Seminormal rings (following Thierry Coquand) (Q2476016)

This paper provides the details of an elementary proof given by \textit{T.~Coquand} [J. Algebra 305 577--584 (2006; Zbl 1102.13005)] of the Traverso-Swan theorem [\textit{C.~Traverso}, Ann. Sc. Norm. Sup. Pisa 24 585--595 (1970; Zbl 0205.50501); \textit{R.~G.~Swan}, J. Algebra 67 210--229 (1980; Zbl 0473.13001)]. The theorem states that a reduced ring is seminormal if and only if the natural homomorphism \(\mathrm{Pic}(A) \to\mathrm{Pic}(A[X])\) is an isomorphism. The authors use the fact that the category of finitely-generated projective modules over \(A\) is equivalent to the category whose objects are idempotent matrices \(P\), \(Q\) with coefficients in \(A\), and whose morphisms from \(P\) to \(Q\) are given by matrices \(H\) such that \(QH = H = HP\). Their major strategy is to replace localisation at a generic minimal prime \(\mathfrak p\) by a tree computation, which arises from the argument that any element \(a\) of the ring is either inside or outside the generic minimal prime. In the six page annex, the authors provide some additional details. In particular, they discuss zero-dimensional reduced rings and \(pp\)-rings (rings where the annihilator of each element is a prinicipal ideal generated by an idempotent) and their relations to fields and integral rings, respectively.