A13.2D2.2A 15.D.D53.A.A$146.2A14.2D18.3D55.A$162.D21.D11.2A42.2A$161.D34.A16.2A$145.2A50.3A14.A$146.A52.A11.3A$143.3A12.2A51.A$143.A14.A16.2A$159.3A 14.A$161.A11.3A$173.A! #C [[ AUTOSTART STEP 8 ]] Congratulations are definitely in order! Now, about that Spartan G-to-H... Scorbie Posts: 1460 Joined: December 7th, 2013, 1:05 am ### Re: The Hunting of the New Herschel Conduits calcyman wrote:That's arbitrarily extensible, and has a nice visual effect since the glider is only very slightly faster than the Herschel (and consequently I think the recovery time increases very slightly as a function of length). But at 11c/46, this is probably the fastest way to send Herschels diagonally without any of that tedious mucking about in gliderspace: Wow, this is as cool as the "Bobsled", although the herschel transcievers and receivers send them slightly faster. (I don't think this is "sending" Herschels anyway) Edit2: What is "tedious mucking about in gliderspace"? Edit: Forgot to add Congrats!!! Best wishes to you, Scorbie simsim314 Posts: 1766 Joined: February 10th, 2014, 1:27 pm ### Re: The Hunting of the New Herschel Conduits dvgrn wrote:Now, about that Spartan G-to-H... Haha I don't think CatForce has matured enough for that. I still struggling to sort out a better report and make efficient recursive deepening. Actually maybe dvgrn or calcyman could give some advice on this topic, I posted my thoughts here. Spartan snark/syringe is definitely on my mind when I develop CatForce. I actually tried to use CatForce to check out alternative to the snark large SL in its place, without any success. dvgrn Moderator Posts: 6977 Joined: May 17th, 2009, 11:00 pm Location: Madison, WI Contact: ### Re: The Hunting of the New Herschel Conduits Scorbie wrote:What is "tedious mucking about in gliderspace"? Well, it's not quite so tedious now that the syringe G-to-H is available. But for Spartan circuitry it's always been kind of a pain to convert Herschel signals into tandem gliders and back. The receiver half of a Herschel transceiver wasn't so bad, but by the time you got the Herschel back it took up an awful lot of space, and the Herschel always seemed to be heading off in the wrong direction. At the transmitter end you had to either build dead spark coils (which wouldn't have been so hard, really, but they never actually got included in any actual self-constructing circuitry) or you had to painfully build a very large and/or very slow Spartan G2, G5, or G6 emitter. And then, no matter what kind of transmitter you tried, the chirality of the tandem gliders would somehow always be wrong, so you'd have to twist the input around in the most awkward possible way to get a mirror-image tandem glider. It always turned out to be much less painful to just use singleton gliders and Silver reflectors. I'll be very interested to see how much of that changes after simsim314's discoveries today...! codeholic Moderator Posts: 1142 Joined: September 13th, 2011, 8:23 am Location: Hamburg, Germany ### Re: The Hunting of the New Herschel Conduits This Spartan conduit must have been found by ptbsearch. The only explanation I have why it wasn't is that ptbsearch has a bug. Similarly, one cannot find the reaction of a honeyfarm predecessor with 3 eaters, that resulted in the period-18 oscillator by 'glider_rider'. Congrats to Michael! Ivan Fomichev Sphenocorona Posts: 489 Joined: April 9th, 2013, 11:03 pm ### Re: The Hunting of the New Herschel Conduits Found a conduit which can act as a period doubler when combined with a syringe... which I realize is pointless (semi-snark plus a syringe accomplishes the same task), but it demonstrates another case of the eater2+loaf+block catalyst mentioned here, and there's a funny variant with no glider output which has a transparent block that popped up out of nowhere. The only reason I post these here is because the loaf spin catalyst might be worth investigation, though it's a bit overshadowed by recent finds... Mostly Useless: Code: Select all x = 31, y = 35, rule = LifeHistory 16.A$15.A.A$15.A.A$13.3A.2A$12.A4.B4.2A$13.3AB2A3B2AB$15.A.2A.4B$20.
3B$20.B2CB$B18.BC2BCB$2B17.2BCBCB$3B17.2BC3B$4B14.2B.4B$.4B13.6B$2.4B .B5.B.11B$3.24B$3.24B$3.25B$3.9B2D13B$4.8B2D11B.B2A$2.2CB2.16B3.BA.A$
.C.CB4.B.11B7.A$.C11.8B8.2A$2C12.7B$11.11B$10.12B$10.12B$10.11B$10.B 3C4B.4B$10.2BC4B4.2A$10.2B3C2B4.A$10.6B6.3A$9.7B8.A$8.4B.B$7.4B! Totally Useless: Code: Select all x = 30, y = 34, rule = LifeHistory 15.A$14.A.A$14.A.A$12.3A.2A$11.A4.B4.2A$12.3AB2A3B2AB$14.A.2A.4B$19.
3B$5.2B12.B2CB$4.4B10.BC2BCB$4.4B10.2BCBCB$3.6B10.2BC3B$2.3B2C2B8.2B. 4B$.4B2C3B7.6B$.8B3.B.11B$2.24B$.25B$27B$11B2D13B$11B2D11B.B2A$22B3.B A.A$5B3.13B7.A$6B6.8B8.2A$6B7.7B$5B5.11B$.4B4.12B$9.12B$9.11B$9.B3C4B .4B$9.2BC4B4.2A$9.2B3C2B4.A$9.6B6.3A$8.7B8.A$7.4B.B!
Also, don't let me forget: Those new spartan conduits are awesome!

