On the uniqueness of the generalized octagon of order \((2,4)\) (Q468711): Difference between revisions

There is a generalized octagon of order \((2,4)\) which belongs to the group \(^2F_4(2)\). Whether this is the unique generalized octagon of this order is still an open question. There is just one vertex transitive one (see [\textit{I. N. Belousov} and \textit{A. A. Makhnev}, Math. Notes 84, No. 3, 483--492 (2008; Zbl 1165.05328); translation from Mat. Zametki 84, No. 4, 516--526 (2008)]) or also just one Moufang octagon [\textit{J. Tits}, Am. J. Math. 105, 539--594 (1983; Zbl 0521.20016)]. In this paper, the authors prove that there is a unique octagon of order \((2,4)\) if it admits a group \(Q\) of automorphisms which fixes a point \(a\) and all lines through \(a\) and acts transitively on the set \(P_a\) of points opposite to \(a\). For the proof, the authors define a subgroup \(Q_0\) of \(Q\) as the group generated by elements \(\alpha_i \in Q\), \(i= 1, \dots , 5\), where \(b\) is a fixed element in \(P_a\), \(\{b,c_i\}\), \(i = 1, \dots, 5\), are the lines through \(b\) and \(b^{\alpha_i} = c_i\). They show that \(2^8 \leq |Q_0| \leq 2^{10}\) and prove many properties of this group. Using these facts, they go over the GAP-list of 2-groups of order \(2^8\), \(2^9\) and \(2^{10}\) and eventually come down with a unique one, which is of order \(2^9\). This then implies uniqueness of the generalized octagon. The authors say that to treat the groups of order \(2^8\) and \(2^9\) it takes 10 CPU-years and for \(2^{10}\) it will be 15 CPU-years. The referee and probably also many readers are not familiar with this kind of measure. Hence, it is not clear whether one is able to verify the proof in reasonable time.