This is an advantage in some ways, but it's a disadvantage in others, as it's harder to perturb an active object that is only moving forwards without disrupting its forwards movement mechanism than to perturb an active object that is moving both forwards and outwards. However, to be fair, only some conduits maintain the forward movement mechanism of the active region. Also, it is possible to turn an R−B or R−H into useful signals.Kazyan wrote: ↑December 4th, 2020, 3:13 pmA Herschel that comes out of a conduit will transports its 'explosion' in a distinct path and direction, and this path vacates the place where it first appeared, without flailing about in such a way that it is nigh-guaranteed to knock over catalysts from the conduit it came from. (Looking at you, century and Pi, but I'm also giving hard side-eye at R.)
While there is a commonly edgeshot predecessor of an R-sequence that goes back the way it came, the same is not true of the century. The century typically comes out either pointing forwards or sideways (specifically, turning to the right), which, in my opinion, is better than the Herschel, which typically only comes from one direction, because this allows one to place catalysts behind or to the right of the input century, just not both. On the other hand, I don't think that there are any conduits forming Herschels forming the front or the side. This isn't very much of a problem for in terms of making conduits that input Herschels because the Herschel only expands in all directions for about 10–12 generations before switching to directed movement, so there weren't many opportunities to perturb the Herschel from behind in the first place, but it is a problem in terms of making/finding conduits that output Herschels that can fit into some input conduit, such as this H→L.Kazyan wrote: ↑December 4th, 2020, 3:13 pmFor any new conduit, if the input is an H, it is almost guaranteed to be able to connect to something as long as you don't put a catalyst directly behind the Herschel. If its output in a Herschel, it has probably (but not always) formed by launching itself 20 cells away from an active region via B-heptomino. It has also probably formed in a way such that it's pointing away from your new conduit, rather than an R or C, which like to do the opposite of that.
R-sequences sometimes form from the side, although that doesn't occur nearly as often. However, even R-sequences that travel backwards can still be extracted. There are several conduits that accept an R from behind, such as RFx36R and RR73H. In addition, I have found several examples of partial conduits intaking backward-forming R-sequences which use a sacrificial object that looks like it can be replaced by a permanent catalyst. Here is a partial RB42R:
Code: Select all
x = 8, y = 9, rule = TripleB3S23
.2D3.2B$.D.2A2.B$2.DEFCB$4.G3$G$3G$.G!
Code: Select all
x = 8, y = 8, rule = TripleB3S23
6.2D$5.CE$4.G.G$4.2G2$.2G$2G$.G!
If nothing else, one can have the B-sequence or Herschel sequence crash into junk that the reaction also created then extract a signal from the R−B or R−H, each of which we know how to do.
While Herschels tend to have better output clearance than other active regions, simply ending in a Herschel is not a proof of connectability. For example, consider this periodic conduit.
Code: Select all
x = 45, y = 46, rule = LifeHistory
7$2A$.A$.A.A$2.2A12.3D$17.D$15.3D8$18.C15.2A$18.3C13.A$19.C15.A$34.2A
2$41.2A$41.A$42.A$41.2A2$28.A2.A9.2A$26.2A4.2A6.A.A$28.A2.A9.A$31.2A.
2A6.3A$33.2A.2A6.A$31.A2.A$29.3A3.3A$31.A$17.2A12.2A.A$17.2A13.A2.A$35.
A$32.A2.A$33.2A2.2A$37.A.A$39.A$39.2A!
Code: Select all
x = 23, y = 36, rule = LifeHistory
16.2A$15.A2.A$15.A.A$16.A11$2.C$3C$C.C$C8$10.3D$12.D$11.3D2$16.2A.A$16.
2A.3A$22.A$16.2A.3A$17.A.A$17.A.A$18.A!
As I have stated before, Herschels may be more likely to be outputted with good clearance than some other regions, but there are Herschel-making conduits whose outputs are inaccessible, and there are also Herschel-intaking conduits whose inputs are inaccessible. Some conduits that accept other active regions but don't connect to any preceding conduits are genuinely unrealistic (although I can think of a ∏→B that ordinarily falls under those lines but is still very useful), but for many, the problem is not that the input is unrealistic but that not enough research has been done into searching for conduits that make those input regions. For example, I found a partial CFx112 that makes a forwards century with wonderful clearance. All that it needs is a catalyst that can replace the sacrificial block.
Edit: gmc_nxtman found a H→B that can be turned into an H→H that is definitely unconnectable because the output Herschel's first natural glider would destroy a fishhook, and any catalyst that suppressed it would block the input Herschel.
Code: Select all
x = 58, y = 46, rule = LifeHistory
47.A$47.3A$37.2A11.A$38.A10.2A$38.A.AB7.5B$39.2AB.3B5.4B$41.7B2.6B$
41.16B$42.15B$41.15B$39.17B$32.2A3.D18B$32.2A2.B2DC16B$36.BDBCBC15B$
29.2A5.3B3C15B$28.B2AB4.5BC15B$29.3B3.13B.6B$28.B.B3.4B5.4B.6B.2B$27.
10B5.4B3.7B2A$25.11B5.4B4.7B2A$24.11B5.4B5.6B.B$22.12B5.4B5.6B$21.12B
5.4B6.6B$3.A16.14B3.4B8.B.B$2.A.A13.17B.4B10.3B$2.A.A12.22B10.B2AB$.
2A.2A10.6B2A14B12.2A$4.B9.8B2A16B$.2AB2AB6.27B$2.A.2A32B.B2A$A3.31B4.
BA.A$2A4.29B7.A$4.29B9.2A$4.28B$4.4B2A22B$2.2AB.2B2A21B$.A.AB.5B.15B.
B.B2A$.A5.3B2.16B2.BA.A$2A6.3B.12B.2B6.A$7.B2AB.11B10.2A$8.2A3.9B$14.
3B.B$13.2B$12.2BAB$13.A.A$14.A!