Personally, I wouldnt bother - you'll only confuse yourself !
stick to 1 or 2 dimensional arrays where the index maths is much easier - if you need more dimensions, have several 1 or 2 d arrays of the same size:
ie:
int x[1000], y[1000], z[1000];
for 3d objects, etc you could use this:
typedef coord struct {int x, y, z };
coord array[1000];
in effect an array of structures.
the coords can then be accessed like this :
px = array[123].x;
etc. or
int curr_ind = 123;
ptr_curr_x = array + curr_ind * sizeof(coord);
ptr_curr_y = array + curr_ind * sizeof(coord) + sizeof(int);
ptr_curr_z = array + curr_ind * sizeof(coord) + 2 * sizeof(int);
This last approach can quite neatly be expanded for any number of dimensions.
why do you need lots of dimensions ? complicated Tensor Analysis ? :-D
