Dear Marco,
I am working on the integration of the dseries class in dynare and I have a question related to your codes. My commits are in a personal repository:
git.ithaca.fr/dynare/src/use-dseries
The main change is that dataset_ is now a dseries object (instead of a structure) and I added another structure dataset_info to store descriptive statistics and informations related to the missing informations.
I think I have no problem with respect to the parallel stuff or gsa, but I miss something in the routines related to identification. In dynare_identification.m, starting line 133 you do something if dataset_ is empty. I didn't figure out yet how to handle this, but could you explain what you are doing in this case ? Is the identification thing working without data ? Also in identification_routine.m (line 144) you put simulated data in the data_info (dataset_) structure if init==1, but I don't understand why you need that.
Best, Stéphane.