Topology of the compactified Jacobians of singular curves (Q2471774)

Inspired by the paper of \textit{S. T. Yau} and \textit{E. Zaslow} [Nucl. Phys. B 471, No. 3, 503--512 (1996; Zbl 0964.81521)], \textit{A. Beauville} [Duke Math. J. 97, No. 1, 99--108 (1999; Zbl 0999.14018)] introduced an invariant of a plane curve singularity \((C,p)\), namely the Euler number of the compactified Jacobian of a projective rational curve \(C\) having \(p\) as its unique singularity and showed that, while counting the number of rational curves in a complete linear system on a \(K3\)-surface, the multiplicity every curve has to be counted with, equals the product of such invariants for each singular point. On the other hand, \textit{B. Fantechi, L. Göttsche} and \textit{D. van Straten} proved [J. Algebr. Geom. 8, No. 1, 115--133 (1999; Zbl 0951.14017)] that for a plane singularity \((C,p)\), the Euler number equals the multiplicity of the \(\delta\)-constant strata in the base of the semi-universal deformation of the singularity. Explicit computations of this invariant are however rather difficult; so far the only known Euler numbers are those of the cusps \(x^p -y^q = 0\) with \(p,q\) coprime. In the paper under review, the author computes the Euler number for unibranched curve singularities with the following characteristic Puiseux exponents: \((p,q)\) with \(p,q\) coprime, \((4,2q,s)\), \((6,8,s)\), \((6,10,s)\) with \(q,s\) odd. He uses the fact that in these cases there exists a cell decomposition of the compactified Jacobian, from which the Betti numbers can be computed by a combinatorial method; moreover some of the Betti numbers can be stated explicitly. This result shows in particular that the Euler number depends only on the Puiseux exponents i.e. on the topological type of the singularity.