Local-global principles for homogeneous spaces over some two-dimensional geometric global fields (Q2054218)

Let $F$ be a number field. A class of $F$-varieties $F$ is said to satisfy the local-global principle if any variety $Z\in F$ that has points in every completion of $F$ has in fact a rational point over $F$. The Hasse-Minkowski Theorem says the most classical example of such a class, will be given by the class of quadrics over $F$. Some varieties fail to satisfy the local-global principle, so in 1970, Manin defined a weakening of the local-global principle, the so-called Brauer-Manin obstruction. Since Manin's contribution, various classes have been proved to satisfy or to fail the local-global principle. Recently, there has been a considerable interest in the study of similar questions for various two-dimensional fields naturally arising in geometry. Let field $F$ be one of the following two types: (a) the function field of a smooth projective curve $X$ over $C((t))$, (b) the fraction field of a local, normal, Henselian, two-dimensional, excellent domain $A$ with algebraically closed residue field of characteristic $0$. The field $F$ then tends to have an arithmetic behavior similar to that of the usual global fields. The author of the paper under review studies the obstructions to the local-global principle for homogeneous spaces with connected or abelian stabilizers over finite extensions of the fields of types (a) and (b). He gives examples that prove that the Brauer-Manin obstruction with respect to the whole Brauer group is not enough to explain the failure of the local-global principle, and he then construct a variant of this obstruction using torsors under quasi-trivial tori which turns out to work. He follows Borovoi's work in the case of number fields. The paper is well done work and is a good sources for researchers in the field.