[qawkwpvv] two grids of mazes

on the left, a PxQ grid of MxN mazes.  on the right, a MxN grid of PxQ mazes.  (M=N=P=Q is of course the simplest.)

on each grid, one maze is highlighted as having focus.  on the two focused mazes, a dot indicates the current positions within each maze.  if the left dot is at position (x,y) within its focused maze, then the maze at position (x,y) in the right-side grid of mazes is focused.  similarly right dot and left focus.

the left controller, perhaps the WASD keys, moves the left dot, obeying the constraints of walls in the left focused maze.  it also correspondingly moves the focus on the right side grid of mazes.  when the right-side focus moves, the dot on the right side moves to the newly focused maze, keeping the same relative location.

as a guide, in addition to the dot in the focused maze, there is a faded dot in all the other mazes on that side which moves in sync with the focused dot, indicating where the focused dot would land if that maze receives focus.

symmetrically, the right controller, perhaps the arrow keys, moves the right dot (within the constraints of the focused right maze) and the left focus.

by manipulating focus away then back, it is possible for a dot to seemingly pass through walls.  maybe it feels like going up, over, then back down on the other side of the wall.

I think (some) standard maze generation algorithms such as Wilson's algorithm will just work to create a spanning tree over all states.  but we probably don't care about all states, just all possible positions of the dots.

it's a limited form of a 4D maze.  future post lrjzlhie: general 4D maze.

simpler variation: all mazes are independently solvable.  the global maze will no longer be a spanning tree.  let the goal be to find short paths.

simpler variation: all mazes on one side are the same.  I suspect standard maze generation algorithms no longer just work.

this can probably be extended beyond 2 grids of mazes.  each maze controls the focus of the next grid in a cycle.

3D (actually 6D) possible: two first person views, split screen.  without the overhead views of 2D, it will likely be extremely difficult to navigate.  x-ray vision might help.