Notes on INS: spectral decompostion in the time variable ?