PEcAn 2.0 list of non-backward compatible changes

[dlebauer]

Package names

all suggestions can be discussed in comments; checked when finalized

• [ ] no .'s that break Rcpp #391
• [ ] of format: "pecan<ACRONYM>"" pecanMA, pecanDB, pecanRTM, pecanVD
• [ ] convert long-names to acronyms (PEcAn.meta.analysis ->

Function names

• [ ] no [generic].s that falsely imply S3 methods, e.g. plot.variance.decomposition (#380)

Other Conventions

Variable names

• [ ] implement consistent use of CF style variable names?
  • Deprecate use of MsTMIP (from @mdietze)?

reorganization

• [x] reorganize packages #1191
• [ ] first release of PEcAn as a docker swarm
  • https://github.com/PecanProject/pecan/projects/7

[ashiklom]

A few BETY-related changes I'd like to see, while they're fresh on my mind:

• Remove the restriction that sites require citations.
• Remove the restriction on author length for citations (BETY PR 573 has been open for ages)
• Add status as a column in the BETY workflows (and runs?) table. Could be a JSON(-binary) field to give us flexibility. That would make it easier to check the status of workflows, and would dramatically accelerate the rendering of our History page (which IIRC currently parses the STATUS text file of every run) .

[robkooper]

• make parameters in workflow be a jsonb field
• add generic key/value to BETY (https://github.com/PecanProject/bety/issues/505)

[github-actions[bot]]

This issue is stale because it has been open 365 days with no activity.

[Aariq]

A suggestion: make it more clear from function names which functions return objects and which are called for their side-effects. E.g. get.trait.data() and do_conversions() return a modified settings object but run.meta.analysis() and model2netcdf.ED2() are called for side-effects. Maybe set_* for things that modify settings and run_* or write_* for functions called for side-effects?