Balanced projectives and axiom 3 (Q1093739)

The paper under review is well-written, informative and important. The paper achieves a unification of Hill's approach to totally projective groups with Warfield's treatment of (global) balanced projective groups. The main result is Theorem 4.2: The following six conditions are equivalent for any abelian group G. (1) There is an \(H(\aleph_ 0)\)- family of K-nice subgroups in G; (2) there is a \(G(\aleph_ 0)\)-family of K-nice subgroups in G; (3) G has a composition series consisting of K- nice subgroups; (4) G is balanced projective; (5) G has a K-basis and there is an \(H(\aleph_ 0)\)-family of nice subgroups in G; (6) G contains a K-basis X with \(G/<X>\) a totally projective group. An \(H(\aleph_ 0)\)-family in G is a family \({\mathcal C}\) of subgroups containing \(\{\) \(0\}\), closed under sums, and for each \(A\in {\mathcal C}\) and each countable subgroup C of G, containing a subgroup B such that \(A+C\subset B\) and B/A is countable. A \(G(\aleph_ 0)\)-family is the same except for requiring only closure under unions of chains rather than arbitrary sums. Thus statement (1) of the theorem is just Hill's ``3rd Axiom of Countability'' with ``nice'' replaced by ``K-nice''. A composition series for G is a well-ordered smooth ascending chain \(\{N_{\alpha}\}\) of subgroups of G whose union is G and such that each quotient \(N_{\alpha +1}/N_{\alpha}\) is either primary cyclic or infinite cyclic. This leaves the ``K-concepts'' unexplained. The development of these concepts make up the bulk of the paper, constitute its greatest achievement but require many other definitions to explain fully. For those with sufficient background a group G is a K-group if for each finite subset S there is a finite rank, free *-valuated subgroup F of G such that \(<S,F>/F\) is finite. A subgroup N of G is K-nice if N is nice in G, for each \(g\in G\) there is a positive integer m such that the coset \(mg+N\) contains an element y which has the same (generalized) p- height as \(mg+N\) for each prime p, and G/N is a K-group. Theorem 2.6 states that every direct summand of a K-group is a K-group. The paper refers frequently to the papers by \textit{P. Hill} and \textit{C. Megibben} [Trans. Am. Math. Soc. 295, 735-751 (1986; Zbl 0597.20047)] and \textit{M. Lane} [Proc. Am. Math. Soc. 96, 379-386 (1986; Zbl 0594.20049)] for proofs or models for proofs. It is a remarkable fact that the results of the paper are independent of the uniqueness and existence theorems previously established for balanced projective groups.