Derived length when a product of characters has two nonprincipal irreducible constituents. (Q2803563)

Let \(\chi\) be a nonlinear irreducible character of a finite group \(G\); then it is known that the principal character \(1_G\in\chi\overline\chi\). Let \(\eta(\chi)\) be the number of nonprincipal irreducible constituents of \(\chi\overline\chi\) and suppose that \(G\) is solvable. \textit{E. Adan-Bante} proved the existence of a linear bound for the derived length of \(G/\ker(\chi)\) in terms of \(\eta(\chi)\), and [in J. Algebra 266, No. 1, 305-319 (2003; Zbl 1034.20008)] she described groups with faithful \(\chi\) such that \(\eta(\chi)\in\{1,2\}\). Her results on the case \(\eta(\chi)=1\) are quite complete, but for \(\eta(\chi)=2\) she only obtained partial results, such as that the derived length of such groups is at most 18.NEWLINENEWLINE The paper under review adds to our understanding of the case \(\eta(\chi)=2\) by proving that, in fact, the best possible bound for the derived length is 8. The proof that 8 is an upper bound requires careful study of linear group actions. To see that 8 cannot be improved, the authors exhibit an interesting example of order 165,888 with faithful character \(\chi\) of degree 8.