Classification of embeddings of the flag geometries of projective planes in finite projective spaces. I, II, III (Q1976324)

[This is a combined review for the three papers Part I--III.] Consider a finite projective plane \(\Pi\) of order \(s\) with point set \(\mathcal P\), line set \(\mathcal L\) and flag set \(\mathcal F\). Then the point-line geometry \(({\mathcal F},{\mathcal P}\cup {\mathcal L})\) where incidence is induced by containment, is a generalized hexagon of order \((s,1)\). Conversely, every finite generalized hexagon of order \((s,1)\) can be constructed in this fashion from a projective plane of order~\(s\). Generalized hexagons of order \((s,1)\) are also called thin generalized hexagons. A thin generalized hexagon \(\Gamma\) of order \((s,1)\) is said to be fully and weakly embedded in \(\text{PG}(d,q)\), considered as a point-line geometry, if \(\Gamma\) is a subgeometry of \(\text{PG}(d,q)\) such that \(s=q\) and the point set of the generalized hexagon generates \(\text{PG}(d,q)\) but for every point \(p\) of the generalized hexagon, the set of points not opposite~\(p\) does not already generate \(\text{PG}(d,q)\). In a previous paper [Des. Codes Cryptogr. 17, 97-104 (1999; Zbl 0942.51004)] the authors showed that in the above situation \(d\in \{6,7,8\}\) and the projective plane associated with the thin generalized hexagon is Desarguesian. In the papers under review, all these embeddings are classified. The combined result is, that all such embeddings are in the known class of classical examples or in the new class of semi-classical examples. The semi-classical examples are only possible for \(d=8\) and came as a surprise. Although the result is rather neat, the proof is lengthy and splits up into three parts. In the first part, the case \(d=6\) is treated completely and the case \(d=7\) is treated under the additional assumption, that the points of \(\Gamma\) not at maximal distance to some fixed line are contained in a 4-dimensional space. In the second part it is shown, that this additional assumption holds if there are four distinct lines incident with the fixed line, such that these four lines generate a 4-dimensional subspace. Finally, in the third part, the authors show that for \(d=7\) this is always the case, while for \(d=8\) there are counterexamples which lead to the semi-classical embeddings. Full and weak embeddings of thin hexagons will be an important tool when investigating embeddings of finite Moufang generalized hexagons. A finite Moufang hexagon contains a thin generalized hexagon or the dual of a thin generalized hexagon. Embeddings of finite Moufang generalized hexagons containing a thin generalized hexagon have been successfully investigated by the authors in two earlier papers [J. Lond. Math Soc. (2) 54, 566-580 (1996; Zbl 0865.51006) and Eur. J. Comb. 19, 733-751 (1998; Zbl 0927.51013)].