On finite \(p\)-groups that are the product of a subgroup of class two and an abelian subgroup of order \(p^3\) (Q509909)

Let \(G=AB\) be a group which is the product of two nilpotent subgroups \(A\) and \(B.\) It is well-known that the group \(G\) must be soluble and it was conjectured that the derived length of \(G\) should be bounded by the sum of the classes of the factors. However, \textit{J. Cossey} and \textit{S. Stonehewer} [Bull. Lond. Math. Soc. 30, No. 3, 247--250 (1998; Zbl 0939.20017)] constructed examples of finite \(p\)-groups of derived length four that can be expressed as the product of an abelian subgroup and a subgroup of class two. The author's work is motivated by the fact that the abelian factor in these counterexamples is quite large, being of order \(p^{p^3 }\) for \(p\) odd. Thus, it is reasonable to ask how small can the abelian factor be in such counterexamples. The author proves that if \(G=A B\) is a finite \(p\)-group that is the product of an abelian subgroup \(A\) of order \(p^3\) and a subgroup \(B\) of nilpotency class two, then \(G\) can have derived length at most three. This is a consequence of some structural result in Theorem 5 and the main contribution of the paper.