A characterization of non-Noetherian BFDs and FFDs (Q2409592)

An integral domain is called \textit{atomic} if every nonzero nonunit is a product of irreducible elements. Over the last three decades, the study of various properties intermediate between atomicity and factoriality has been one of the main focuses of factorization theory. One of the main driving forces behind this movement was the paper [\textit{D. D. Anderson} et al., J. Pure Appl. Algebra 69, No. 1, 1--19 (1990; Zbl 0727.13007)], which is an excellent place to start if one is looking to get into factorization theory. Two kinds of domains studied in that paper were \textit{bounded factorization domains} (BFDs), which are domains in which every nonzero nonunit has a finite upper bound on the lengths of its factorizations into nonunit elements, and \textit{finite factorization domains} (FFDs), which are domains in which every nonzero element has only finitely many divisors up to associates. It is easy to see that FFD \(\Rightarrow\) BFD \(\Rightarrow\) ascending chain condition on principal ideals (ACCP). Nice characterizations of Noetherian FFDs are known [\textit{D. D. Anderson} and \textit{B. Mullins}, Proc. Am. Math. Soc. 124, No. 2, 389--396 (1996; Zbl 0851.13006)] and Noetherian domains are always BFDs. These properties are less well-understood in the non-Noetherian case, which is the focus of the paper under review. The author gives characterizations of BFDs and FFDs with a ``topological'' flavor. As an application, he gives new information about two of A. Grams's examples of non-Noetherian Prüfer domains that satisfy the ACCP [Proc. Camb. Philos. Soc. 75, 321--329 (1974; Zbl 0287.13002)], showing that these examples are in fact respectively a BFD and an FFD. Out of all the ``topological'' notions introduced in the present paper, the following are two of the most interesting. The author calls a domain \(D\) \textit{connected} if no nonzero element has an infinite sequence of comaximal divisors (not to be confused with ``connected'' in the usual ``indecomposable'' sense). He calls \(D\) \textit{finitely coverable} if for every \(0 \neq x \in D\) there is a finite union of maximal ideals that contains every divisor of \(x\). The author proves that (i) ACCP \(\Rightarrow\) connected and (ii) a finitely coverable almost Dedekind domain is a BFD. (This makes a pleasing comparison with the well-known fact that a finite character almost Dedekind domain is Dedekind.) It is worth observing that one can make some adjustments to the author's argument to generalize (ii): if \(D\) is finitely coverable and locally a BFD, then \(D\) is a BFD.