d means the chain and the CFT for the same class, and one docstring warns about the other's required value #157

The collision

For one and the same universality class — the infinite-randomness fixed point of the
1D random transverse-field Ising chain — two sibling lookups take a keyword called d
and require different values of it:

InfiniteRandomness's own docstring already warns about the second value:

d is the chain's own dimension, not its classical image's. A random
transverse-field Ising chain is d = 1. Atlases hand out d = 2 for its clean
critical point's 2D classical image, so 2 is the number nearest to hand and it is
wrong: φ comes out 3.618 rather than the golden mean 1.618, unflagged.

So the warning names a hazard that the sibling lookup for the same class requires.
A consumer who reads the effective central charge with d = 2 — the only value that
lookup accepts — and carries that d into InfiniteRandomness gets φ = 3.618 with
no error, which is what the docstring says happens.

Why this is a vocabulary issue here

SpatialDimension is this package's quantity, and InfiniteRandomness is where the
constraint 0 < x_m < d and the derived φ = (d − x_m)/ψ live. The two meanings are
both legitimate; what is missing is that they are not the same variable and nothing
says so at the call site.

Options

1. Name them apart. The CFT dimension is d + z at a conformal point and has no
   counterpart at an infinite-randomness one (the docstring already observes there is
   no finite d + z to confuse it with). Giving the CFT lookup its own keyword would
   end the collision at the source.
2. Cross-reference. Keep both, and have InfiniteRandomness's warning name the
   sibling lookup explicitly, so the hazard is greppable from the side that triggers it.

Option 1 is a breaking change in the atlas; option 2 is documentation. Either is a
decision for this package, since it owns the quantity.