Survey of Oka theory (Q554263): Difference between revisions

The paper under review gives an account of the current state of what is now called Oka theory. The beginning actually dates back to K.~Oka's results on the second Cousin problem over domains of holomorphy in 1939. They imply that continuous sections of holomorphic fiber bundles with fibre~\(\mathbb{C}^*\) over such domains can be deformed to holomorphic sections. They were later extended by H.~Grauert to more general fiber bundles over arbitrary Stein spaces. The subject then took a new start in 1989, when M.~Gromov considered elliptic bundles and put the emphasis on the homotopy-theoretic aspect. The story goes on today: recently, the work of F.~Forstnerič gave birth to the notion of a complex Oka manifold, with several characterizations. After an introductory section, the authors write a short review of the theory of Stein manifolds in Section 2, recalling, in particular, four equivalent definitions. One of them, Stein's original definition actually, makes clear that Stein manifolds carry many holomorphic functions, i.e. maps to \(\mathbb{C}\). Oka manifolds are sort of dual: they have many holomorphic maps from~\(\mathbb{C}\). To make this statement precise, the authors introduce in Section 3 several properties, like the convex approximation property, the convex interpolation property, or parametric versions of them, to tell if holomorphic maps from convex compact domains of~\(\mathbb{C}^n\) or from submanifold of~\(\mathbb{C}^n\) can be uniformly approximated by global holomorphic maps from~\(\mathbb{C}^n\) or even extended to such global maps. All the different properties turn out to be equivalent: they define Oka manifolds. Section 3 ends with a list of explicit examples of Oka manifolds. The following section is devoted to a proof of the fact that the convex approximation property implies the others. In Section 5, M. Gromov's notion of an elliptic manifold is introduced: it is a manifold \(X\) that admits a dominating spray, i.e. a holomorphic map \(s: E \to X\) from the total space of a vector bundle such that the restriction to any fiber induces a submersion at 0. Such a manifold is Oka. The authors give several examples as well as applications, for example to Forster's conjecture: any connected Stein manifold of dimension \(n\geq 2\) admits a proper embedding to \(\mathbb{C}^m\), where \(m=[3n/2]+1\). Section 6 deals with a relative situation: the definition of being Oka is extended to maps of complex manifolds. Then Section 7 quickly mentions a homotopy-theoretic viewpoint: the category of complex manifolds and holomorphic maps may be embedded in a model category (in the sense of D. Quillen), where Stein manifolds, Oka maps, etc. find a natural interpretation. In the last section, the authors give a list of open problems.