From: Raymond Toy <raymond.toy@er...>  20081030 13:21:13

Sam Steingold wrote: > Raymond Toy wrote: >> Sam Steingold wrote: >>> (loop :for n :upfrom 3 >>> :for tg = (handlercase (os:tgamma n) >>> (floatingpointoverflow () 'floatingpointoverflow)) >>> :while (and (numberp tg) (= 1 (/ (! (1 n)) tg))) >>> :finally (print (list n tg))) >>> >> (172 FLOATINGPOINTOVERFLOW) >> >> Does that mean tgamma computed 171! correctly to the last bit? For >> every integer from 3 to 171? Nice! > > no, not to the last bit, just up to the relative doublefloatepsilon. That's what I meant. Obviously it can't be the exact value since it's a floating point number and the exact value has way more bits than are available in a float. Ray 