Python as a success story for breaking backward compatibility is dubious.

When Python 3 was first promoted to the default download, many projects and libraries still hadn’t been upgraded from Python 2, some were coded against pre-releases of Python 3 and wouldn’t build using newer versions. Confusing and frustrating for those part-time Pythonistas not following the whole backstory.

Even now some projects and libraries aren’t compatible with Python 3, and here’s the reason why: making them compatible requires extra non-trivial work with no outward benefit. It means spending time not writing features to update the code for Python 3. Depending on complexity, it might be worth the cost, otherwise, fuck it, stick with Python 2.

Of course, projects usually have dependencies, and those dependencies also have dependencies. So before your project can go Python 3 you have to either wait until everything down the tree decides to make the effort, or sack up and do it yourself (which may include first learning C).

This is why we’re all still running multiple versions of Python, both 2 and 3, and all major Linux distros have packages for both. Maintenance is ongoing (eg. back porting security patches, kernel/module changes, etc.), and it’s been like this for years.