Real homogeneous spaces, Galois cohomology, and reeder puzzles (Q314256)

The main object of the paper under review is a homogeneous space \(X=G/H\) where \(G\) is a simply connected absolutely simple algebraic group defined over the field of real numbers \(\mathbb R\) and \(H\) is simply connected semisimple \(\mathbb R\)-subgroup of \(G\). The authors are interested in computing the number of connected components of \(X(\mathbb R)\), which is equal to the number of \(G(\mathbb R)\)-orbits on \(X(\mathbb R)\), which is equal to the cardinality of the kernel of the map \(H^ 1(\mathbb R, H)\to H^ 1(\mathbb R,G)\). To compute the latter number, the authors propose a method allowing one to describe explicitly the first Galois cohomology of any simple simply connected real group. It is based on combinatorial considerations going back to \textit{M. Reeder} [J. Algebra 285, No. 1, 29--57 (2005; Zbl 1157.20330)]. The authors generalize Reeder's techniques by extending his method to non-simply-laced Dynkin diagrams and, even further, to coloured Dynkin diagrams which correspond to non-compact inner forms of compact groups (here the authors use another combinatorial notion, the so-called Kac diagram, introduced by \textit{M.~Borovoi} and \textit{D.~A.~Timashev} [``Galois cohomology of real semisimple groups'', Preprint, \url{arXiv:1506.06252}]). The cardinalities of the first Galois cohomology sets of all simple simply connected real groups were earlier computed by \textit{J.~Adams} [``Galois cohomology of real groups'', Preprint, \url{arXiv:1310.7917}] using different methods.