> 50% of Python developers have less than 2 years of professional experience
That seems insane. Do juniors start with Python but then don't write it again once they grow up? or is it just who fills out responses to a survey like this.
Or is Python just growing that rapidly? (There was a time when computer programming as a whole was growing that fast. Is Python there now?)
Or is Python picking up lots of people whose job isn't actually programming? They want to do data science, or biology, or whatever, and they need to do something with some data, and it's too complicated for an Excel spreadsheet, so they reach for Python. But when they get the data mangling part done (or at least working so that they don't have to keep fiddling with it), they quit being a programmer of any kind, and go back to doing whatever it is that they actually do. If people like that were, say, half of the Python programmers, and they only did that for a couple of years each, you'd see exactly this result.
> is Python picking up lots of people whose job isn't actually programming? They want to do data science, or biology, or whatever
Important figures in the community seem to believe this is the case. Which explains a lot about why the packaging discourse goes the way it does; there's a fair bit of emphasis on serving the needs of people who are just not going to be interested in contributing packages that work with the existing packaging system, and just want their friends to be able to run their code using existing third-party packages, and don't necessarily want their friends to even have to know what Python is. And then there's the concern for people who do everything in "notebooks" and have never created a .py file from scratch. And so on and so forth. And there's a looming sense that we don't even have a good idea of what the most common existing workflows look like, outside of the ones that were explicitly designed for the needs of "traditional" developers.
If you're familiar with the official Discourse forums at discuss.python.org, I recommend rummaging around the posts of Paul Moore (one of the main devs for pip) a bit. You'll get a feeling for it pretty quickly.
Huh. The most common existing workflows aren't necessarily those of traditional developers. Or, stronger, the most common needed workflows - the ones that aren't addressed yet, but there's a need for.
The problem is, it's also hard to do research into what other users actually do, or what they find useful. Above and beyond the usual pitfalls of trying to get design advice from end users, you don't even know how to find and communicate with everyone. Certainly they don't generally seem aware of the official forum, and weren't aware of the mailing lists it replaced (heck, I didn't know about the forum until years in, having previously tried using the mailing list!), or the many relevant GitHub issue trackers, etc. So the next obvious thing is to try to post solicitations for feedback somewhere else — but where?
That seems insane. Do juniors start with Python but then don't write it again once they grow up? or is it just who fills out responses to a survey like this.