?

This is another generative experiment made with Nodebox based on the studies of Bruno Munari (http://en.wikipedia.org/wiki/Bruno_Munari) on how elements of a texture are perceived as a shape when they're close together and instead the areas where these same points are thinned are read as background.
How close the points should be before we can see a question mark? Is it different between different person? I keep wondering that... :)