Update on 3-folds

The division of compact Riemann surfaces into 3 cases K_C<0, g=0, or K_C=0, g=1, or K_C>0, g>=2 is well known, and corresponds to the familiar trichotomy of spherical, Euclidean and hyperbolic non-Euclidean plane geometry. Classification aims to treat all projective algebraic varieties in terms of this trichotomy. The model is the treatment of surfaces by Castelnuovo and Enriques around 1900. Because the canonical class of a variety may not have a definite sign, we usually have to modify a variety by a minimal model program (MMP) before it makes sense to apply the trichotomy; this involves contractions, flips and fibre space decompositions. The classification of 3-folds was achieved by Mori and others during the 1980s. New results over the last 5 years have added many layers of subtlety to higher dimensional classification. These include: -> extension of MMP to dimension 4 (Shokurov), -> rational connected varieties (Campana and Koll\'ar--Miyaoka--Mori), -> explicit classification results for 3-folds (Corti, Kawakita and others), -> Calabi-Yaus, mirror symmetry and string theory applications, -> resolution of orbifolds and McKay correspondence, -> the derived category as a characteristic of varieties (Bondal, Orlov, Bridgeland and others), -> birational rigidity (Sarkisov, Corti, Pukhlikov and others). The study of 3-folds also yields a rich crop of applications in several different branches of algebra, geometry and theoretical physics. My lecture surveys some of these topics.