Good points of a hypersurface (Q1110593)

Let (R,F,G,H) be a 4-tuple, with R a regular local ring, F,G,H pairwise coprime ideals in R such that GH has a normal crossing at R. The problem is to show that, upon applying a suitable finite succession of blowing ups, FGH will have only normal crossings. Here F is of primarily interest, while G and H play an auxiliary role. To study this problem the author, inspired by Zariski's concept of equisingular point of the hypersurface of local equation F, introduces and studies in great details the concept of a good point of F. Roughly speaking, this is a point of F with the property that to simplify it, it takes the same effort as it does to simplify a point of a plane curve. The prototype of a good point in dimension 2 is the singularity at 0 given by \(Z^ d-X^ aY^ b=0\), with \(a+b\geq d>a'+b'\), where \(a'\) and \(b'\) are the residues of a and b modulo d. Although good points are cruder than equisingular points, the author shows that they are better adapted for the desingularization problem. The paper has three parts. In the first part (the longest one) the author develops the local theory in dimension 2 aiming to prove that after sufficiently many monoidal transformations applied to a d-fold point of a surface, all the resulting d-fold points are good. The second part is a globalization of the first one, and, as a rather easy application, the author obtains a proof of the desingularization theorem for surfaces (via good points). The last part discusses some open problems. The author has developed this theory of good points in 1966, and the reason of publishing it after more than twenty years is that in the meantime he succeeded in finding another proof of the desingularization theorem in arbitrary dimension; in this sense the present paper could be useful as a good introduction to higher-dimensional theory of good points.