On a question of Kollár (Q627460)

The action of a group \(G\) on a topological space \(X\) is ``cocompact'' if there exists a compact subset \(K \subset X\) such that \(G \cdot K=X\). The author shows that, if \(X\) is a locally compact metric topological space and \(G\) acts by isometries, then \(X\) is a complete metric space and for every isometric embedding \(i : X \hookrightarrow Y\) of \(X\) into a metric topological space \(Y\), the image \(i(X)\) is closed in \(Y\). Among many interesting corollaries of this result, one gets the following classic result: A homogeneous bounded domain in a Stein manifold is itself Stein. On the other hand, if \(D\) is a connected complex space and \(\Gamma\) is a group of automorphisms of \(D\) acting cocompactly on \(D\), if \(Z\) is a non-empty \(\Gamma\)-invariant subset of \(D\), then every bounded holomorphic function on \(D\) which vanishes on \(Z\) must be identically zero. As a consequence, if \(D\) is a complex space which can be embedded in a reduced Stein space \(X\) as a relatively compact connected open subset and if there exists a group of automorphisms of \(D\) acting properly discontinuously on \(D\) with compact quotient, then \(D\) must be smooth. This constitutes a negative answer to the question whether there exist compact spaces with a bounded singular domain as universal covering.