Constructive finite free resolutions (Q663308)

The authors find interesting ways to simplify a number of results in [\textit{D. G. Northcott}, Finite Free Resolutions. Cambridge Tracts in Mathematics. 71. Cambridge etc.: Cambridge University Press. (1976; Zbl 0328.13010)]. To use the authors' words, these proofs become now elementary in a logical sense in that minimal prime ideals have been replaced by their local-global principle in Lemma 2.3: Let \(J\) and \(I=\langle a_1,\dots,a_m\rangle\) be two finitely generated ideals. If \(I\) is regular (the only element that annihilates each \(a_i\) is 0) and \(J=0\) in each of the localizations \(R[1/a_i]\), then \(J=0\) in \(R\). If \(I\) is regular and \(J\) is regular in each localization \(R[1/a_i]\), then \(J\) is regular. Another idea was to define grade via homology, which enables them to prove directly the Auslander-Buchsbaum-Hochster (Theorem 2, Chapter 6, ibid.) along with a number of results on conditions for exactness of a complex (no localizations at primes are used in these proofs). While acknowledging that relevant Northcott's ideas are still useful in certain situations, the authors strengthen the old results as well as generalize them. They show that the greatest common divisor of elements can be algorithmically computed, based on the fact that if the ideal generated by \(a_1,\dots a_n\) has a finite free resolution, then these elements have a (regular) greatest common divisor.