Seminormality and polynomial rings (Q734757)

A ring \(A\) has been called seminormal if \(b^{2}=c^{3}\) implies that there exists \(a \in A\) such that \(b=a^{3}\) and \(c=a^{2}\). For commutative rings \(A \subseteq B,\) the seminormal closure of \(A\) in \(B\) is the smallest subring \(A_{1}\) of \(B\) containing \(A\) such that \(x \in B, x^{2} \in A_{1},\) and \(x^{3} \in A_{1}\) implies \(x \in A_{1}.\) Let \(\mathbb{C}\) be a ring and \(f=a_{0}+a_{1}X+ \dots a_{d}X^{d}\) a polynomial in \(\mathbb{C}[X].\) Let \(A\) be the ring generated by the coefficients of \(f^{2}\) and \(f^{3}\). Let the ring generated by the coefficients of \(f\) be denoted by \(B\). An algorithmic proof is given to show that there are finitely many elements \(c_{1}, \dots c_{m} \in B\) such that \(c_{i+1}^{2},c_{i+1}^{3} \in A[c_{1}, \dots c_{i}]\) \((i \in \{1, \dots , m-1 \})\) and \(B=A[c_{1}, \dots c_{m}].\) This implies that if a ring \(A\) is seminormal, then so is \(A[X]\). Combining this result with earlier results by \textit{S. Barhoumi} and \textit{H. Lombardi} [J. Algebra 320, 1531--1542 (2008; Zbl 1153.13019)], together with induction, an algorithmic proof for the statement: ``A reduced ring \(A\) is seminormal if and only if the canonical map \(\mathrm{Pic} ~A \rightarrow \mathrm{Pic} A[X_{1}, \dots X_{k}]\) is an isomorphism'' is obtained. This latter result has been proved before, but constructively [\textit{T. Coquand}, J. Algebra 305, 577--584 (2006; Zbl 1102.13005)].