New Trapedoizal Tail integration method

[Namir]

(05-13-2021 06:17 AM)Paul Dale Wrote:  I'm failing to see any advantage here. Two function evaluations per interval and first order convergence. Simpson's rule has the same number of evaluations with second order convergence.

Pauli

The algorithm is an "open" type and does not calculate values at the beginning or end of the interval. As such it can avoid runtime errors in cases where the bounds of the interval generate runtime error (i.e Simpson rule will bomb!)

Namir