A combinatorial characterization of extremal generalized hexagons (Q2112585)

In this clever paper, the author establishes a necessary and sufficient condition for a finite generalized hexagon to be extremal, that is, to have parameters \((s,s^3)\). An analogous result is since long known for generalized quadrangles, but remained open for the other classes of thick finite generalized polygons, namely hexagons and octagons. The condition generalizes the one for finite generalized quadrangles. It says the following: Let \((x,y,z)\) be a triple of points, with \(y\) and \(z\) at mutual distance \(2\) (distance in the point graph), with \(u\) the unique point collinear to both, and both \(y,z\) at distance \(3\) from \(x\). The author's condition requires that the number of points collinear to \(x\) and at distance \(2\) from both \(y\) and \(z\), is a constant, only depending on the distance from \(u\) to \(x\) (which can be \(2\), call it Case 1, or \(3\), call it Case 2). Moreover, the condition also requires that for each triple \((x,y,z)\) at mutual distance \(3\), the number of points collinear to one of these points, and at distance \(2\) from the other two points, is a constant (in principle only depending on \(x,y,z\) and not on the order in which you take them). The author also verifies by computer whether in the small known hexagons of order (s,s), the given numbers are constant, and he finds that in Case~1, this is indeed true in the dual split Cayley hexagons of order \((2,2)\) and \((4,4)\), and in Case~2, this is true for the split Cayley hexagons of order \((2,2), (3,3)\) and \((4,4)\), and for the dual split Cayley hexagon of order \((2,2)\). The latter is easy to verify by hand, only using the (small) order (and the constant is \(1\)). For all the dual split Cayley hexagons, Remark~6.3.5 of \textit{H. van Maldeghem}'s book [Generalized polygons. Basel: Birkhäuser (1998; Zbl 0914.51005)], implies that in Case~1, we indeed have a constant, and this constant is \(2\). Furthermore, the distance-\(2\) regularity implies directly that, for all split Cayley hexagons, in Case 2 we also have a constant, and this constant is~\(1\). The same tokens also show that the third condition can never be satisfied for (dual) split Cayley hexagons. Finally, the author obtains along the way a related combinatorial characterization of the regular extremal finite near hexagons.