Semistrict models of connected 3-types and Tamsamani's weak 3-groupoids (Q995633)

One of the goals of algebraic topology is an equivalence of homotopy categories, one of which is a category of topological spaces and the other an algebraic category whose description alone explains the complexity of classification problems up to homotopy. An \(n\)-type is a path-connected topological space \(X\) for which \(\pi_i(X) = \{0\}\) for \(i > n\). If we denote by \(\text{ Top}_*^{(n)}\) the category of \(n\)-types, then Loday introduced \(\text{ cat}^n\)-group theory, generalizing the notion of crossed modules of Whitehead (\(n=2\)). The author investigates \textit{Z. Tamsamani}'s model [\(K\)-Theory 16, No. 1, 51--99 (1999; Zbl 0934.18008)] for homotopy \(n\)-types which is given by weak \(n\)-groupoids. In the case \(n=3\), two natural questions are: (1) how do the models of Tamsamani and Loday compare? and (2) can one generalize from path-connected spaces to connected spaces? The author's solution is a subcategory of weak 3-groupoids which are not as weak as in Tamsamani's case. The key notion is that Segal maps in one direction are isomorphisms rather than equivalences. An analysis of bisimplicial sets (from the \(\text{ cat}^2\)-groups) underlies the main methods together with a construction called discretization. A comparison for 3-types to Gray groupoids with one object is also carried out.