31 Aug
2012
31 Aug
'12
1:45 p.m.
On 8/31/12, W. Edwin Clark <wclark@mail.usf.edu> wrote:
Here's what I got for N > 28:
29 -668472148627227148288 30 5815090244206769209344 31 73063423016960000000000 32 607251019095909989351424 33 -6839365759938434944729088 34 -44960394053009336168349696 35 -573365205894336707192422400 36 -6807508522480635921404264448 37 -54110675162696323264893943808 38 439152787578280613557573779456 39 6321982650569688652881686691840 40 64398429946593839846774588571648
--Clark's exact values agree exactly with my current (integer form) output for N<=31; only disagreements when N>=32. Of course, this exact agreement could have been just by luck in the last few cases.