Is it (or will it become...) possible to fit a mixed non-linear SSmicmen model assuming log-normal error? I have tried to use a log transformed micmen model but it does not seem so powerful. Any other ideas about possible approaches are more than welcome! Irene