release version 2.0.0

[edwardhartnett]

Description

As soon as #906 is merged, I will do a GitHub release of fv3atm. This will have no impact on other developers.

Additional context

The release process will be:

• adding a VERSION file.
• updating release notes.
• generating documentation and uploading to gh-pages.
• doing the GitHub release.

Output

There are no code changes and there will be no change of output data.

Testing:

By inspection on GitHub after release.

Dependent PRs:

N/A

[DusanJovic-NOAA]

What is the purpose of this release? How are users/developers supposed to use it? Will there be corresponding releases of the subcomponents and/or ufs-weather-model?

[edwardhartnett]

It's purpose was as described in the meeting in College Park; there was a presentation. As described in that meeting we will be doing quarterly releases in 2025. This is the first.

Users are not expected to use it.

Developers (i.e. us) will use it as a baseline to measure performance and progress in the SENA effort to bring fv3atm code under unit testing. Future releases in the series (but probably not this one) will be used in spack-based testing.

There will not be corresponding releases of sub-components or the ufs-weather-model. Nor, typically, is there in multi-repo software projects. For example where HDF5 does a new release, there is usually not a corresponding new netcdf-c release, and vice-versa. The decoupling of the projects is some extra effort, but really pays off in the long run.

As discussed in the meeting, doing a release is a low-cost process that does not result in extra work for programmers. This will have no impact on existing work.

[DusanJovic-NOAA]

I find it strange that the GitHub release, which is publicly available, is not expected to be used by public (users). Is public release what should be used to track the progress in SENA effort, whatever that is. Are there no other project management tools available to track the progress. This is the first time I see software being released that is not supposed to be used. There should be a big bold warning in the release description notifying users that this release is not to be used for anything.

The example of HDF5/netcdf-c is irrelevant. HDF5 library can be used without netcdf, fv3atm can not be used without the ufs-weather-model. Opposite is also true for HFD5/netcdf-c, netcdf can be used without HDF5, while fv3atm can not be used for anything without ccpp, atmos_cubed_sphere, upp, etc.

[edwardhartnett]

Excellent points. I have added this to the release notes:

Note that this release represents a snapshot of the fv3atm project and is not intended for release in operational systems without extensive testing. This release is the beginning to a transition to a more agile process on the fv3atm project. Future releases will include better testing and documentation of this important code base.

You can see the draft release notes under the releases page.