The set that POV-Ray uses currently doesn't look like that either. (see my
post, pbi "Re: >1600 radiosity samples: right costheta distribution?"). "My"
distribution has very similar properties to that of POV-Ray, except for the
ordering that I still have to fix.