The principle of localization for the \([J,f(x,y)]\) means (Q1101636)

The \([J,f(x,y)]\) transform is the two-dimensional analogue of the \([J,f(x,y)]\) transform introduced by \textit{A. Jakimovski} [Bull. Res. Council, Israel, Sect. F 8, 135--154 (1960; Zbl 0094.26501)]: for \(f:{\mathbb{R}}^ 2\to {\mathbb{R}}\), and \(C^{\infty}\) variable, write \[ f^{(mn)}(x,y)=\partial^{m+n}f(x,y)/\partial x^m\partial y^n;\] then the \([J,f(x,y)]\) transform of a double sequence \((s_{mn})\) is defined as \[ t(x,y)=\sum_{m,n}\frac{(-x)^m}{m!}\frac{(-y)^n}{n!}f^{(mn)}(x,y)s_{mn}. \] If the defining series converges for all large \(x\) and \(y\), and tends to the limit \(t\) as \(x,y\to \infty\), then \(t\) is called the \(J\)-limit of \((s_{mn})\). The object is to apply this transform to the partial sums of a double Fourier series. A summary is given of different notions of bounded variation for a function of two variables, and the following theorem is proved: Let \(h(s,t)\) be periodic, of period \(2\pi\), BV with respect to \(s\) on \(I_1(\delta)=[\delta,\pi;-\delta,- \pi]\) and with respect to \(t\) on \(I_2(\delta)=[\pi,\delta;-\pi,- \delta]\), and let \(h(s,t)\) be bounded on \(I_1(\delta)\cup I_2(\delta)\). Then the \([J,f(x,y)]\) transform of the Fourier series of \(h(s,t)\) displays the principle of localization at \((0,0)\).