simsim314
Posts: 1766
Joined: February 10th, 2014, 1:27 pm

### Re: The Hunting of the New Herschel Conduits

codeholic wrote:This Spartan conduit must have been found by ptbsearch.
One possible thing is that the beehive reappears quite some time after disappearance.

Another point is that ptbsearch is unsupported and pretty complex tool with bugs, hard to operate as end user, and the code is pretty hard to follow (there are somewhere around 5K lines of code as I remember).

This is why independent tool with different approach is so beneficial - it shows us bugs in the current tools and allows new search approaches.

I would be very glad if someone could support it from the "inside", as it is nice piece of code that gave us many nice conduits, and CatForce is potentially less efficient for some implementations than ptbsearch (as any tool it has its own strong and weak points).

EDIT Thx for the congrats everyone. It was actually very straight forward result from CatForce.

dvgrn
Moderator
Posts: 6977
Joined: May 17th, 2009, 11:00 pm
Contact:

### Re: The Hunting of the New Herschel Conduits

calcyman wrote:I suspect we now need a variant of Hersrch which allows multiple different transients (especially now that the number of indecomposable Herschel conduits is now infinite).
Definitely. We just need Karel Suhajda to come out of retirement and build JHersrch, as he was hoping to do a couple of years back. That was before Snarks, let alone syringes... Now it would be nice to be able to say, e.g.,

"I have a SW-traveling tandem glider at (0,0,T=0), or it could be just a single glider instead. I want a north-traveling Herschel at (-30,53,T=753+8N). Here are the obstacles attached to the start point, and here are the obstacles that will move with the end point. What's the simplest way to get the signal there with Spartan/non-Spartan [delete whichever is inapplicable] circuitry?"

Almost half of the time the solution for the above might be a simple arrangement of Snarks and a syringe... and the other half of the time the output Herschel will be the wrong color, so it will need just one more conduit at the end, unless that gets in the way of some obstacle somehow, in which case things get more complicated.

I'd really love to have a precalculated lookup table available for each combination of input and output (object, orientation), so that it would be possible to drag the output object around in a GUI, maybe somewhat along the lines of the Seeds of Destruction Game, and see the candidate conduit adjust itself in real time.

However, it might still be a good intermediate step to build all the composite H-to-H conduits that include one or more B-to-Bs and fit inside 100x100 or so, and add those to Hersrch along with all the other recent new discoveries. Some usable information will come out of that, about which conduits can precede or follow the new additions.

The other intermediate step is to finally collect all lane and timing statistics for all of the known H-to-Gs, in a form that a script could use -- because really one of the most common situations is the opposite of the above scenario: I have a Herschel, and what I want is exactly the right output glider.

I have the H-to-G classifier script nicely started now, but I'm out of spare time again (funny how that keeps happening). So it may be several weeks before I get the H-to-G collection all sorted out.

Kazyan
Posts: 984
Joined: February 6th, 2014, 11:02 pm

### Re: The Hunting of the New Herschel Conduits

Yeah, I think all the Spartan self-constructing circuitry magic is going to be quite different from here on out. It's an impressively cheap and useful conduit.

