Complete intersections and mod \(p\) cochains (Q1938757)

The authors state their philosophy: ``The idea is to formulate conditions in commutative algebra so that they are homotopy invariant and then to say that a space \(X\) has property \(P\) if the commutative ring spectrum \(C^*(X;k)\) has property \(P\).'' The paper does this with the complete intersection (ci) condition, focusing mainly on the case where \(X\) is the classifying space \(BG\) of a finite group \(G\), and \(k\) is a field of characteristic \(p\). Several versions of the complete intersection condition are formulated, and implications made between them. Let \(R\) be a commutative Noetherian local ring. Say that \(R\) is \textit{sci} if it is the quotient of a regular local ring by a regular sequence. Say that \(R\) is \textit{gci} if \(\mathrm{Ext}_R^*(k,k)\) has a resolution with polynomial growth, where \(k\) is the residue field of \(R\). Say that \(R\) is \textit{mci} if any finitely generated \(R\)-module has an eventually periodic resolution. \textit{D. J. Benson} and \textit{J. P. C. Greenlees} [``Complete intersections and derived categories, (2009), \url{arXiv: 0906.4025}], constructs a homotopy invariant version of the mci condition, the \textit{zci} condition. For classical commutative local rings, the sci, gci, and zci conditions are equivalent [\textit{T. H. Gulliksen}, Math. Scand. 47, 5--20 (1980; Zbl 0458.13010)]. To formulate these conditions for spaces, one needs a commutative model for the cochains, and this necessitates the use of commutative ring spectra. Thus here \(R = C^*(X;k) := \mathrm{map}(X,Hk)\) is the function spectrum from \(X\) to the Eilenberg-MacLane spectrum \(Hk\). The first six sections of the paper under review are devoted to this formulating, which is too technical to detail here. By Section 7, we have notions of regular (a space is said to be \textit{\(g\)-regular}), Noetherian (a \textit{normalisable} space), sci, gci, and eci (another version of the mci condition for rings). Sections 7 -- 12 are then mainly devoted to showing, for a normalisable space, and working over \(k=\mathbb{F}_p\), that sci implies gci, which is equivalent to eci. Section 13 gives examples. Similar work was recently done in [\textit{J. P. C. Greenlees, K. Hess} and \textit{S. Shamir}, J. Pure Appl. Algebra 217, No. 4, 636--663 (2013)] for the rational homotopy theory of simply connected spaces; in this setting one has a strictly commutative DGA model for the cochains coming from the simplicial de Rham complex. Under a Noetherian hypothesis (that \(H^*(X;\mathbb{Q})\) is Noetherian), they show that zci implies sci implies gci, and under a stronger Noetherian hypothesis gci implies sci.