Curvettes and clusters of infinitely near points (Q644554): Difference between revisions

From the author's abstract: ``The aim of this paper is to revise the theory of clusters on infinitely near points for arbitrary fields.'' Let \(K\) be a field, and let \(\Omega(K)\) be the set of two-dimensional regular local rings which have \(K\) as field of quotients; for \(R\in\Omega(K)\) the rings \(S\in\Omega(K)\) with \(S\supset R\) are said to be infinitely near to \(R\). Let \(R\in \Omega(K)\) and \(\mathfrak m\) be the maximal ideal of \(R\); every homogeneous principal prime ideal \(p\in \text{gr}_{\mathfrak m}(R)\) determines \(S_p\in \Omega(K)\) with \(S_p\supset R\); \(S_p\) is said to be a quadratic transform of \(R\), and the family of quadratic transforms of \(R\) is said to be the rings in the first neighborhood \(N_1(R)\) of \(R\) (one sets \(N_0(R)=\{R\}\)). Let \(R\subsetneqq S\) lie in \(\Omega(K)\); then there exists a uniquely determined sequence \(R=R_0\subset R_1\subset\cdots\subset R_h=S\) where, for \(i\in\{1,\dots,s\}\), \(R_i\) is a quadratic transform of \(R_{i-1}\); this sequence is called the quadratic sequence from \(R\) to \(S\). Now \(R\) determines an ideal in \(S\), the exceptional divisor; in Lemma 2.17 the author proves that the result which is well-known in the classical case -- where \(R\) is essentially of finite type over an algebraically closed field -- holds also in this more general setting. The order function \(\text{ord}_R\) of the maximal ideal of \(R\) determines a discrete valuation ring \(V\) of \(K\); \(S\supsetneqq R\) is proximate to \(R\), \(S\succ R\), if \(S\subset V\). Let \(p\) be as above, and let \(n_p\) be the order function of \(p\,\text{gr}_{\mathfrak m}(R)\). Then \(\nu_p(-)=(\text{ord}_R(-),(n_p(-))\) defines a valuation of \(R\). In Prop.\ 2.27 the author determines \(\nu_p(f)\) for \(f\in R\setminus\{0\}\) in terms of the points infinitely near to \(R\); a similar result for finite-colength ideals of \(R\) was shown by \textit{J. Lipman} [in: Algebraic geometry and commutative algebra, Vol. I, 203--231 (1988; Zbl 0693.13011), Prop.\ 2.3 and Lemma 2.4]. In section 3 the author introduces the notion of a cluster \(\mathcal C\) (for the notion of a cluster in the classical case (cf.\, e.g., [\textit{A. Casas-Alvero}, Singularities of Plane Curves. London Mathematical Society Lecture Note Series. 276. Cambridge: Cambridge University Press (2000; Zbl 0967.14018)]), and defines, following Lipman [loc.~cit.], the refined proximity matrix \(P'_{\mathcal C}\) [there is a typo in the definition: if \(S\prec T\), then \(p'_{S,T}=[T:S]\)] and the total proximity matrix \(\widetilde R_{\mathcal C}\) [\,there is a typo: if \(S\prec T\), then \(\widetilde p_{S,T}=[T:R]\)]. Now let \[ X=X_s\overset{\pi_s}\rightarrow X_{s-1}\cdots X_1\overset{\pi_1}\rightarrow X_0\eqno(*) \] be a sequence of point blow-ups of two-dimensional regular schemes; more precisely, let \(x_0\in X_0\) be a closed point, let \(X_1=\text{Bl}_{x_0}(X_0)\) be the blow-up of \(X_0\) in \(x_0\), choose a closed point \(x_1\in X_1\) lying over \(x_0\), let \(X_2={\text{Bl}}_{x_1}(X_1)\) be the blow-up of \(X_1\) etc. Setting \(R_i:=\mathcal O_{X_i,x_i}\) for \(i\in\{0,\dots,s\}\), we get the quadratic sequence \[ R_0\subset R_1\subset\cdots\subset R_s; \] this sequence is called the cluster associated to \((*)\). In Theorem 4.8 the author determines the elements of the intersection matrix \(N_{\mathcal C}\) associated to this cluster [in all the formulae on pp. 453--454 one must replace \(\widetilde p_{U,S}\) by \(\widetilde p_{S,U}\) and \(p_{U,T}\) by \(p_{T,U}\)]. In section 5, the author defines the Hamburger-Noether tableau of a pair \((x,y)\) of elements in \(V=k[\![\,t\,]\!]\), \(k\) a field, following \textit{P. Russell} [Manuscr. Math. 31, 25--95 (1980; Zbl 0455.14018)] (there is no need to use the algebraic closure of \(k\); the elements \(a_i\) lie in \(k\)). In the last two sections the author works with theses tableaus.