On reconstructing separable reduced p-groups with a given socle (Q1101830): Difference between revisions

Assuming G.C.H. the author proves the following interesting theorem. Suppose G is a pure subgroup of the torsion-completion of an (unbounded) direct sum of cyclic p-groups. Then he can find another pure subgroup H with the same socle \(H[p]=G[p]\) but not isomorphic to G. The proof uses heavily the combinatorial machinery as \(\lambda\)-sets and \(\lambda\)-systems. It would be desirable (e.g. a good master thesis) to derive a more algebraic proof from the author's arguments. Clearly, there are other ``invariants'' which could be prescribed at the same time and still \(G\not\cong H\).