No label defined (Q3977566)

[For the entire collection see Zbl 0688.00012.] The ideal boundary \(M(\infty)\) of a finitely connected, oriented, complete open Riemannian 2-manifold \(M\) admitting total (Gaussian) curvature \(c(M)\) is the metric space associated to the following pseudometric \(d_{\infty}\) defined in the set of all rays in \(M\). For rays \(\sigma\), \(\gamma\) in \(M\) let \(D\subseteq M\) be a half plane with boundary \(\partial M\approx\mathbb{R}\) coinciding with \(\sigma\) for large positive \(t\in\mathbb{R}\) and with \(\gamma\) for large negative \(t\in \mathbb{R}\). Set \(L(\sigma, \gamma) =\pi-k(D)-c(D)\) and \(d_ \infty (\sigma, \gamma)= \hbox{min} \{L(\sigma, \gamma), L(\gamma,\sigma)\}\) where \(k(D)\) is the geodesic curvature integral along \(\partial D\). Theorem A says that if \(c(M)=-\infty\) then \(M(\infty)\) is isometric to a disjoint union of closed connected subsets of \(\mathbb{R}\). \(M\) is called a visibility surface if any two different points in \(M(\infty)\) can be joined with a straight line in \(M\). Theorem B expresses visibility in terms of \(M(\infty)\) (as a metric space) and of the Busemann functions.