Chiming back in to say that, while everything you're saying is correct (i.e. Python was not a ubiquitous language until "relatively" recently), it doesn't change the point: the best packaging solutions, done right, need to be done early in a language's history.

To illustrate the point with an example, you could invent cargo for python yesterday or in 2005, but it wouldn't have solved the problem, because you would still have decades-worth of third-party libraries that wouldn't comply to py-cargo's packaging requirements.

In contexts like these, it's the package manager with the fewest hard-asks (i.e. pip, or npm for node) that wins.

Go, for example, endured major controversies over migrating away from GOPATH-managed-with-third-party-dep-managers to go modules. Even though `go mod` would have been the best solution to start with from scratch, inertia and breaking changes are a real thing.

Where is gen220 rewriting Python history?

Rust is a pretty obscure language now in pretty much the same way that Python was an obscure language then.

Of course the world of programmers was smaller in the 1990s. But if your baseline is the entire world, then probably every programming language outside of Basic, C/C++, and Pascal was obscure in the 1990s. Just like Rust is now.

It feels very much like you have shifted baselines to determine what "obscure" means.

From my view, Python's popularity took off around 2000. That's when I no longer had to tell people what Python was, and when people in my field (cheminformatics) started shifting new code development from Perl to Python. It's also about when I co-founded the Biopython project for bioinformatics. And SWIG in the mid-1990s included Python support because Python was being used to steer supercomputing calculations at LANL.

So your statement that Python's popularity and use in science in general started only in 2010 sounds like revisionism which distorts the actual history with an artificial baseline.

You wrote "with a few web frameworks which were not that popular".

Ummm.... what? Zope was quite popular. The 2001 Python conference had its own Zope track, and the 2002 conferences felt like it was 50% Zope programmers.

Quoting its Wikipedia entry, "Zope has been called a Python killer app, an application that helped put Python in the spotlight". One of the citations is from 2000, at https://web.archive.org/web/20000302033606/http://www.byte.c... , with "there's no killer app that leads people to Perl in the same way that Zope leads people to Python."