If \(S\times T\) is semiperfect, is \(S\) or \(T\) perfect? (Q1592613)

An abelian \(*\)-semigroup \(S\) is said to be semiperfect if any positive definite function on \(S\) admits a representing measure on the dual semigroup of \(S\). If the measure is unique, \(S\) is said to be perfect. The author proves the following two results. Theorem 1. If \(S\) is a nonperfect semigroup, then \(S\) has a nonperfect semiperfect minor which is a subsemigroup of a rational vector space with the identical involution. Theorem 2. Let \(S\) be a discrete subsemigroup of a countable rational vector space \(X\) with the identical involution. If \(S\) is semiperfect, then \(S\) is \(\{0\}\) or isomorphic to \(N_0\) or \(Z\). The author also discusses a route that may lead to an affirmative answer to the following question which was posed by himself earlier: If \(S\) and \(T\) are *-semigroups such that \(S\times T\) is perfect, does it follow that \(S\) or \(T\) is perfect?