CCPP Framework Meeting Minutes 2021 04 27

Agenda

• capgen.py progress
• Transition to capgen.py
• Process for creating a master cross-lab standard name resource
• Other business

capgen.py process

• No significant updates.

Transition to capgen.py

• Dom's ccpp-physics PR #611 is in the commit queue (use assumed-size array declarations, fix inconsistencies between metadata and Fortran code), will be merged in the second week of May

Process for creating a master cross-lab standard name resource

• Weiwei debugging her Python script, hopefully done soon

Other business

• When is a variable intent(out) in a CCPP physics scheme?
  • It is the responsibility of the scheme to set the variable entirely, not just partially
  • Compilers may set the incoming data to whatever they want
• Error handling: ccpp-physics needs to produce error messages that are compliant with NCO standards
  • This shouldn't be a big deal (fairly standard format), but we need to make sure that other users of CCPP are ok with that
  • Alternatively, a more sophisticated way to configure the "logger" for error messages in CCPP
  • See https://github.com/NCAR/ccpp-physics/pull/641
  • Steve's error handler could be augmented to allow specifying prefices for the different error levels that can be set by the host model
  • DTC also has a deliverable to improve the error handling in CCPP in the next period of performance (07/21-06/22)
  • Need to define a set of error levels and provide instructions for CCPP scheme writers
  • Can we talk to NCO to get information on what levels they use (and what format)?