[vtkusers] dividing cubes algorithm

[Tim Fooy]

Hi all,

I have a question about the dividing cubes algorithm, as described in The 
Visualisation Toolkit book, p. 303.
The book states (p. 304) that the recursive variant of the algorithm possibly 
generates more pixels than the procedural one.  Isn't that wrong?  Both 
variants divide voxels to such a depth that a solid surface can be rendered, 
depending on the screen resolution.  So, in the end, the size of the 
subvoxels they create is equal.  But, the recursive variant does not go to 
this depth everywhere in the original voxel, only in the part where the 
surface passes through.  So as i see it, i am inclined to think that the 
recursive variant produces less points.

Can somebody point out where my thinking goes wrong?

Thanks in advance,

Tim