Date: Wed, 31 Dec 80 12:10:00 -0500 (EST)
From: David C. Plummer <DCP@MIT-MC
>
Subject: the 5x5x5 [133 lines]
OK, folks, I'm considering going further than 4x4x4 and entering
the realm of the 5x5x5.
Cubies:
C := Corner
X := aXis (center)
E := Edge (outside center)
L := Left (external edge)
R := Right (external edge)
D := Diagonal (internal [to the face] corner)
A := Adjacent (to the center, thanks to WER)
(internal [to the face] edge)
A 3-D view would look like this
z=-5 +---+---+---+---+---+
/ / / / / /|
/ C / L / E / R / C / |
+---+---+---+---+---+ |
/ / / / / /|C +
/ R / D / A / D / L / | /|
+---+---+---+---+---+ |/ |
z=0 / / / / / /|R + |
/ E / A / X / A / E / | /|L +
+---+---+---+---+---+ |/ | /|
/ / / / / /|E + |/ |
/ L / D / A / D / R / | /|D + |
+---+---+---+---+---+ |/ | /|E +
/ / / / / /|L + |/ | /|
/ C / R / E / D / C / | /|A + |/ |
y,z=5 +---+---+---+---+---+ |/ | /|A + |
| | | | | |C + |/ | /|R +
| C | L | E | R | C | /|D + |/ | /|
| | | | | |/ | /|X + |/ |
y=3 +---+---+---+---+---+ |/ | /|D + |
| | | | | |R + |/ | /|C +
| R | D | A | D | L | /|A + |/ | /
| | | | | |/ | /|A + |/
y=1 +---+---+---+---+---+ |/ | /|L +
| | | | | |E + |/ | /
y=0 | E | A | X | A | E | /|D + |/
| | | | | |/ | /|E +
y=-1 +---+---+---+---+---+ |/ | /
| | | | | |L + |/
| L | D | A | D | R | /|R +
| | | | | |/ | /
y=-3 +---+---+---+---+---+ |/
| | | | | |C +
| C | R | E | L | C | /
| | | | | |/
y=-5 +---+---+---+---+---+
x=-5 -3 -1 1 3 5
LOVE THAT ASPECT RATIO !!!!
All in all there are 6 aXis faces 8 Corners 12 Edges 24 Left/Right type edges 24 Diagonals 24 Adjacents -- 98 = 5^3 - 3^3 = 125-27 visible cubies
Computation (inaccurate, but within a couple orders of magnitude)
of the number of reachable positions:
Axes: lets not hack the extended problem yet -> 1 Corners:8 of them anywhere -> 8! each can take 3 orientations -> 3^8 parity of the corner -> 1/3 Edges: 12 of them anywhere -> 12! each can take 2 orientations -> 2^12 position/orientation restriction -> 1/4 L/R: 24 of them anywhere -> 24! orientation defined (they cannot flip) -> 1 parity (cannot swap only two) -> 1/2 (I think) Adjac: 24 of them anywhere: -> 24! one edge always touches a face center -> 1 parity -> 1/2 (at least) Diags: 24 of them anywhere -> 24! one corner always touches a face center -> 1 parity -> 1/2 (at least) It may not be accurate, but this computation gives 1.291318 * 10^90
A slice through the center (z=0) probably looks something like
y=5\
/ ..XXXXXXXXXX++++++++++EEEEEEEEEE
..XXXXXXXXXX++++++++++EEEEEEEEEE
.....XXXX++++++++EEEEEEEEEEEEEEE
.....XXXX++++++++EEEEEEEEEEEEEEE X is an axis cubie
y=4\ .....XXXX++++++++EEEEEEEEEEEEEEE E is an edge cubie
/ .....XXXX++++++++EEEEEEEEEEEEEEE + is one adjacent cubie
.....XXXX++++++++EEEEEEEEEEEEEEE ~ is another adjacent
.....XXXX++++++++EEEEEEEEEEEEEEE
.....XXXX++++++++EEEEEEEEEEEEEEE
y=3\ .....XXXX++++++++EEEEEEEEEEEEEEE
/ .....XXXX++++++++EEEEEEEEEEEEE~~
.....XXXX++++++++EEEEEEEEEEEEE~~
.....XXXX++++++++EEEEEEEEEEEEE~~
.....XXXX++++++++EEEEEEEEEEEEE~~
y=2\ .....XXXX++++++++EEEEEEEEEEEEE~~
/ .....XXXX+++++++/~~~~~~~~~~~~~~~
.....XXXX++++++/~~~~~~~~~~~~~~~~
.....XXXX+++++/~~~~~~~~~~~~~~~~~
\....XXXX++++/~~~~~~~~~~~~~~~~~~
y=1\ .\...XXXX+++/~~~~~~~~~~~~~~~~~~~
/ ..\..XXXX++/~~~~~~~~~~~~~~~~~~XX
...\.XXXX+/~~~~~~~~~~~~~~~~~~~XX
....\XXXX/~~~~~~~~~~~~~~~~~~~~XX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
y=0\ XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
/ XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
..../XXXX\....................XX
.../.XXXX.\...................XX
y=-1\ ../..XXXX..\..................XX
/
/\ /\ /\ /\ /\
x=-1 0 1 3 5
This time the central axis is rigid in the sense that it does
form a cross, but each of the six spokes can rotate as in the
3x3x3 cube. The curvature and tolerances of some of the pieces
gets a little hairy, but I'm working with graph paper and looking
at the other slices through the cube. Wish me luck -- I have
thoughts of construction.