On \(k\)-stage Euclidean algorithms for Galois extensions of \(\mathbb{Q}\) (Q1922561)

The ring \(R=O_K\) of integers of an algebraic number field \(K\) is called Euclidean with respect to the norm if, given \(a,b\in R\), we can find \(q,r\in R\) such that \(a-bq=r\) and \(|N(r) |<|N(b) |\). In 1975, Cooke and Weinberger studied \(k\)-stage Euclidean number fields, where the condition that the norm should decrease in each division step is replaced by demanding that it should decrease after at most \(k\) steps. \textit{G. Cooke} and \textit{P. J. Weinberger} [Commun. Algebra 3, 481-524 (1975; Zbl 0315.12001)] showed that, assuming the truth of the generalized Riemann hypothesis, number fields with a real embedding and class number 1 are 3-stage Euclidean. In this paper, the author proves slightly weaker results without assuming GRH, by using the method developed in a previous paper [[1]\textit{D. A. Clark} and \textit{M. Ram Murty}, J. Reine Angew. Math. 459, 151-162 (1995; Zbl 0814.11049)]. Consider a real normal number field \(K\) with odd class number and degree \(n\). Adjoin the \(n-1\) square roots of a system of fundamental units and call the resulting field \(L\). The author shows that \(L\) is not contained in the ray class field of \(K\) with defining modulus \((\pi)\), where \((\pi)\) is a prime element in \(R\), and claims that the remainder of the proof parallels exactly that given in the cited paper [1]. In the reviewer's opinion, however, there are a few points that need to be clarified: 1. The main theorems of both papers are stated for integral domains whose quotient field is a finite real normal extension of \(\mathbb{Q}\); this includes e.g. rings of \(S\)-integers in such number fields. The proofs, on the other hand, are only worked out for the maximal order of such fields. 2. In the proof of [1], one needs the fact that \(L\cap F=K\), whereas the author only proves \(L\nsubseteq F\). 3. The field \(L\) defined in the paper differs from the one used in [1]; in the paper under review, the rank of \(\text{Gal} (L/K)\) is one less than in [1], which means that the argument used in [1] is only valid for number fields of degree at least 3. 4. The prime element \(\Pi\) in the proof of Lemma 5 in [1] has to be replaced by a prime ideal, since the field under consideration may have class number bigger than 1. 5. The upper bound for the number of integer \(s\)-tuples stated (without proof) in the proof of Lemma 6 in [1] is actually supposed to be a lower bound, and it is not valid for all values of \(s\) and \(y\). The idea behind the proof is taken from \textit{R. Gupta} and \textit{M. Ram Murty} [Invent. Math. 78, 127-130 (1984; Zbl 0549.10037) Lemma 2], where the special case \(s=3\) and \(K=\mathbb{Q}\) is treated. References for unproved claims are not given. Readers are advised to read \textit{R. Gupta}, \textit{M. Ram Murty}, and \textit{V. Kumar Murty} [CMS Conf. Proc. 7, 189-201 (1987; Zbl 0618.12006)] before studying [1] or the reviewed paper.