A calculus for flow categories (Q2089645)

\textit{Framed flow categories} are combinatorial models of topological spectra, first introduced in [\textit{R. L. Cohen} et al., Prog. Math. 133, 297--325 (1995; Zbl 0843.58019)]. One of the succesful applications of these models is the \textit{Khovanov spectrum}, a link invariant with values in spectra discovered by \textit{R. Lipshitz} and \textit{S. Sarkar} [J. Am. Math. Soc. 27, No. 4, 983--1042 (2014; Zbl 1345.57014)]. These link invariants have the particularity of having small cohomological width (their non-trivial cohomology is concentrated in a few consecutive degrees) relative to the number of crossings of their link diagrams. The first purpose of this article is to give a combinatorial viewpoint on the question of whether two framed flow categories yield homotopy equivalent spectra. The authors introduce four ``moves'' between framed flow categories: perturbation, stabilization, handle cancellation and extended Whitney trick. Any two framed flow categories related by any of these moves yield weak equivalent spectra, and they conjecture that the converse holds: Any two weak equivalent framed flow categories are related by a sequence of moves. The main theorem of the article proves this conjecture when the cohomology of the framed flow categories is supported in four consecutive degrees. On the other hand, homotopy types of topological spectra whose cohomology is supported in four consecutive degrees can be classified using the work of \textit{H.-J. Baues} and \textit{M. Hennes} [Topology 30, No. 3, 373--408 (1991; Zbl 0735.55002)]. The authors give minimal framed flow categories modeling these homotopy types and construct an algorithm that turns any framed flow category of cohomological length four into a minimal one. Remarkably, this algorithm gives an independent proof of the classification of spectra of cohomological length three. If the framed flow category has bigger cohomological length, the algorithm can be still used to give effective computations of the first three Steenrod squares. This is the fourth paper in a series of articles exploring combinatorial models of weak equivalences (see [\textit{D. Jones} et al., Indiana Univ. Math. J. 66, No. 5, 1603--1657 (2017; Zbl 1394.57012); \textit{A. Lobb} et al., Algebr. Geom. Topol. 18, No. 5, 2821--2858 (2018; Zbl 1418.57023); Exp. Math. 29, No. 4, 475--500 (2020; Zbl 1464.57020)]).