Generalized matric Massey products for graded modules (Q640290)

There is a theory for generalized matric Massey products, which includes cup-products, for computing local formal moduli, see \textit{O. A. Laudal}'s work [Lect. Notes Math. 1183, 218--240 (1986; Zbl 0597.14010)]. In the paper under review, the author proves that this theory applies to compute the local ring of the moduli of a graded \(A\)-module \(M\), \(A\) being a standard graded algebra over an algebraically closed field of characteristic zero. If \(M\) is cyclic, i.e., \(M \simeq S/I=A\) and \({\mathbb P}^n={\text{ Proj}}\,S\), then the mentioned local ring is up to completion isomorphic to the local ring of the Hilbert scheme \({\text{ Hilb}}^{p}( {\mathbb P}^n)\) at \(({\text{ Proj}}\,A)\) provided \({\text{ depth}} A > 1\), cf. \textit{G. Ellingsrud} [Ann. Sci. Éc. Norm. Supér. (4) 8, 423--431 (1975; Zbl 0325.14002)]. The author of this paper indicates a proof for the generalization of this result to non-cyclic modules, claiming that the local deformation functor of the graded \(A\)-module \(M\) and the local deformation functor of the corresponding coherent sheaf \(\mathcal M\) are isomorphic under the assumptions \(\Gamma_{*}({\mathcal M}) \simeq M\) and \({ }_0{\text{Ext}}^1(M,M) \simeq {\text{ Ext}}^1({\mathcal M, M})\) where \({ }_0{\text{Ext}}\) is the degree zero piece of the graded \(A\)-module \({\text{ Ext}}\) and \({\text{ Ext}}({\mathcal M, M})\) is global \({\text{ Ext}}\) of sheaves, cf. the Remark below. Several pages of this paper is devoted to a non-trivial and interesting example of computing the hull of the graded deformation functor of a cyclic module \(M\). The author nicely describes the hull \(H\) modulo the third power of the maximal ideal of H, i.e. the part which is given by the cup-products. To include the higher order Massey products the author cuts the hull by hyperplanes (i.e. puts several of the generators of \(H\) equal to zero). Then he uses the algorithm of the paper to compute higher order Massey products. It is illustrating to see how this algorithm works, even though the higher order Massey products turns out to vanish. The conclusions in the final proposition of the paper are, however, too ambitious and not correct as stated because the hull is changed in cutting by hyperplanes. The example had been more interesting if he had only cut the hull by those hyperplanes that leaves the hull unchanged up to isomorphism (there are quite a lot of such hypersurfaces in this situation). Reviewer's remark. In the reviewer's opinion, an assumption corresponding to an injection of their obstruction spaces has to be included to get the mentioned comparison result. The assumptions \(\Gamma_{*}({\mathcal M}) \simeq M\) and \({}_0{\text{Hom}}(M,H_{*}^1({\mathcal M}))=0\) imply an injection of obstruction spaces as well as \({}_0{\text{Ext}}^1(M,M) \simeq {\text{ Ext}}^1({\mathcal M, M})\) by Remark 2.6 (b) in [Can. J. Math. 62, No. 5, 1131--1154, (2010; Zbl 1206.14033)] of the reviewer, whence we get the comparison result by the cohomological argument referred to in Remark 2.6 (b).