Hi @StrivnX,
One example of when this sort of thing will come in handy is for grid-based pathfinding.
To move one space diagonally, you have two options;
Move along 1 cell and up 1 cell (for a distance of 2 units)
Move diagonally 1 cell (for a distance of √2 units)
As expected, diagonal movement has a shorter distance (√2 < 2) and a smaller number of cells traversed (1 < 2).
However, if you only work by the number of cells traversed (1 instead of √2) then you may start to encounter problems when you start introducing things like variable movement costs, such as roads increasing movement speed or mud decreasing it.