Landau's theorem for holomorphic curves in projective space and the Kobayashi metric on hyperplane complements (Q652920)

The main result of the paper is an effective version of a theorem of \textit{J. Dufresnoy} [Ann. Sci. Ec. Norm. Supér., III. Sér. 61, 1--44 (1944; Zbl 0061.15205)]: For any set of \(2n+1\) hyperplanes in general position in complex projective space \({\mathbb P}^n\) there is an explicit constant \(K\) such that for every holomorphic map \(f\) from the unit disc in the complex plane with values in the complement of these hyperplanes in \({\mathbb P}^n\) satisfies \(f^\sharp(0)\leq K\), where \(f^\sharp\) denotes the norm of the derivative of \(f\) with respect to the Fubini-Study metric. More precisely, the result is the following: Theorem. Let \(H_0,\dots,H_{2n}\) be \(2n+1\) hyperplanes in general position in \({\mathbb P}^n\). Let \[ G=\max_{0\leq j_0<\cdots<j_n\leq 2n} \max\Big\{\log\Lambda(H_{j_0},\dots,H_{j_n}), \log(n+1)-\log\lambda(H_{j_0},\dots,H_{j_n}) \Big\} \] and \[ G^\sharp=\min_{0\leq j_0<\cdots<j_n\leq 2n} \dfrac 1{\lambda^\sharp(H_{j_0},\dots,H_{j_n})}. \] Let \(B=\binom {2n+1}{n+1}\) be the binomial coefficient and \[ K=12,672\big(2.6\cdot 10^7\log B+10^8\big)^{6(4\log B+20)} GG^\sharp. \] Let \(f\) be a holomorphic map from the unit disc in \(\mathbb C\) to \({\mathbb P}^n\) omitting the hyperplanes \(H_0,\dots,H_{2n}\). Then \[ f^\sharp(0)\leq K. \] The norm of \(f'\) with respect to the Fubini-Study metric is \(f^\sharp=\|f\wedge f'\|/\|f\|^2\) and the functions \(\lambda\), \(\lambda^\sharp\), \(\Lambda\), and \(\Lambda^\sharp\) are defined in the following way: Let \([X_0:\cdots:X_n]\) be homogeneous coordinates in \({\mathbb P}^n\) and let \(H_0,\dots,H_n\) be \(n+1\) hyperplanes in general position in \({\mathbb P}^n\) given by linear defining forms \(H_j(X_0,\dots,X_n)=a_{j0}X_0+\cdots+a_{jn}X_n\) normalized so that \(\|H_j\|^2=|a_{j0}|^2+\cdots+|a_{jn}|^2=1\). Consider the \((n+1)\times (n+1)\) matrix \(A=(a_{jk})\) and let \(0<\lambda_0\leq \lambda_1\leq \cdots\leq \lambda_n\) be the \((n+1)\) eigenvalues of the matrix \(AA^*\) listed in non-decreasing order. Then define \(\lambda(H_0,\dots,H_n)=\lambda_0\), \(\lambda^\sharp(H_0,\dots,H_n)=\sqrt{\lambda_0\lambda_1}/\lambda_n\), \(\Lambda(H_0,\dots,H_n)=\lambda_n\), and \(\lambda^\sharp(H_0,\dots,H_n)=\sqrt{\lambda_{n-1}\lambda_n}/\lambda_0\). For the proof of the theorem the authors use many interesting results from potential theory. The paper is very well written.