Regarding the coordinate-dropping approach, have you already seen the technical report of Voelker et al.? As for sampling over a ball, it would be interesting to compare with the approach of Barthe et al., which uses a (multi)normal distribution and an exponential distribution together, and the approach of Harman and Lacko, which takes a decompositional approach. (And of course, the normal variate approach is a Marsaglia original.)
Thanks for these great references!
-- you have earned Featured Contributor Badge Your exceptional post has been selected for our editorial column Staff Picks http://wolfr.am/StaffPicks and Your Profile is now distinguished by a Featured Contributor Badge and is displayed on the Featured Contributor Board. Thank you!
