Existence of rational points on smooth projective varieties (Q1024257)

It is proved (Theorem 1.1) that if, for a fixed number field \(k\), there exists an algorithm for deciding whether a regular projective geometrically integral \(k\)-variety has a \(k\)-point, then there is an algorithm for deciding whether an arbitrary \(k\)-variety has a \(k\)-point and also an algorithm for computing \(X(k)\) for any \(k\)-variety \(X\) for which \(X(k)\) is finite. The main ingredient in the proof is the following fact (Theorem 1.3): Let \(k\) be a number field. Let \(X\) be a projective variety. Let \(U\subseteq X\) be an open subvariety. Then there exist a regular projective variety \(Y\) and a morphism \(\pi: Y\to X\) such that \(\pi (Y(k))=U(k)\). Moreover, there exists an algorithm for constructing \((Y,\pi )\) given \((k,X,U)\). The key special case of Theorem 1.3 (\(X=\mathbb P^1\), \(U=\mathbb A^1\)) gives rise to a striking example of a family of smooth projective geometrically integral varieties in which every rational fibre but one has a rational point. The proof of Theorem 1.3 is involved enough. It starts with a generalization of Iskovskikh's example of a conic bundle surface over \(\mathbb Q\) violating the Hasse principle to the case of an arbitrary number field \(k\) (which is of independent interest). This gives rise to a one-parameter algebraic family of Châtelet surfaces such that exactly one of the fibres fails to have a \(k\)-point. The author heavily relies on ideas and results of the seminal paper by \textit{J.-L.~Colliot-Thélène, J.-J.~Sansuc} and \textit{Sir P.~Swinnerton-Dyer} [J. Reine Angew. Math. 373, 37--107 (1987; Zbl 0622.14029); J. Reine Angew. Math. 374, 72--168 (1987; Zbl 0622.14030)]. He also uses Faltings' theorem to prove the ``noneffective'' part of the theorem and then applies Parshin's argument [\textit{L.~Szpiro}, Astérisque 127. Publié avec le concours du Centre National de la Recherche Scientifique. Paris: Société Mathématique de France. 275--287 (1985; Zbl 0588.14028)] to make the proof effective. Some eventual generalizations to the case of global fields of positive characteristic and related open problems are also discussed.