The canonical contact structure on the link of a cusp singularity (Q743664)

A normal surface singularity is called a cusp singularity if the exceptional set of the minimal resolution is a cycle \(E=\cup_{i=1}^n E_i\) of non-singular rational curves, i.e., each \(E_i\) intersects its two neighbours transversally at one point, respectively, and there are no other crossings. It is known by \textit{U. Karras} [Proc. Symp. Pure Math. 30, 37--44 (1977; Zbl 0352.14007)] and \textit{H. B. Laufer} [Am. J. Math. 99, 1257--1295 (1977; Zbl 0384.32003)] that the link of an embedded cusp singularity in \({\mathbb C}^3\) is a Sol manifold, i.e., the mapping torus \(T_A\) of a torus diffeomorphism given by some hyperbolic matrix \(A\in SL(2,{\mathbb Z})\). The suspension flow on \(T_A\) is an Anosov flow with stable and unstable foliation defined by some \(1\)-forms \(\beta_\pm\) and it is known that \(\beta_++\beta_-\) yields the canonical contact structure on the link of the embedded cusp singularity. The paper under review gives two new ways of seeing this contactomorphism. One is an interpretation of unpublished work of Furukawa and uses the moment polytope, i.e., the image of the canonical map \(S^5\to {\mathbb R}^3\). The other approach uses \textit{F. E. P. Hirzebruch}'s resolution of Hilbert modular cusps [Enseign. Math. (2) 19, 183--281 (1973; Zbl 0285.14007)] (and proves that the latter are cusp singularities) to describe the associated contact structures.