On the Alexander-Hirschowitz theorem (Q2475119)

It is a classical interpolation problem in algebraic geometry to determine the dimension of the vector space of homogeneous polynomials of a given degree \(d\) in \(k[x_0,\dots,x_n]\) (\(k\) being a field) having multiplicity at least \(m_1,\dots,m_s\) at generic points \(P_1,\dots,P_s\in {\mathbb P}^n\). The problem is still open even for \(n=2\), but in the case of double points (\(m_1=\dots =m_s=2\)) it has been solved be J. Alexander and A. Hirschowitz which determine that the dimension is always the expected one, with a series of known exceptions. Alexander-Hirschowitz theorem was proved in a series of papers, and then simplified proofs were put out by themselves and K. Chandler. In this paper a nice exposition of the proof is given, along with details of its history which goes back to results of J. J. Sylvester, F. Palatini and A. Terracini. The theorem is also viewed in its connections to the study of higher secant varieties of Veronese varieties and to Waring problem for forms. In proving the theorem by induction, the initial case for \(d=3\) is treated here in a new way, which turns out to be more elegant and short than the previous ones.