Topological vector spaces, compacta, and unions of subspaces (Q2518755)

By a topological vector space (for short, t.v.s.), we mean a locally convex topological vector space over the field \(\mathbb R\) of real numbers. It is well-known that a t.v.s. \(L\) is compact only when \(L\) contains the zero vector only, and that every finite-dimensional t.v.s. is \(\sigma\)-compact, that is, the union of a countable family of compacta. However, an infinite-dimensional t.v.s. can also be \(\sigma\)-compact. For instance, the \(\sigma\)-product of \(\omega_1\) copies of \(\mathbb R\), denoted by \(P_\sigma\), is a t.v.s. of this type. Note that \(P_\sigma\) cannot have a locally finite cover by compacta, since any t.v.s. having a locally finite cover by compacta must be finite-dimensional. Thus, it is an interesting and natural question to study when a \(\sigma\)-compact space admits a ``combinatorially nice'' covering by compacta. In Section 2 of this paper, the author investigates when a \(\sigma\)-compact space can be covered by a countable point-finite family of compacta. It is shown that every metrizable \(\sigma\)-compact space \(X\) covered by countable family of metrizable compacta can be covered by a countable point-finite family of metrizable compacta. In the class of topological vector spaces, the situation can be reversed, that is, a t.v.s. \(L\) covered by a countable point-finite family of compacta can be covered by a countable point-finite family of metrizable compacta. In Section 3, the author studies another interesting question: When can a compact space \(X\) be represented as the union of two (or finitely many) subspaces that are homeomorphic to topological vector spaces? The main result asserts that such spaces must be metrizable (and thus separable). As usual as in many of his other papers, the author also poses several interesting open problems.