This code snippet just won't compile in a conforming compiler:

Code:
template<> struct series<0.0, 0, 0>
because of the floating point template parameter....
If this "specialization" is useful, how come MS would have discarded it?
Thanks
Jack