\(p\)-integral harmonic sums (Q1182862): Difference between revisions

Let \(H(n)=\sum_{k=1}^ n{1\over k}\) (\(n\geq 1\)) be the harmonic sums. It is a classical result of J. Kürschák that \(H(n)\) is never an integer. In order to generalize this fact, the authors write \(H(n)=a(n)/b(n)\), with \((a(n),b(n))=1\) for \(n\geq 0\), where \(a(0)=0\), \(b(0)=1\). Clearly \(a(n)\not\equiv 0\pmod 2\), \(b(n)\equiv 0 \pmod 2\) for \(n>1\). For a prime \(p\), let \(J(p)=\{n\geq 0\): \(a(n)\equiv 0 \pmod p\}\). The authors prove that \(J(p)\) is finite for \(p\leq 7\) and conjecture that this is the case for all primes. However they cannot show even that this is true for \(p=11\). A prime \(p\) is called harmonic whenever \(J(p)=\{0,p- 1,p(p-1),p^ 2-1\}\). The harmonic primes below 200 are determined, and it is conjectured that the set of these primes is infinite.