Relative Riemann-Zariski spaces (Q1758922): Difference between revisions

The author constructs and studies a relative version of the classical Zariski-Riemann space of a field. Let \(f : X \rightarrow Y\) denote a separated morphism of quasi-compact quasi-separated schemes. Consider factorizations of \(f\) into a composition of a dominant morphism \(Y \rightarrow Z\) and a proper morphism \(Z \rightarrow X\). The relative Zariski-Riemann space for \(f\) is defined as the projective limit (in a natural sense) of the spaces of all such factorizations; it is a locally ringed space (If \(X\) is integral and \(Y\) its generic point, this recovers the usual Zariski-Riemann space). The first result is a description of the space (as topological space) in terms of valuations. Next, the author gives an application: every such morphism \(f\) factors as composition of an affine and a proper morphism. In its turn, this implies a new proof of Nagata's compactification theorem (stating that such \(f\) of finite type factors as an open immersion followed by a proper morphism; the proof is very suggestive of generalizing to algebraic spaces or stacks) and the result from Thomason-Trobaugh that such \(f\) factors into an affine morphism followed by a separated morphism of finite presentation.