On the Bellman front, here's a Pi-to-2C that works different from the Pi-QB-2C converter. It likewise cannot be hooked up to anything yet:

Code: Select all

x = 21, y = 22, rule = Life
6b2o5b2o$6bo7bo$8bo3bo$4b4obobob4o$4bo2bobobobo2bo$2o6bo3bo6b2o$2o17b
2o13$9b3o$9bobo$9bobo! Notice that the hexagonal keys do the same thing as two eaters, but are tucked in by one cell so that the reaction does not later obliterate them, where two eaters cannot do the same. There are at least four century-producing conduits that are currently dead ends: my two previous C-to-C's, this one, and the Pi-QB-2C converter. So, I'm going to run a century through CatForce now and see what happens. --- On the subject of Hersch, would it be possible/easy to make it so that Hersch accepts the list of conduits from an external file instead of having everything built-in? Tanner Jacobi Coldlander, a novel, available in paperback and as an ebook. simsim314 Posts: 1766 Joined: February 10th, 2014, 1:27 pm ### Re: The Hunting of the New Herschel Conduits Kazyan wrote: I'm going to run a century through CatForce now and see what happens. If you intend to run CatForce, please try add combine-results yes To the .in and run with 1-2 catalysts. This flag combines all the possible catalysts and uses filters and survival time only in the end. For 1-2 catalysts it's great feature. Runs few minutes and gives all the combined results of many catalysts. I want to add some more features to this option, and will report about it. Please make sure you take the really latest version (I posted this feature as it should work just few minutes ago). If you run from 3 on, check out the fit-in-width-height flag, as it can speed up the search significantly. With such searches the combine-results feature is not working so well yet. As a side note if you don't want to use some catalyst for the CatForce just add additional char (like #cat instead of cat) to the line with the unwanted property. If you have some questions and/or bugs please report. dvgrn Moderator Posts: 6977 Joined: May 17th, 2009, 11:00 pm Location: Madison, WI Contact: ### Re: The Hunting of the New Herschel Conduits Kazyan wrote:On the subject of Hersch, would it be possible/easy to make it so that Hersch accepts the list of conduits from an external file instead of having everything built-in? I've made the necessary changes to the current build... in other words, no changes are needed: that's already how Hersrch works. The external file is called "data.ini", in a subfolder called "data" in Hersrch's root directory. Recent builds (i.e., from 2012) are set up so that you can run Spartan searches instead of standard searches, by renaming "data" to "standard-data" and renaming the "Spartan-data" folder to just plain "data". I'll probably ship a batch file to do that trick automatically in the next release. Anyway, you can make whatever changes you want to data.ini. If you add any newly-discovered components, and run enough tests to make sure Hersrch is still working correctly, then please post the "u" and "v" files and the data.ini additions to this thread, and I'll roll them into the next semi-official release of Hersrch. The link is to an introductory posting that explains all the non-obvious details about Hersrch that I could think of. Please post questions there if anything is still confusing. simsim314 Posts: 1766 Joined: February 10th, 2014, 1:27 pm ### Re: The Hunting of the New Herschel Conduits @dvgrn A bit off topic: have you considered to use .in file like bellman and CatForce instead of plain command line? I think this format makes the input format so much cleaner. Instead of unpleasant parsing with meaningless shortcuts and parameters, the user just writes meaningful "codes" as prefixes. It's so much more readable and user friendly, and requires so little on the development side. Like instead of: -p 250 -o result.rle -f 1.log -n 25 -s -u 100 -t 10000 The user creates file with: period 250 output-file result.rle log-file 1.log max-results 25 sort-result-by-track-size tree-size-bytes 10000 dvgrn Moderator Posts: 6977 Joined: May 17th, 2009, 11:00 pm Location: Madison, WI Contact: ### Re: The Hunting of the New Herschel Conduits simsim314 wrote:@dvgrn A bit off topic: have you considered to use .in file like bellman and CatForce instead of plain command line? Sure, it's a good idea. What I was actually planning to do at some point -- several times actually -- was write a Python-script front end to gather most of that information, and then have the script call Hersrch and display candidate results in context. I'm still a lot more likely to do that than to bother rewriting Hersrch code to change where it gets its input. Hersrch is very good at what it does, but I don't think the basic search engine can be salvaged, now that many connections will be made with glider+Snark+syringe, and signal fanouts will often be done that way too rather than with tandem-glider Herschel splitters. Your incredibly versatile new B-to-B conduit was really just the final straw. Hersrch could never really handle transceivers either, by the way, except on a case-by-case basis -- years ago we probably missed a lot of small asymmetric transceiver-based guns. Not that that matters much at this point, since they'd all have been replaced with chris_c's project anyway! Hersrch really assumes it will have a reasonably small number of plain-vanilla Herschel-to-Herschel conduits to work with. It will pretty definitely have to be rewritten from the ground up to deal with elementary conduits instead. That might be an opportunity to address some of Hersrch's other limitations -- build something that can look for ways to avoid specific obstacles, for example, instead of just assuming that a smaller bounding box is always better. So given that the old Hersrch is something of a dead end at this point, I was figuring on doing another maintenance release or two of the data files eventually, but I wasn't planning on doing anything at all with the executable. If anyone wants to tackle a rewrite of the input parser, or wants to make a cross-platform version of Hersrch that's better integrated with Golly... well, hopefully the new data.ini will work just as well as with the current Hersrch.exe. Kazyan Posts: 984 Joined: February 6th, 2014, 11:02 pm ### Re: The Hunting of the New Herschel Conduits Turns out that CatForce memory-leaks. Oops. I saved the interesting results from the first half of the run, though! A handful of transparent objects, two new useless C-to-C's for the collection I'm apparently making, and a C-to-Pi not in my records that is probably known, but I don't know where it's been recorded. Code: Select all x = 220, y = 248, rule = B3/S23 43b2o$3bo27bo11b2o22bo$2bobo25bobo33bobo21bo$3bo27bo35bo20b3o$87bo70bo$87b2o68bobo$142bo15bo$142b3o$145bo$144b2o3$14b3o25b3o33b3o$13bo2bo24b
o2bo32bo2bo$13b2o26b2o34b2o4$149b3o$148bo2bo$148b2o31$140bo59bo$140b3o
16bo40b3o$143bo13b3o43bo$89b2o51b2o12bo45b2o$88bobo65b2o$89bo$50b3o 152b2o$2o4b3o36b2o2bo2bo152bobo$2o3bo2bo36b2o2b2o155bo$5b2o3$147b3o43b 3o$146bo2bo42bo2bo$82b3o61b2o44b2o$81bo2bo$81b2o39$200bo$200b3o$203bo$202b2o6$9b2o$8bobo87b2o59b2o$8bo7bo81b2o59b2o$7b2o5b3o190b3o$13bo192bo
2bo$13b2o53bo93b2o42b2o$66b3o93bobo$65bo98bo$60b2o3b2o97b2o50b2o$59bo 2bo31b3o51b3o65bobo$60b2o31bo2bo50bo2bo67bo$93b2o52b2o69b2o$100bo$4b3o 92bobo$3bo2bo92bobo$3b2o95bo2$55b3o$54bo2bo$54b2o24$14bo$12b3o$11bo$
11b2o3$100bo$100b3o$54bo48bo$54b3o45b2o$57bo$2b3o51b2o$bo2bo$b2o103bo$51bo53bobo$50bobo52bobo$11b2o38bo6b3o45bo$11bobo43bo2bo$13bo43b2o38b3o$13b2o81bo2bo$96b2o42$21bo$21b3o$14b2o8bo$15bo7b2o$15bobo$16b2o$28b2o$29bo$29bobo$30b2o2b2o54bo39b2o$34b2o53bobo39bo$90bo40bobo$132b2o$137b 3o$83bo52bo2bo$82bobo51b2o$83bo3b3o$86bo2bo$86b2o$26b3o$25bo2bo$25b2o 3$35b2o$35bobo$37bo\$37b2o!
Tanner Jacobi
Coldlander, a novel, available in paperback and as an ebook.

