No label defined (Q4953606)

This article provides an excellent summary of the history of homology manifolds from Bing decompositions to the current non-resolvable examples of the Bryant, Ferry, Weinberger and the author. A topological space \(X\) is an n-dimensional homology manifold if for every point \(x\in X\), \(H_*(X,X\smallsetminus x)\simeq H_*(\mathbb R^n,\mathbb R^n \smallsetminus 0).\) The work of Cannon and Edwards in the late seventies supported the Resolution Conjecture: Every ENR homology of dimension at least five is resolvable. In the mid-eighties Quinn produced an integer obstruction to resolution of homology manifolds, but there was still no example of a non-resolvable homology manifold until 1993. In 1993 Bryant, Ferry, Mio and Weinberger produced a large class of examples: For every integer and every simply connected topological manifold they construct a homology manifold with the given resolution obstruction in the homotopy type of the given topological manifold. Bryant, Ferry, Mio and Weinberger also prove a version of Browder, Novikov, Sullivan, Wall surgery theory for homology manifolds. The author wisely does not explain much about this new surgery theory where the structure set is made up of homotopy equivalences from homology manifolds, but he does discuss the most interesting point: Siebenmann Periodicity holds in the new theory, because the homology manifold structures fill in the structures missing from topological manifold surgery. Both the ordinary surgery for topological manifolds and controlled surgery for topological manifolds are used in the BFMW construction explained in section 6 of this paper. The reader who is inexperienced with manifold surgery is warned that section 3 on controlled surgery theory is better postponed until after reading the first half of section 4 in which some ordinary surgery is reviewed. The paper concludes with some suggestions for applications of the study of homology manifolds: as limits of manifolds in Gromov-Hausdorff space, ends of groups, or as a route to a disproof of the existence part of the Borel conjecture. For the entire collection see [Zbl 0933.00057].