On 2001-5-15 22:40, Anton Sherwood wrote:
> ... Meanwhile, another idea: if (dx,dy,dz) were replaced with
> (dx+dy,dx+dz,dy+dz), the `crystal' would be a rhombic dodecahedron.

Or not.  Less than 13 years later, as I was thinking about something 
unrelated to ray-tracing, it hit me that it would have to be
	(dx+dy, dx-dy, dx+dz, dx-dz, dy+dz, dy-dz).

-- 
*\\* Anton Sherwood *\\* www.bendwavy.org

Post a reply to this message