Talk:Hardy-Littlewood Circle Method

the assumption that a positive radius of convergence exists for an arbitrary generating function is invalid. a bound on the number of partitions of n is required. one could argue that P(n) < S(n)=n! since P(n) numbers the conjugacy classes of the symmetric group. also a simple argument requiring s to tend to zero should be inserted to deal with the O(s) term.