Closed ExpHP closed 5 years ago
Still low priority, just wanted to mention that the best solution is likely to interpret normal
as a Miller index. (i.e., it is the normal of the plane in units of the reciprocal vectors, not the real lattice vectors)
Triage: this was added already O_o
the algorithm I originally wrote in
assign_layers
is only correct when the desired normal vector is both parallel to one lattice vector, and perpendicular to the other two.this is obviously insufficient for general lattices when taking into consideration equivalence of the lattice under unimodular transformations. There may also be other limitations that can be lifted if one is clever enough. I don't know.
not a terrible issue at the moment because it checks for bad cases and throws errors.