Non-compact subsets of the Zariski space of an integral domain (Q2408446)

Let \(D\) be an integral domain and \(\mathrm{Zar}(D)\) be the Zariski space of the valuation overrings of \(D\) contained in the fraction field \(K\) of \(D\). Two important properties for subspaces of \(\mathrm{Zar}(D)\) to investigate are the properties of compactness and Noetherianess (a topological space is Notherian if its open sets satisfy the ascending chain condition). Denote by \(\mathrm{Over}(D)\) the set of overrings of \(D\) contained in \(K\). The Zariski topology of \(\mathrm{Over}(D)\) is generated by the sets of the form \(\{T\in \mathrm{Over}(D)\mid x_1\in T, \dots, x_n\in T\}\), as \((x_1,\dots,x_n)\) ranges among the finite subsets of \(K\). Let \(D[{\mathcal F}_f]\) be the set of finitely generated \(D\) algebras of \(\mathrm{Over}(D)\). First, the author proves that if \(\Delta\subseteq \mathrm{Zar}(D)\) dominates \(D[{\mathcal F}_f]\), then \(\Delta\) is compact if, and only if, it contains \(\mathrm{Zar}_{\mathrm{min}}(D)\) (the set of minimal elements of \(\mathrm{Zar}(D)\)). It follows that if \(\mathrm{Zar}(D)\backslash \{V\}\) is compact, where \(V\in \mathrm{Zar}_{\mathrm{min}}(D)\), then \(\dim(V)\leq 2\dim(D)\) and \(V\) is the integral closure of the localization \(D[x_1,\dots,x_n]_M\) for some \(x_1,\dots,x_n\) in \(K\) and \(M\) a maximal ideal of \(D[x_1,\dots,x_n]\). At the end of the paper he proves that \(D[{\mathcal F}_f]\) is dense in \(\mathrm{Over}(D)\), with respect to the constructible topology, and that there is no spectral subspace \(X\subsetneq \mathrm{Over}(D)\) such that \(D[{\mathcal F}_f]\subseteq X\). A criterion based on semistar operations on \(D\)-submodules is introduced in some proofs. Such an operation can be seen as a kind of closure. The author studies the valuative dimension \(\dim_v(D)\) of \(D\), i.e. the supremum of the dimensions of the overrings of \(D\) contained in \(K\) (we have \(\dim(D)\leq \dim_v(D)\)). He deduces that if \(D\) is a Prüfer domain with Noetherian spectrum, then \(\mathrm{Zar}(D)\) is Noetherian. He also constructs domains whose Zariski spaces are Noetherian, but with \(\dim(D)<\dim_v(D)\). We can quote more theorems or propositions proved in this paper. Let \(\Delta\) be a set of pairwise incomparable prime ideals of \(D\). Assume that \(D\) is local and that \(D_Q\) is a valuation domain, where \(Q\) is the intersection of all ideals in \(\Delta\) (so \(Q\) does not belong to \(\Delta\), since the ideals in \(\Delta\) are pairewise incomparable). Then \(\mathrm{Zar}(D)\) is not Noetherian and, for every overring \(V\) of \(D\) contained in \(D_Q\), \(\mathrm{Zar}(D)\backslash \{V\}\) is not compact. If \(D\) is not a field, then \(\mathrm{Zar}(D[X])\) is not a Noetherian space. As a corollary, let \(F\subset L\) be a transcendental field extension. If \(\mathrm{trdeg}_F(L)=1\) and \(L\) is finitely generated over \(F\), then \(\mathrm{Zar}(L|F)\) is Noetherian. If \(\mathrm{trdeg}_F(L)>1\) , then \(\mathrm{Zar}(L|F)\) is not Noetherian. Then the author gives an application of these criteria to the study of Kronecker function rings. Finally, assume that \(D\) is Notherian and let \(X_1=\{V\in\mathrm{Over}(D) \mid V\) is Noetherian and of dimension at most \(1\}\), \(X_2=\{V\in\mathrm{Over}(D) \mid V\text{ is a Dedekind domain (}K\) included)