It has nothing to do with loyalists and everything to do with practicalities. I was responsible for a fairly big project in 2011 and we decided that Python was the language of choice for some of the components. We started out with Python 3... but then soon realized that a lot of libraries and tooling that were available for Python 2 were not available for Python 3. So, we had to convert our code to Python 2 and continue because we had already spent some time with it. In hindsight, we should have picked some other language than this god awful mess. It pains me to see that Python is the most used language today given their history of how they handled Python 2/3.

Why does that pain you? Given that they know how painful that move from 2 to 3 was, you can be sure that they will never even consider a break like that again.

Trust me I have no interest in using python 2. My issue is exactly what you stated: breaking lots of programs that relied on it being there. The fact that it's being removed in a minor update is even worse.

Yup. Photographers and the like are being slowly brickwalled by DisplayCAL (UI for colord, ArgyllCMS, used for many thing like cross-platform measuring/calibrating monitors) in an unmaintained by usable state being removed from package repos everywhere because of Python 2.7 removal.

> I find it amazing the lengths python 2 loyalists take to keep using it

You don't need to spin issues that can be purely technical in terms of personal traits like stubbornness.

Just keep in mind that software doesn't change itself, and if you have a python27 script written yesterday then it won't be magically ported to python37 during the night. Unless people do the work, that won't change. Given there are far more people using software than writing it, then I don't understand what leads you to believe that software is perpetually maintained and ported to newer releases.

Python 2 was originally going to be EOL'd in 2015, the announcement for which was in 2008. Only later was the EOL pushed back to 2020. So while it's only been officially "dead" for two years now, the writing had been on the wall for 12 years before that.

At this point it is stubbornness - I know software doesn't magically get ported and updated as time goes on, but I also acknowledge that if no one cares to update the software (or provide a replacement) in a 14-year window then that is on that package's community rather than everyone else. The baseline should not be dragged down to accommodate legacy software.

MacOS removing Python 2 support in a minor OS version bump was the wrong way of handling things, but in general you should not expect it to be available on modern platforms or be surprised at its removal (outside of pledged long-term support e.g. CentOS). And you can still use a VM or something, same as you would use DosBox to run old 16-bit software.

> what leads you to believe that software is perpetually maintained and ported to newer releases.

Apple has never hesitated deprecating and then removing apis. Right now they have translation software so you can run x86 apps on their arm hardware but guess what? In two years a release will come out that doesn’t have that anymore and apps that are unmaintained and haven’t been recompiled or ported to arm won’t work anymore.

The only argument there is is that it isn’t too nice to do this in a point release. But it wouldn’t surprise me if their hand got forced by some security issue that couldn’t be fixed. Obviously they aren’t going to make the huge investment required to fix it themselves, it’s dead technology.

> I find it amazing the lengths python 2 loyalists take to keep using it, and the ferocity when it was EOL'd (the second time).

I can see where they are coming from. I like Python, I dislike its old boys club of core team. By EOLing Python 2.7, they have officially stopped mucking with it for good, thus the language is finally "stable" in the sense that it won't keep changing under you while still being called Python. Some people just might happen to like it.