Universal formulas for the number of partitions (Q1787168)

Explicit formulas for the partition function, attempting to improve on Euler's pentagonal number recurrence, have been of interest since Ramanujan. Notable contributions are the astonishing differential Hardy-Ramanujan-Rademacher formula involving complex numbers, the trace formula of \textit{K. Bringmann} and \textit{K. Ono} [Proc. Am. Math. Soc. 135, No. 11, 3507--3514 (2007; Zbl 1156.11040)] that involves algebraic integers, and the combinatorial formula of \textit{Y. Choliy} and \textit{A. V. Sills} [Ann. Comb. 20, No. 2, 301--316 (2016; Zbl 1353.05013)] that counts partitions by Durfee squares. In this paper the author gives explicitly a number of quasipolynomials (which are referred to in the paper as fractal polynomials) giving the values of \(p(n,m)\), the number of partitions of \(n\) into exactly \(m\) parts, for \(m \leq 7\). Researchers following the citations to the \textit{A. V. Sills} and \textit{D. Zeilberger} paper in the references [Adv. Appl. Math. 48, No. 5, 640--645 (2012; Zbl 1242.05020)] will find mentions of sources with formulas up to \(m=15\) or more. Other theorems are ``counting'' formulas in only the most technical sense: Theorem 3, for instance, is simply indexed by nearly every partition of \(n\), collecting as terms of the summation only those that differ in just their first two parts. That these quasipolynomials are of period \(\operatorname{lcm}(1,\dots,m)\) is known; the author seems to be unfamiliar with the relevant work of Kronholm and Munagi on \(p(n,m)\). The hope is expressed that a general form can be found for the coefficients of the polynomials, which by summation would lead to a much shorter formula for \(p(n)\).