On the number of generators of ideals in polynomial rings (Q2630926)

\textit{M. P. Murthy} asked [in: Conf. commut. Algebra, Kingston196--211 (1975; Zbl 0354.14015)] whether an ideal \(I\) in a polynomial ring over a field, \(R\), can be generated by the same number of elements required to generate \(I/I^2\). If we denote by \(\mu\) the minimal number of generators, the question asks whether \(\mu(I)=\mu(I/I^2)\) for such ideals. It is not difficult to see that \(\mu(I/I^2)\leq\mu(I)\leq\mu(I/I^2)+1\). The reviewer had answered this question in the affirmative with the additional assumption that \(\mu(I/I^2)\geq\dim R/I+2\) [Invent. Math. 46, 225--236 (1978; Zbl 0395.13009)]. The article under review claims to prove this in full generality. Unfortunately, there are gaps in the proof and the author's errata is expected to appear soon [\textit{J. Fasel}, Ann. Math. (2) 186, No. 2, 647--648 (2017; Zbl 1372.13015)]. The author and others have recently proved a number of beautiful results in this general area using \(\mathbb{A}^1\)-homotopical methods developed by \textit{F. Morel} and \textit{V. Voevodsky} [Publ. Math., Inst. Hautes Étud. Sci. 90, 45--143 (1999; Zbl 0983.14007)]. This article, while inspired by those theories, uses only what the author calls, naive homotopies, a far simpler concept. The article proves that one can lift any set of generators of \(I/I^2\) to a set of generators for \(I\). But, this has been known to be false due to an example of Mandal and Murthy. This example appears in an errata by S. Mandal, to appear in Journal of Algebra, which corrects \textit{S. Mandal}'s paper [J. Algebra 458, 156--170 (2016; Zbl 1358.13022)] which generalized Fasel's paper, and thus the final results are in doubt, since it uses these results. It is believed that both Theorem 1.0.6 and Lemma 3.2.3 need more justification.