Not easy to modify cod OM as a user

[k-doering-NOAA]

While working on addressing #18, I learned that it is not intuitive on how to change the default cod model (part of package external data) to make new scenarios. For example, just trying to increase the M value lead to many problems (M needed to be changed in both control and par, or changed in control and the model needs to be run with no estimation; ultimately that M scenario is too high and crashed the population, so is probably not a realistic OM).

I think we could add functionality into SSMSE to make this process easier for users:

• Make a check on the OM files for the SSMSE user: do the control and ss.par files say the same thing? Do the original ctl and control.ss_new say the same thing? Warn or exit on error if they don't. The user needs to be clear on where the model values are coming from. For example, changing the M value control.ss_new file for the cod model alone and leaving the original ss.par file lead to the original M in the ss.par file to be retained, even thought the starter file specified that the values should come from the control file originally (SSMSE changes it to read from the par).
• Some sort of helper function to allow users to test out new OMs? I'm not sure how exactly to do this, but maybe running the OM through SSMSE and finding it crashing is not the way to go?
• Expand documentation on changing OMs to make multiple scenarios with different OMs. This will be done to some extent in the readme, but we may need a long form guide eventually.

Probably some other ideas will be obvious as we spend more time trying to use the SSMSE code with more OMs.

[k-doering-NOAA]

@nathanvaughan-NOAA suggests changing the iter_list argument for run_SSMSE from a list of vectors to a vector. E.g.,

iter_list = list(1:5, 1:5), # run with 5 iterations each

to

iter_list = c(5,5) # run with 5 iterations each

Pros: makes input simpler; cons: SSMSE will need to do a bit more gymastics to find what to name folders (probably just look for the highest numbered folder in the path it is looking at).

[k-doering-NOAA]

@nathanvaughan-NOAA I was thinking about working on this issue (starting with your suggestion to change the iter_list argument, which I am thinking to change to iter_vec). Would this be ok to do without your changes merged into master from your dev branch? Or should I wait until you merge in so there are less conflicts?

[k-doering-NOAA]

the .ss_new model files are used for the OM, but the original files are used for an EM. I think this makes sense, but needs to be documented

[k-doering-NOAA]

I think this is sufficiently dealt with now.