A counterexample for some problem for de Groot dual iterations (Q837634)

Given a topological space \((X,\tau)\), a set \(A\subset X\) is called saturated if \(A=\bigcap {\mathcal A}\) for some family \({\mathcal A}\subset \tau\). The family of all compact saturated subsets of the space \(X\) is a closed base for a topology \(\tau^d\) which is called the de Groot dual of \(\tau\). It is proved in the paper that there exists a topological space \((X,\tau)\) such that \(\tau\) does not coincide with \(\mu^d\) for any topology \(\mu\) on the set \(X\) and, besides, \(\tau^d=\tau^{ddd} \neq \tau^{dd}\) while \(\tau^{dd} \subset \tau\) and \(\tau^{dd}\neq \tau\).