Posting a Wikipedia link without commentary or context is essentially a text meme. It doesn’t invite discussion.

What are you trying to say?

(Not that I disagree with your point, but it’s unclear.)

This was not my intention —

But I enjoy the irony.

Invoking the concept of “thought terminating cliche”… as a thought terminating cliche.

> Posting a Wikipedia link without commentary or context is essentially a text meme. It doesn’t invite discussion.

You literally just replied and we are engaging discussion — empirically the wiki link above has done literally the opposite of what you are saying! At any rate, the intended relevance of the wiki link is the following.

When something bad happens to someone, they tend to overestimate how bad it is relative to something experienced as objectively equally good; it is thus important to be aware of such a bias — in addition to using regular old critical thinking.

In this case: “although this is a ‘migration’, what do we mean when we say that? and is it really the same thing as py2 to py3?”

I would tentatively propose “no” to the latter question; and would additionally propose that calling this a “migration” is not terribly useful as although it’s not incorrect, it’s insufficiently specific and seems to invite a category error.

In addition to the loss aversion bias mentioned above.

Category error: https://en.wikipedia.org/wiki/Category_mistake

What are some good examples of meaningful migrations of similar scale that went better?

Objective C to Swift.

.NET Framework to dotnet core

Maybe I am off base here, but big migration efforts feel like they would be significantly easier in a compiled language. Potentially not a fair comparison when the tooling can automatically push so much code without errors.

This pep only affect the C world. It will not require change in Python code.

No it wasn't and not by a long shot.

What's your top three?

He, the Perl 5 to 6 migration is already forgotten? I think this says something about how well it went, because at the time Perl was more popular than Python.

There was no Perl 5 to Perl 6 migration. Perl 6 was announced, a bunch of design work happened on it, and then it became a different language run by different people rather than a version of Perl. People are still writing Perl code extensively, and Perl 5 is still maintained.

I would expect Python to follow a similar language split. People were disappointed at how few breaking changes occurred in 2->3. If there was going to be a 3->4 migration, will be a large number of proposals to correct other deficiencies in the language.

"How bad was the migration?"

"It didn't happen"

It didn’t happen, but it still killed the language.

... I mean if you want to compare that it had exactly same problem so it is another evidence it's not something specific to Python, it's just utterly terrible way to go ahead.

Only difference is that they noticed it's a bad idea midway and went back to developing Perl 5

And I still can write "use v5.8" in my Perl script and write script that works in Centos 5 instance from 2007, if I needed to do something on legacy system, and it will work the same on latest Perl.

> because at the time Perl was more popular than Python.

I don't believe that was true at the time as P6 took like decade to decide on semantics and only then the implementations started to happen. More existing code, sure, but much of the web dev moved to PHP/Ruby and many other things to Python.

And unlike Py3 it was initially also much slower than P5, making migration kinda worthless.

A lot of old Java people have recently got into Python, and enjoy the types of py3