If $X$ is a closed $n$-manifold, a *fake $X$* is another closed manifold homotopy equivalent to $X$. There is some interest in classifying manifolds (up to, say, homeomorphism) homotopy equivalent to a given manifold; the Poincare conjecture is the special case of $X = S^n$ (there are no topologically fake $S^n$s). I'm interested in this in the case of $X = \Bbb{KP}^n$, where $\Bbb K = \Bbb {R,C,H,O}$ (where in the $\Bbb O$ case we fix $n=2$; there are no other objects which deserve to be named "$\Bbb{OP}^n$".) One reason to find this interesting is that fake projective spaces are precisely the manifolds that have a sphere bundle over them with total space a sphere, with as usual the unfortunate exception that this is not true for $\Bbb{OP}^2$. (I would note that this is different than the algebro-geometric usage of "fake projective plane", where to my understanding what was sought was a classification of compact complex manifolds with the same Hodge diamond as complex projective space.)

Both real and complex fake projective spaces have been classified: both computations are carried out in Wall's book on surgery theory (see sections 14C and 14D for the classifications of fake complex and real projective spaces, respectively), and there are nice, briefer descriptions of the real and complex cases on the Manifold Atlas.

I've had some difficulty in finding a classification of fake quaternionic projective spaces or fake octonionic projective planes. Has this been carried out, and if so, what is a reference?