The Toronto problem (Q388820)

A \textit{Toronto space} is a topological space that is homeomorphic to each of its full-cardinality subspaces. The \textit{Toronto problem} asks whether there exists a non-discrete uncountable Hausdorff Toronto space, and was first posed by J.~Steprāns [``Steprān's problems'', in: J. van Mill (ed.) et al., Open problems in topology. Amsterdam etc.: North-Holland. (1990; Zbl 0718.54001)]. Toronto spaces are analogous to fractals, in that they satisfy a ``self-similarity'' condition, not unlike one that crops up in continuum theory: both the arc and the pseudo-arc are continua that are homeomorphic to each of their nondegenerate subcontinua. One of the main results of the paper under review is a classification of the non-\(T_1\) Toronto spaces of any given infinite cardinality \(\kappa\): up to homeomorphism, there are exactly three; namely the indiscrete topology, and the lower and upper topologies on the linearly ordered set \(\kappa\). (\(T_1\) topologies include the co-finite, the co-countable, etc.) The other main focus in this paper is on what a non-discrete Hausdorff Toronto space \(X\), of cardinality \(\aleph_1\), would look like if it were to exist. (If \(X\) did exist, it would have to live in a model of set theory where the Continuum Hypothesis fails.) For example: (1) \(X\) would be scattered, and each subspace would have just countably many isolated points; (2) \(X\) could not be first countable; (3) \(X\) would be neither locally compact nor paracompact; (4) \(X\) would be hereditarily separable but not Lindelöf (an S-space); and (5) \(X\) could not be both regular and countably compact.