A delay theorem for pointlikes (Q5952235)

For a fixed pseudovariety \(\mathbb{V}\) of semigroups (or monoids), a subset \(P\) of a finite semigroup \(S\) is said to be \(\mathbb{V}\)-pointlike if \(P\) relates to a single element under every relational morphism between \(S\) and a semigroup (or a monoid) in \(\mathbb{V}\). One says that \(\mathbb{V}\) has decidable pointlikes if, given a finite semigroup \(S\) and a subset \(P\subseteq S\), one can decide whether or not \(P\) is \(\mathbb{V}\)-pointlike. Let \(\mathbb{D}\) denote the pseudovariety of all finite semigroups in which every idempotent is a right zero. The main result of the paper (Theorem 1.2) shows that a monoid pseudovariety \(\mathbb{V}\) has decidable pointlikes if and only if the semidirect product \(\mathbb{V}*\mathbb{D}\) has decidable pointlikes. This generalizes the so-called Delay Theorem by \textit{B.~Tilson} [J. Pure Appl. Algebra 48, 83-198 (1987; Zbl 0627.20031)]. In fact, the author gives a nice description of the \(\mathbb{V}*\mathbb{D}\)-pointlike sets of a given semigroup \(S\) in terms of the \(\mathbb{V}\)-pointlike sets of a certain category formed from the idempotents of \(S\). The main technical tool of the paper is the notion of a profinite category.