ADR Suggestion: Standardize on Python version support strategies

[rozyczko]

We should have a way to determine when to stop supporting older python versions and when to start supporting new versions.

This is to a large extent dependent on our main dependency - scipp.

We could possibly mirror scipp's support management for Python versions and important packages like numpy, scipy etc.

However, there is a non-trivial issue with modules like PySide (and to a lesser degree - pyinstaller), which are very cranky wrt. version dependencies.

[rozyczko]

We need to also consider requirements for the calculators we use in our projects

• cryspy
• pycrysfml

[damskii9992]

Is this really a corelib discussion or an organization-level discussion? And did we agree to follow the version-support of Scipp?

[rozyczko]

This should be an org-level discussion.
We did discuss following scipp for version support, at least for main modules like python/numpy.

[damskii9992]

Moved the discussion to the organization level and will close it and make a proper ADR