Life: glider gun origin
Dean Hickerson
and sc...@tekcrl.LABS.TEK.COM gave a plausible explanation. I sent Scott's
message to Bill for his comments. In the following, lines beginning with
">>" are from Scott's posting, those beginning with "> " are Bill's replies.
>> I'll venture a reasoned speculation that the glider gun was discovered
>> in 3 steps.
>>
>> First, the queen bee was found (that's the arrowhead-shaped moving
>> part, of which 2 are in the standard glider gun). The queen bee
>> arises fairly often in random fields and is pretty easy to spot,
>> so it's a reasonable guess she was discovered in random fields.
>
> Observed, but not captured. We thought (correctly) that we saw it during
> the r pentomino (gen 700+), but the software was so slow and cumbersome,
> and our confidence in what we saw so low, that we didn't rerun the r to
> see what was really happening. (The bee arises phenomenally often during
> rabbits, but I wonder if we could have run rabbits, even if we knew it.)
Rabbits is this 9 bit object, discovered by Andrew Trevorrow in 1986:
o...ooo
ooo..o.
.o.....
It runs for 17331 generations before stabilizing as 136 blinkers,
109 blocks, 65 beehives, 18 loaves, 18 boats, 7 ships, 4 tubs, 3 ponds,
2 toads, and 40 gliders.
Gosper continues:
> Then one day, an analytical chemist friend of ours wandered in, and
> diddling around, typed two rows 15 dots vertically separated by an
> intervening row of blanks. This makes two shuttles which separate,
> lay eggs, collide elastically in the middle, and then succumb to
> dystocia. When I saw the "shuttle" thus captured, I said I bet I
> would find a gun in the next couple of days.
>
>> Second, shuttles were constructed from the queen bee. (Shuttles have
>> one or more parts, such as queen bees, moving in a back and forth
>> periodic motion). The queen bee in isolation isn't periodic, because
>> she leaves some debris when she turns around and is destroyed by
>> the debris on her next pass. But there are many queen bee configurations
>> in which the debris is neutralized, including placement of a block or
>> eater near the bee's turnaround point, or placing two queen bees
>> in a line or at right angles in various positions and phases.
>
> Right on. Since this predated the phenomenon of eating, I never imagined
> that something as simple as a block would kill the eggs, and immediately
> went after a 90 degree collision, so as to close the disposal problem.
> This proved easy, and the first period 30 oscillator had eight bees!
> (We called them shuttles, even before they were.)
> Schroeppel, I think, observed that it could be reduced to four, but I
> prepared to exhaustively search all the x,y,t dislocations of the
> headon and perpendicular collisions of the eight bee configuration.
> At this point, the PDP-1 hackers came up with the block mechanism,
> and I decided that, for a factor of 4, I'd do the headons and
> perpendiculars separately.
>
>> In most queen bee shuttles, turnaround debris neutralizes by
>> evaporationg. But in one position and phase of a 2-bee shuttle (the
>> standard glider gun), the debris evolves into a glider and sails away
>> before the bees return. It's a good guess the standard gun was found
>> while methodically investigating 2-bee shuttle configurations.
>> --
>> Scott Huddleston
>
> Exactly.
I also asked Gosper how thoroughly he explored the combinations of 2 or 3
shuttles.
> I had grimly resolved to look at the threes (and combos with
> pentadecathlons) if the twos had failed. Luckily, I did the headons
> first. (Escape looked likelier.) I won't swear the gun is unique, but I
> think it is. I looked at fairly many perpendiculars while seeking the
> ovolysis, and the winning geometry is fairly near the end of the plausible
> search space. Again, re perpendicular bees, gliders move so slowly that
> only one of the four escape directions is likely to work.
>
> Actually, we cheered prematurely more than once, as a number of
> headon geometries made prompt gliders, only to exit toward the wrong
> shuttle.
>
> I don't believe I ever came close to exhausting the p46 shuttle collisions.
> For all I know, some of them could make spaceships. At least one makes
> two gliders/whack.