A study on convergence and ideal convergence classes (Q1646536)

It is well known that topologies can be described by specifying which nets should converge to which points in the underlying set, see \textit{J. L. Kelley} [Duke Math. J. 17, 277--283 (1950; Zbl 0038.27003); General topology. Reprint of the 1955 original published by van Nostrand. Mineola, NY: Dover Publications (2017; Zbl 1358.54001)]. In this paper the authors consider a few modifications of Kelley's ideas. First they put a mild restriction on subnets in one of Kelly's axioms and show that this does not really change anything: these so-called semi-convergence classes yield exactly the same topologies as standard convergence classes. The second modification, applied to both types of convergence classes, is that one can consider ideals on the domains of the nets and specify that convergence of a net~\(\langle x_d\rangle_{d\in D}\) to~\(x\) requires that \(\{d:x_d\notin U\}\) belongs to the ideal whenever \(U\)~is a neighbourhood of~\(x\). The ideal modification of semi-convergence classes can be put into one-to-one correspondence with the standard convergence classes in such a way that paired classes yield the same topology. This does not quite work in the case of modifications of standard convergence classes.