Singularities of Rees-like algebras (Q2223521)

If \(X\) is a non-degenerate, irreducible, embedded projective variety over an algebraically closed field \(k\) corresponding to a homogeneous prime ideal \(P \subseteq S = k[x_1,\dots,x_n]\), the Eisenbud-Goto conjecture predicts an upper bound for the regularity of \(X\): \(\hbox{reg}(X) \leq \deg(X) - \hbox{codim}(X) + 1\). This was open for many years, with special cases proved. Thus it was a bit of a shock when Peeva and the second author gave counterexamples to this conjecture, producing irreducible projective varieties with regularity much larger than their degrees. The two main new ideas in their work were so-called Rees-like algebras and step-by-step homogenization. All of the varieties thus produced are singular, and the current paper studies the singularities and their geometry. The authors compute the codimension of the singular locus of a Rees-like algebra over a polynomial ring, and then show that the step-by-step process can decrease the codimension of this singular locus. Thus the authors introduce prime standardization, as an alternative to step-by-step homogenization that preserves the codimension of the singular locus. They then look at the regularity of certain smooth hyperplane sections of Rees-like algebras and show that they all satisfy the Eisenbud-Goto conjecture. They also give a characterization of Rees-like algebras of Cohen-Macaulay ideals, and more generally they characterize when Rees-like algebras are seminormal, weakly normal, and in the case of positive characteristic, F-split.