On the idempotency of some composite functors (Q1758972): Difference between revisions

In [\textit{I. Chatterji}, ed., Guido's book of Conjectures, Monographies de l'Enseignement Mathématique 40. Genève: L'Enseignement Mathématique (2008; Zbl 1143.00006)], E. Dror-Farjoun raised the following question: let \(L\) and \(CW\) be, respectively, a localization and a cellularization functor, which are homotopy idempotent by definition. Are the composition functors \(L\circ CW\) and \(CW\circ L\) also homotopy idempotent? In this paper the author answers the above in the negative by proving the following result: Let \(CW\) be the cellularization with regard to the wedge \(B{\mathbb Z}/2\vee \bigl(\vee_{p\,\,\text{odd prime}} \,\,M({\mathbb Z}/p,2)\bigr)\), and let \(L\) be the localization functor which extracts the \(n\)th Postnikov section of a given space for an appropriate \(n\). Then, \(L\circ CW\) and \(CW\circ L\) are not homotopy idempotent. Along the way, interesting results about the homotopical behavior of the classifying space of the Suzuki simple group \(Sz(8)\) are obtained.