FYI, Eugene McDonnell contributed an article in Vector McDonnell, Eugene, At Play with J: Blists in OLEIS, Vol 17.1, 110 and a technical correspondence in Vol 17.2, 112 It shows Bell number is the row sum of Stirling subset numbers.
Joseph De Kerf discussed calculation of Bell number using floating point in APL, De Kerf, Joseph, Bell Numbers and APL, Vol 17.1, 131 -- regards, bill ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
