V 1.0, shall we consider the API as stable?

[kikocorreoso]

Hi,

I'm following the development of the usual suspects, floris, pywakeand foxes, and what I can see is that in general some minor versions have changes breaking some code, in floris we have version 4 when version 3 is being developed and not all the changes have been included in the new one, etc.

Please, do not misinterpret me. It is ok to break things when it is needed. It is ok to follow rapid development to include many improvements, etc.

I'm very grateful for your work being done as open source. We can better understand what it is done. We can validate results against other packages (open or closed source). We can help with the development, examples, documentation, etc.

So, I'm wondering if foxes APIs could be considered "stable" with the release of version 1.0.

[SchmJo]

Thanks for the question. I am also hoping for some API convergence, and I am quite positive that the main API will be rather stable now (e.g. the basic components like engines, states, farm, mbook, algo will remain).

However, the project as a whole is still very active, since it is our main tool for wake modelling and every project brings in new development ideas. For example, the recent idea to run foxes through parameter files instead of Python scripts.

Generally, we are trying to follow this versioning conventions: Versions are either A.B.C or A.B.C.D, where

• A: Major changes in the core, affecting API and results.
• B: Minor changes that might affect parts of the API and results.
• C: Sub-minor changes, new model implementations, and bug fixes that should not impact the API too much.
• D: Meta data, documentation or other changes that do not alter the code base. No need to follow those for typical users.

Did that somehow help to clarify things? I am always interested in any kind of feedback, thank you.