Stellar stratifications on classifying spaces (Q2312917)

The paper under review studies stratifications on classifying spaces of acyclic topological categories. Using the notation of the authors, let \(C\) be a category, \(C_0\) the objects of \(C\), \(C_1\) the morphisms, and \(C(x,y)\) the morphisms from \(x\) to \(y\). A topological category is called acyclic if 1. For any pair of distinct objects \(x,y\) of \(C\), either \(C(x,y)\) or \(C(y,x)\) is empty. 2. For any object \(x\) of \(C\), \(C(x,x)\) consists only of the identity morphism. 3. Viewing \(C_0\) as the subspace of identity morphisms in \(C_1\), we have \(C_1=C_0\sqcup (C_1\setminus C_0)\). To gain an idea of the uses and applicability of the result contained in the paper, the introduction contains several helpful questions that the contents of the paper answers. One of the main objects the authors use are cylindrically normal stellar stratified spaces (CNSSS), a generalization of a CW complex. The main result is Theorem 1.5. Given an acyclic topological category \(C\) along with some finiteness conditions, ``there exists a CNSSS structure on the classifying space \(BC\) whose face category is isomorphic to \(C\) as topological categories.'' Several corollaries of this main result follow. One result that I found particularity intriguing, Theorem 1.8, concerns discrete Morse theory. In a recently published paper [\textit{V. Nanda} et al., Adv. Math. 340, 723--790 (2018; Zbl 1405.57034)], it was shown that the homotopy type of a finite regular CW complex \(X\) can be reconstructed from a discrete Morse function \(f\) on \(X\). This result involves constructing a certain category, denoted \(C(f)\), based on the discrete Morse data and then taking the classifying space \(B^2\) of this category. As impressive as this result is, the classifying space can actually be just as complicated as the original \(X\), at least in terms of the number of cells. Theorem 1.8 yields a much nicer cell structure on the classifying space and consequently, on \(X\). A fairly simple corollary of Theorem 1.5, in part the theorem states Theorem 1.5. Let \(f\colon X \to \mathbb{R}\) be a discrete Morse function. Then there exists a CNSSS structure on the classifying space \(B^2C(f)\) such that strata are indexed by the set of critical cells of \(f\). For the entire collection see [Zbl 1411.55001].