Selection principles and Sierpinski sets (Q2644344)

This paper studies selection principles in the real line with the density topology. The density topology is a means to consider measure-theoretic properties as topological ones. In the density topology, compact sets are finite, and it makes sense to consider more general covering properties. A classical result of \textit{F. D. Tall} [``The density topology'', Pac. J. Math. 62, 275--284 (1976; Zbl 0305.54039)] tells us that a set of real numbers is Lindelöf in the density topology if, and only if, it is a Sierpiński set. One of the main results of this paper is that Sierpiński sets actually satisfy the generally stronger covering property of Menger: For each countable family of open covers, there is a cover consisting of finitely many elements from each of the given covers. The author considers various additional covering properties from the literature, whose systematic study was initiated in [\textit{M. Scheepers}, ``Combinatorics of open covers. I: Ramsey theory'', Topology Appl. 69, No.~1, 31--62 (1996; Zbl 0848.54018)]. Each of these properties is stated as a possibility to diagonalize countable families of open covers, via a certain procedure, or \textit{selection principle}. Originally, these properties were studied for sets of reals with the standard topology of the real line, and here these properties are considered for the density topology. Each of quite a few considered properties is shown to be equivalent to being \(\kappa\)-Sierpiński for a suitable cardinal \(\kappa\). (A set of reals is \textit{Sierpiński} if its intersection with each Lebesgue measure zero set is countable. It is \textit{\(\kappa\)-Sierpiński} if all such intersections have cardinality less than \(\kappa\).) In particular, the Continuum Hypothesis implies that all these properties coincide. This is in sharp contrast to the situation for the standard topology, where all properties differ in the presence of the Continuum Hypothesis. One of the considered properties, namely \(\mathsf{S}_1(\Omega,\Omega)\), stands out as more difficult to characterize. The author supplies for this one a construction and an interesting problem.