Half-Moufang implies Moufang for generalized quadrangles (Q4449495)

A simple path \((x,y,z)\), \(x\neq z\), of length two in a generalized quadrangle is a Moufang path, if the group of all collineations fixing every element incident with one of \(x\), \(y\), \(z\) acts transitively on the set of apartments containing \(x\), \(y\), \(z\) and two fixed elements \(u\), \(v\), with \(u\neq y\) incident with \(x\), and \(v\neq y\) incident with \(z\). If all simple paths of length two are Moufang paths, then the quadrangle is called a Moufang quadrangle, and these have been classified by \textit{J.Tits} and \textit{R. M. Weiss} [Moufang polygons (Springer Monographs in Mathematics, Springer, Berlin) (2002; Zbl 1010.20017)]. However, there are two kinds of simple paths of length two: the ones with a point in the middle, and the ones with a line in the middle. If all simple paths of (at least) one kind are Moufang paths, then we call the quadrangle half Moufang. \textit{J. A. Thas}, \textit{S. E. Payne} and the reviewer showed in [Invent. Math. 105, 153--156 (1991; Zbl 0708.51010)] that finite half Moufang quadrangles are automatically Moufang quadrangles. In the paper under review, the author shows the same result for all half Moufang quadrangles, without the finiteness restriction. It is a marvellous result. The proof has some history: the author first used a lemma on permutation groups that she borrowed from a colleague. But the lemma turned out to be wrong, as remarked by R. Weiss. The reviewer then repaired the proof in the frame of a course that he gave on the subject. It turns out that, up to one argument of the author of the paper under review, the proof of the implication ``half Moufang to Moufang'' can be reduced to the repairing argument of the reviewer, as shown in the paper ``Characterizations of Moufang quadrangles'' (joint with F. Haot), to appear in the Glasgow Journal of Mathematics.