simsim314
Posts: 1766
Joined: February 10th, 2014, 1:27 pm

### Re: The Hunting of the New Herschel Conduits

Kazyan wrote:Turns out that CatForce memory-leaks.

There was very old memory leak, and there is a new memory intensive operation that was reduced.

The current version should work fine.

Kazyan
Posts: 984
Joined: February 6th, 2014, 11:02 pm

### Re: The Hunting of the New Herschel Conduits

There was very old memory leak, and there is a new memory intensive operation that was reduced.

The current version should work fine.
Awkward--I got the code, but forgot to compile it, and am using the old CatForce.

Also, I just realized I shouldn't use the "*" symmetry character for symmetrical still lifes, which...should make things go faster.
Tanner Jacobi
Coldlander, a novel, available in paperback and as an ebook.

simsim314
Posts: 1766
Joined: February 10th, 2014, 1:27 pm

### Re: The Hunting of the New Herschel Conduits

Kazyan wrote:Also, I just realized I shouldn't use the "*" symmetry character for symmetrical still lifes
Please have a look at 3.in as it's not only takes into account all spartan SLs, but with "forbidden" flag that removes all the common eaters configurations from the results.

In general I wouldn't touch the Catalyst section of it, as it's very well defined. And unless you want to add special non spartan catalysts, I would just leave it as it is. I personally never touch this section in day-to-day searches.