“…The crux is, as we saw in the reasoning above, that for non analytic boundaries we have to let the interpolation points approach the boundary, while for analytic boundaries, the interpolation points can be separated from the boundary. When we let the interpolation points approach the boundary, it is natural to expect (and formally proven in [3,Theorem 5]) that we cannot obtain convergence for any analytic function on D. Analytic functions can behave very wildly near the boundary and we may get ''bad information'' by interpolating at such points.…”