Openness of the regular locus and generators for module categories (Q2000804)

Let \(R\) be a commutative Noetherian ring, the regular locus \(\operatorname{Reg}R\) be the subset of \(\mathrm{Spec }R\) consisting of prime ideals \(\mathfrak{p}\) such that \(R_{\mathfrak{p}}\) is regular. A complex in the bounded derived category \(D^b\pmod{R}\) is called a generator if each complex in \(D^b\pmod{R}\) can be build from it by finite direct sums, direct summands, and exact triangles. It is known that such a generator exists in \(D^b\pmod{R}\) implies that \(\operatorname{Reg}R\) is an open subset of \(\operatorname{Spec}R\). The authors of the paper under review conjecture that the converse does not hold. They proved a weak version: \(\operatorname{Reg}(R/\mathfrak{p})\) contains a nonempty open subset for each \(\mathfrak{p}\in\operatorname{Spec}R\) if and only if the derived category \(D^b\pmod{R/\mathfrak{p}}\) has a generator for each \(\mathfrak{p}\in\operatorname{Spec}R\). In additional, they also consider the relations between the openness of \(\operatorname{Reg}R\) and the existence of generators for the category of finitely generated \(R\)-modules and the singularity category of \(R\).