The structure of projective maps between real projective manifolds (Q1678553)

A real projective manifold \(M\) is a manifold equipped with a real projective structure, which is an open cover \(M=\bigcup_{\alpha}U_{\alpha}\) along with coordinate charts \(\phi_{\alpha}:U_{\alpha}\to\mathbb P(\mathbb R^{d+1})\) such that each transition function \(\phi_{\alpha}\circ\phi_{\beta}^{-1}\) coincides with the restriction of an element in \(\mathrm{PGL}_{d+1}(\mathbb R)\) to \(\phi_{\beta}(U_{\alpha}\cap U_{\beta})\). A projective map \(f:M_1\to M_2\) between two projective manifolds is a map where \(\varphi_{\alpha}\circ f\circ\varphi_{\beta}^{-1}\) is the restriction of a projective map for any coordinate chart \(\varphi_{\beta}\) of \(M_1\) and \(\varphi_{\alpha}\) of \(M_2\). In this paper, the author studies the set of projective maps between compact properly convex real projective manifolds. He shows that this set contains only finitely many distinct homotopy classes and each homotopy class has the structure of a real projective manifold. When the domain is irreducible and the target is strictly convex, the results imply that each non-trivial homotopy class contains at most one projective map.