The implementation of the author's algorithms for the manipulation of power series [ibid. 15, 27-36 (1986; reviewed above)] is here discussed in detail. For the efficiency important computer science aspects, such as the organization of loops and the storage of the coefficients are treated carefully. The resulting routines, which are expressed in a form that can be translated easily into a higher computer language, are remarkably compact and general. They can use weighted truncation with an arbitrary weight function, and they can efficiently manipulate power series in an arbitrary number of variables.