Survey on the topology of real algebraic sets (Q1061798)

The paper is a very lively presentation of the subject. First the author discusses the restriction on the topology of real algebraic sets: a real algebraic set is triangulable hence is a cone on a space X. Sullivan proved that X has even Euler characteristic. Since the one point compactification of a real algebraic set can be made algebraically, Sullivan's condition gives some extra informations. Further more complicated restrictions have been given by S. Akbulut and the author. - Then the author studies sufficient conditions for a set to be algebraic. After an historical discussion he indicates Tognoli's result: any smooth compact manifold is diffeomorphic to a real algebraic set. A similar result is true for manifolds with boundaries. - The classification of real algebraic sets with isolated singularities, which is now complete, is then discussed. In the case of non isolated singularities, the idea is to get a topological resolution of singularities, make the resolution algebraic and then blow down algebraically. In the two dimensional case it is known that Sullivan's condition is sufficient. In the case of dimension three a counterexample by S. Akbulut and the author shows that Sullivan's condition is not sufficient. One says that the homology of a real algebraic set is totally algebraic when every homology class can be represented by a real algebraic set. It is known that in some cases (the Grassmannian for example) all the homology is algebraic and that in some other cases it is not. S. Akbulut and the author conjectured that any compact manifold is diffeomorphic to a real algebraic set with all its homology algebraic. Benedetti gave recently a counter-example to this conjecture.