Desingularization preserving stable simple normal crossings (Q2339611)

The paper under review is about ``partial'' resolution of singularities of an algebraic variety (over a characteristic zero field). More precisely, the authors say that a variety \(X\) has a \textit{stable simple normal crossing} (s-snc) at a point \(a\) if, locally near \(a\), there is a closed embedding of \(X\) into a smooth variety \(Z\) such if \(X^{(1)}, \ldots, X^{(m)}\) are the irreducible components of \(X\) that contain \(a\) and \(X_a = \bigcap _{i=1}^{m}X^{(i)}\), then \(X_a\) is smooth at \(a\) and \(\sum_{i=1} ^{r} {\mathrm{codim}} (X^{(i)},Z)={\mathrm{codim}} (X_a,Z)\). Then the embedding dimension of \(X\) at \(a\) equals the dimension of \(Z\). They give several equivalent versions of this definition. They also work with pairs \((X,D)\), where \(X\) is as above and \(D\) a \({\mathbb Q}\)-Weil divisor on \(X\) and they similarly define the notion ``a pair \(((X,D)\) has s-snc at \(a \in X\)''. Given a pair \((X,D)\) and a smooth closed subvariety \(C\) of \(X\), there is a new pair \((X_1,D_1)\), the \textit{transform} of \((X,D)\) with center \(C\), where \(X_1\) is the the blowing-up of \(X\) with center \(C\) and \(D_1=D'+E'\), with \(D'\) the birational transform of \(D\) and \(E'\) the exceptional divisor. Their main result is: {Theorem 1.} Given a pair \((X,D)\) as above, there is a sequence of blowings-up \(X=X_0 \leftarrow \cdots \leftarrow X_t\) with smooth centers \(C_i \subset X_i\), \(i=0, \ldots, t-1\), such that if \((X,D)=(X_0,D_0)\) and for all \(i \geq 1\) \((X_i,D_i)\) is the transform of \((X_{i-1},D_{i-1})\) with center \(C_{i-1}\), then \((X_t,D_t)\) has only s-snc singularities, and each blowing-up \(X_{i-1} \leftarrow X_i\) is an isomorphism over the set of s-snc points of \(X_{i-1}\). The assignment of the sequence above to a pair \((X,D)\) is functorial with respect to smooth morphisms \(X' \to X\) that preserve the number of irreducible components at each point. In the special case where \(X\) is smooth, Theorem 1 gives a log resolution of the divisor \(D\). They also give a definition of {simple normal crossings singularity} and present an example showing that the analog of Theorem 1 is not valid for snc singularities. Theorem 1 generalizes the main theorem of \textit{E. Bierstone} and \textit{F. V. Pacheco} [Rev. R. Acad. Cienc. Exactas Fís. Nat., Ser. A Mat., RACSAM 107, No. 1, 159--188 (2013; Zbl 1285.14014)]. The notion of snc point given in the present paper is more general than that found in the article just mentioned article, or in [\textit{E. Bierstone} and \textit{P. D. Milman}, Adv. Math. 231, No. 5, 3022--3053 (2012; Zbl 1257.14002)], where a a snc point is necessarily a hypersurface singularity, which is not the case with the new definition. Due to the structure of the proof, the authors find it more convenient to show a similar theorem but for triples \((X,D,E)\), where \((X,D)\) is a pair as above and \(E = a_1 H_1 + \cdots + a_r H_r\) is snc divisor on \(X\) (the \(H_i\) taken in a given order), which easily implies Theorem 1. The proof of this result involves, among other things, the use of the function inv, introduced by Bierstone and Milman in their mentioned article, the Hilbert-Samuel function and, several times, the technique of \textit{presentations}, i.e., the introduction of an auxiliary object involving a sheaf of ideals, whose cosupport is related to values of an invariant. Of course, these techniques must be adapted to the present situation, and a number of subtle points are addressed by the authors. In spite of its technical nature, the paper is relatively self contained and well written.