Warfield invariants in Abelian group algebras. (Q848646)

Let \(FG\) be a group algebra of an Abelian group \(G\) over a perfect field \(F\) of characteristic \(p>0\) and let \(V(FG)\) be the group of normalized units of \(FG\). Suppose that \(q\) is a prime number and \(\alpha\) is an ordinal. In this paper the author calculates the invariants \(W_{\alpha,q}(V(FG))\) introduced by Warfield under essential restrictions on the group \(G\). The main result is the following: Theorem 6. Suppose that \(F\) is a field of characteristic \(p\neq 0\) and \(G\) is an Abelian group. Then, for every ordinal number \(\alpha\), (1) \(W_{\alpha,p}(V(FG))=|G_t/G_p|\cdot W_{\alpha p}(G)\), provided \(F\) is perfect and either \(|G_t/G_p|\geq\aleph_0\) or \(p\neq 2\) and \(|G_t/G_p|=2\); (2) \(W_{\alpha,q}(V(FG))=W_{\alpha,q}(G)\) for all primes \(q\), provided \(G\) is \(p\)-mixed. We note that the reviewer has given counterexamples to cases 2 and 3 of the main theorem (Theorem 10) of a previous paper of the author [\textit{P. V. Danchev}, Extr. Math. 20, No. 3, 233-241 (2005; Zbl 1117.16022)]. Now the author writes that the result obtained in this paper supplies the main result of the above article because of ``some bugs which are corrected below'' (page 255, line 3 from below). This is not quite true. Indeed, the indicated counterexamples are not valid now for Theorem 6. However, Theorem 6 in the present paper does not replace the above mentioned Theorem 10, since in Theorem 6, in comparison with Theorem 10, there are essential restrictions on the conditions. Namely, Theorem 6, as we see from its formulation, does not give an answer to the question for \(W_{\alpha,p}(V(FG))\) when \(2<|G_t/G_p|<\aleph_0\), where \(G_t\) is the torsion subgroup of \(G\) and \(G_p\) is the \(p\)-component of \(G\). Besides the author considers a perfect field \(F\) of characteristic \(p\) instead of a perfect commutative ring of characteristic \(p\) without zero divisors, as in Theorem 10. Additional remarks. (i) The condition \(p\neq 2\) in case 1) of Theorem 6 is superfluous, since it follows from the condition \(|G_t/G_p|=2\). (ii) In the proof of Theorem 6 (page 258, lines 2-1 from below) there is an undefined ring \(R\) which is a technical mistake.