-
Notifications
You must be signed in to change notification settings - Fork 106
Description
I was thinking about usability of the parameter files. And currently they are structured as a long list of names, with the names sometimes helping to group them. I think this largely comes from them being NetCDF files where this is how they are handled. However, the move to JSON in #1493 opens up a new capability to add a level to sort by "group". Where group could be the general scientific process it corresponds with or specific packages or options. This would help map the parameters to the science and to the code.
Here's my comment where I introduce this idea.
I was thinking about this and I think it would help to have one additional level in here. And that's a groupname level. So there would be something like this:
"parameters": {
"fates_history": {
"comment": { "FATES History parameters"},
"fates_history_ageclass_bin_edges": {
...
}
"fates_generic":{
"comment": { "Generic FATES parameters that are always used"},
"fates_pftname": {
...
}
}
"fates_alloc":{
"comment":{ "FATES parameters having to do with allocation"},
"fates_alloc_organ_id": {
...
}
"fates_allom":{
"comment":{ "FATES parameters having to do with Allometry (FatesAllometryMod.F90)"},
"fates_allom_agb1": {
...
}
}
"fates_hydro": {
"comment":{ "FATES parameters for the fates_hydro parameterization (ONLY used when fates_hydro is on)"},
"fates_hydro_organ_name": {
...
}
"fates_spitfire": {
"comment":{ "parameters used by the spitfire model, and only used when it's active (in the fates/fire directory)"},
"fates_fire_alpha_SH": {
...
}
}
}}
Originally posted by @ekluzek in #1493 (comment)