Arks

[Tropylium]

Pentadecathlon.com lists 5 arks (puffers composed of two switch engines). Is this list supposed to be even halfway complete? The wiki states probably on this basis that "only 5 are known", but I could without much effort quickly find a whole lot more… I can list them if it turns out this is an area that has not been explored much (but I dout it).

On a related note: how sure are we that the block-laying and the glider-laying type are the only stabilizations for a single switch engine? Clearly there isn't a third natural one, but seeing how the engine itself has period 96, I wonder if smaller stabilizations are engineerable.

[calcyman]

how sure are we that the block-laying and the glider-laying type are the only stabilizations for a single switch engine?

This is technically false, as a switch engine can be supported by a universal constructor following at c/12. However, we haven't ever seen a third orbit arise naturally in any experiment.

[Tropylium]

I'm not sure what you mean. How do you figure universal constructors will be able to travel at c/12?

(Now I can begin to imagine an inherently mobile universal constructor with a body composed of Corderships, but that would be a whole different topic…)

I don't think naturality settles anything. Hands up who's ever seen, say, a natural sidecar?

I suppose all of the *WSS-nosed or B-nosed ships could be argued to simply comprise different stabilizations of a single engine (in which case a sidecar would still be a 2-engine case) — but even then, largish single-nose c/2 ships do exist. Analogous switch engine structures would probably make use of some distinct c/12 cleanup tagalongs.

[calcyman]

How do you figure universal constructors will be able to travel at c/12?

A single switch engine moves in front of the UC, as the UC provides a constant stream of gliders to 'feed' the switch engine, and occasionally sends another glider to produce some random junk near the switch engine. The UC then releases a long salvo of gliders to synthesise itself at the location of the random junk, repeating the process ad infinitum.


Anyway, a similar technique can be used to allow a UC to travel at speeds arbitrarily close to the speed limit, for any rational slope. However, the proof breaks down when 'arbitrarily close to' is replaced with 'at'; I cannot, for example, prove the existence of a (1,2)c/6 knightship.

[Tropylium]

How do you figure universal constructors will be able to travel at c/12?

A single switch engine moves in front of the UC, as the UC provides a constant stream of gliders to 'feed' the switch engine, and occasionally sends another glider to produce some random junk near the switch engine. The UC then releases a long salvo of gliders to synthesise itself at the location of the random junk, repeating the process ad infinitum.

Isn't the speed limit for this procedure not that of spaceships, but that of junk-pushing reactions? The UC needs some initial seed to start construction, just a stream of gliders off to infinity won't do any good.

Now granted, I have no idea what the fastest known such reactions are, but those I've seen seem to be slower than c/12.

…No, wait: sufficiently fast puffers with ignitable output would work too. Remind me, are any of these known to be synthesizable? (Plain puffers probably wouldn't work; there's no way to "push" off them without deleting the entire output up to a point, at which point we do essentially have a fuse.)

[Tropylium]

Err, nevermind the preceding, didn't catch on first reading the part where you said it would be possible to just generate the seeds off the SE itself, so that the UC always stays behind it.

I'm still not sure if c/2 is approachable, however…

But hey, about those switch engines, huh?

[calcyman]

No, wait: sufficiently fast puffers with ignitable output would work too. Remind me, are any of these known to be synthesizable?

Yes, there's a c/2 blinker puffer with output that burns faster than c/2. That's how my hypothetical almost-c/2 spaceship would operate.