Having implemented the Lerch transcendent and Riemann zeta, now it's time for the Hurwitz zeta. Technically speaking, the Lerch transcendent is a generalisation of the Hurwitz zeta, so that ζ(s,n)=L(0,n,s)=Φ(1,s,n); However, my implementation of Lerch phi (which is still not as efficient as I'd like...) computes the upper incomplete Gamma function value as a factor in the final result, and when z=1 a!=1 we stumble upon a funny case where the upper incomplete Gamma function has a complex pole /yet/ the Lerch phi is defined at this point (as of course the Hurwitz zeta).
The game plan now is to implement a somewhat general Euler-MacLaurin summation function and derive the formula for the n-th derivative of the Hurwitz zeta function with respect to s (which should obviously be trivial) to speed up the "general" method.
This will have an interesting consequence: We can compute an arbitrary derivative of the Hurwitz zeta at any point we wish, meaning that computing the Glaisher constant defined in terms of the derivative of zeta at some integral point will become attainable.
The pieces of puzzle in SciJava are slowly coming together.
