On log surfaces (Q1931915)

The paper under review is an announcement of the Minimal Model Program for (log) surfaces over an algebraically closed field in any characteristic. The main result says that a log surface pair \((X,\Delta)\) has a minimal model on which abundance holds (if \(K_X+\Delta\) is pseudoeffective) or a Mori fibre space (if \(K_X+\Delta\) is not pseudoeffective), if either \(X\) is \(\mathbb Q\)-factorial (with no additional assumptions on the pair \((X,\Delta)\)) or \((X,\Delta)\) has log canonical singularities. The result also holds in the relative setting, when there exists a projective morphism \(X\to S\) to another variety \(S\). The result is based on two papers by the authors: [\textit{O. Fujino}, Publ. Res. Inst. Math. Sci. 48, No. 2, 339--371 (2012; Zbl 1248.14018)] in the case of characteristic zero, and [\textit{H. Tanaka}, ``Minimal models and abundance for positive characteristic log surfaces'', \url{arXiv:1201.5699}] in the case of positive characteristic. Also, a simplified proof of the contraction theorem due to Artin and Keel is given. The paper is a well-written survey which highlights differences between methods used in characteristic zero and positive characteristic, and also the differences between the geometry over \(\overline{\mathbb F}_p\) and over other ground fields of characteristic \(p>0\).