If we define a cross product as, a⊗b=ba
consider,
1m⊗a=a1m
1n⊗b=b1n
1k⊗c=c1k
and
c1k=a1m+b1n
where 1m+1n+1k<1
Don't panic, it's only maths.
Are 1m, 1n and 1k confined strictly within the space of a one eighth of a sphere of radius r=2π (surface of sphere not included)? Yes. As required.
This will have finite numbers of triplets, (am,bn,ck) but their values can be high.
Does a, b and c being co-prime make them (their axes) orthogonal here? As illustrated, positive and orthogonal.
{1k,c}, is defined on the plane extended by (1n⊗b) and (1m⊗a). The cross product (1k⊗c) lift {1k,c} into a perpendicular plane. (1m⊗a), (1n⊗b) and (1k⊗c) forms a rectangular axis, right handed.
What happen to the angle in between with cross product?
It is in 1n, etc in actual 1n should be represented by a vector,
2πeiθ, where θ=1nπ2 the arc length 1n divided by the radius, 2π. This radius constant is in all a, b and c and cancels eventually in the equation,
am+bn=ck
Consider, y=eiθ
ln(y)=i1nπ2
−i2ln(y)ln(a)π=1nln(a)
y−i2ln(a)/π=a1n
eiθ−i2ln(a)/π=a1n
eiθa−i2/π=a1n This is wrong.
eiθ=a1n+i2/π
Re[eiθ]=cos(θ)=a1n
This is where cos(θ) for the cross product is hiding.
This means the illustration is consistent. Fermat–Catalan Conjecture, on hold.