@todd_a_jacobs There's no SemVer fail because Python doesn't follow SemVer (it predates it). Python minor versions are not backwards compatible, things are removed in minor versions:

https://docs.python.org/3/faq/general.html#how-does-the-python-version-numbering-scheme-work

For example, here are backwards incompatible removals in 3.13:

https://docs.python.org/3/whatsnew/3.13.html#removed-modules-and-apis

(Conda also doesn't use SemVer but CalVer)

@hugovk Once I knew what to look for, I also found PEP-2026, targeting the Python3 v3.14 release notes.

It contains reasons for CalVer, but IMHO it still makes dependency resolution more complicated. I switched to Ruby a long time ago because it was more *consistent* (not better) than Python where len() is still a built-in function rather than an object method..

Still, it works for the Python community and they expect it. Now I do, too. Thanks for the explanation!

https://peps.python.org/pep-2026/

@todd_a_jacobs You're welcome!

PEP 2026 would have replaced 3.15 with 3.26, but was rejected this week. Yeah, it wouldn't have changed much regarding resolution, it aimed not to change workflows too much. I'm very familiar with this PEP, I know its author well ;)