The minimal resolution of the ideal of a general arrangement of a big number of points in \(\mathbb{P}^ n\) (Q5961445)

If one chooses a generic set \(\{P_1,\dots, P_s\}\) of points in projective space \(\mathbb{P}^n\) over an algebraically closed field \(k\), there is a natural conjecture what the minimal graded free resolution of its vanishing ideal \(I=I(P_1,\dots,P_s) \subseteq S=k[x_0,\dots,x_n]\) should look like. This conjecture is known as the minimal resolution conjecture and was formulated explicitly by \textit{A. Lorenzini} in her Ph.D thesis [see also \textit{A. Lorenzini}, J. Algebra 156, No. 1, 5-35 (1993; Zbl 0811.13008)]. For the case \(n=2\), it was shown to hold by \textit{A. V. Geramita} and \textit{P. Maroscia} [J. Algebra 90, 528-555 (1984; Zbl 0547.14001)], and the case \(n=3\) was solved by \textit{E. Ballico} and \textit{A. V. Geramita} [in: Algebraic geometry, Proc. Conf., Vancouver 1984, CMS Conf. Proc. 6, 1-10 (1986; Zbl 0621.14003)]. Furthermore, the conjecture has been proved in a number of special cases, some of which were settled by computer calculation [cf. \textit{S. Beck} and \textit{M. Kreuzer}, ``How to compute the canonical module of a set of points'', in: Algorithms in algebraic geometry and applications, Proc. MEGA-90 Conf., Santander 1994, Prog. Math. 143, 51-78 (1996)]. Recently, however, D. Eisenbud and S. Popescu have shown that for small numbers of points \(s\leq 2n\), there are infinitely many counterexamples to the minimal resolution conjecture [cf. \textit{D. Eisenbud} and \textit{S. Popescu}, ``Gale duality and free resolutions of ideals of points'' (preprint 1996)]. In the paper under review, the authors demonstrate that the conjecture is true for large numbers of points \(s\gg n\). Unfortunately, as they point out at the end of the paper, their method is not very effective and gives a bound of something like \(s=6^{n^3\log n}\). Nevertheless, their ``methode d'Horace'' is interesting in its own right and has been applied successfully by J. Alexander and the first author in other contexts as well [cf. \textit{J. Alexander} and \textit{A. Hirschowitz}, J. Algebr. Geom. 1, No. 3, 411-426 (1992; Zbl 0784.14001)]. More precisely, the authors reduce the minimal resolution conjecture to the claim that the restriction homomorphism \(H^0 (\mathbb{P}^n, \Omega_{\mathbb{P}^n}^p (\ell))\to \bigoplus_{i=1}^s \Omega_{\mathbb{P}^n}^p (\ell)|_{P_i}\) has maximal rank for \(0\leq p\leq n\) and all \(\ell\geq 0\). This claim is proved via a complicated and subtle induction argument. For the start of this induction the authors cite the thesis of \textit{F. Lauze} [in: ``Sur la resolution des arrangements génériques de points dans les espaces projectifs'' (Thesis Univ. Nice 1994)]. The methode d'Horace is then used to lift maximal rank statements like the desired one from a nonsingular divisor of a smooth projective variety to the variety itself.