1Here's a list of TODO items... 2 3--> still a bit unclear on how to generate the initial beam 4 conditions... 5 6--> maybe at a later point have the reserved memory in routines 7 like fit_waveform be static so we don't have to reserve + free 8 it call after call, might speed things up... 9 10---> in fit_waveform.c: 11 - Probably need to put in some box constraints for better behaviour at a certain point, 12 can use the routine nr_lmder_bc for this. 13 - Maybe program in something to replace the matrix inversion based upon the LU algorithm 14 in nr_levmar.c. The original version had some LAPACK functionality, but as we 15 wanted this standalone, i've cut it out. 16 - Best way to use it probably to take the parameters from previous fit of last pulse and have 17 them as initials for the current fit. 18 19--> in fit_fft.c 20 - properly check fitting algorithm, works... but question is how well... 21 22 23--> in process_waveform.c 24 - if the waveform is saturated, we now set the tSample to the time of the last 25 unsaturated sample, still need to shift this in order to have a decent sampling 26 window... 27 28 29