Ah. It sounds as if Carleson dealt with infinite Fourier series of
finite-length waves, while the earlier theory was finite Fourier series of
infinitely long waves. That point was missing from the original article.


--