09/28/2009, 07:29 AM
(09/28/2009, 06:08 AM)mike3 Wrote: OK, that big failure was just because I forgot to change conj(L) to L in the OneIteration routine when running it over the constant-L array... it was still assuming it was tetrationBut when that's done it keeps the array mostly constant (more or less up to the error of the Simpson rule), so I'm still perplexed.
If it keeps the constant L, good!
Could you plot please, the deviation the function at the first iteration from the constant?
Could you plot it along the imaginary axis from -iA-i to iA+i ?
We'll see, where the approximation is worse, at the center or at the end.
If it keeps at least 4 digits, you may run the second teat I suggested:
for the precise initial, plot the initial and the first iteration in the range -iA-i to iA+i .