some time ago I’ve come across Mark Shannon’s ideas to speed up CPython. He proposed how this project should be funded, but I think I’ve got a better idea. I’ve posted a different proposal for funding, which you can find here. TL;DR: core devs are mostly working for free. This was fine till now and I’m simply astonished how they made Python the top 1 tool for AI and one of the top 3-5 languages. I just can’t figure out, how Python can keep its momentum. What makes Python special: it’s created by the community for the community. I’d like to propose to set up some crowdfunding to fund CPython development. There’s a lot of people, who would like to drop a buck from time to time. You don’t need to promise us anything. You don’t need to deliver. You’ve done so much and we trust you. There’s Patreon, Kickstarter, Buy Me A Coffee, etc. If we can’t squeeze speeding up Python or other big projects into its development cycle, we could try to fund such projects with our community.
As the Reddit suggested, there’s already a place to donate to Python: https://psfmember.org
I’ve been coding in Python for years and never seen this. I’ve also googled for Python crowdfunding and donations and didn’t see it as well. Maybe we push the idea of donations too little? Maybe moving it to one of modern support/crowdfunding platforms could help?
But it still brings up the question: I’ve read that most of the core devs are working for free. If the foundation is receiving some cash, also from big corpos, we can’t still afford to pay the devs?
I’ve took a look at PSF sponsors and annual reports. The numbers in those also contain funds from the corpo sponsors? I see we get around 5M of those. If that’s what a cropos can do, I think we could do just as much, if we market it wider in the community.
Thanks for bringing up this topic. COVID-19 and cancellation of PyCon US has negatively affected our income. Additional revenue streams like crowd funding could be a viable option for the PSF.
Personally I prefer corporate sponsorship. It’s usually more stable than crowd funding. It’s also more ethical. I don’t want to receive money from an employee that uses Python at their day job. Instead I would rather get money from their employers.
In general it’s also easier to offload sponsorship and donations to the PSF. Direct, personal sponsorship has legal and tax implications. Dealing with W8N forms, German taxes, exchange rates, and international money transfer is complicated and time consuming. Since I spend some of my work time on OSS, I would either have to clearly separate work OSS time from personal OSS time or clear any income with my employers legal department and financial department.
Maybe you misunderstood me. I didn’t want to fund specific core devs. A single piggy box, which goes to PSF, would suffice. As we’ve been discussing this on Reddit, there’s an issue that PSF has little marketing impact and a lot of devs don’t know about it. Corporate sponsorship is great and I don’t want to change that. I just wanted to bring the idea of crowdfunding specific, big CPython projects like removing GIL, which could take a lot of time.
Could you clarify your last sentence? I think I was following what you’re trying to say, but the last one threw me off. Not sure if it’s the grammar or maybe I was misunderstanding your entire idea. Also you’ve mentioned Reddit multiple times, but could you get more specific? Who are “we”? Are you talking about Reddit the company, Reddit users as a whole, or a specific group? There’s so much context missing here.
I’ve donated to Pypy and projects directly, but I’ve never donated to PSF there because of this sentence: “Your donation helps support sprints, meet ups, and community events.”
When I read about “community events”, I think of talks on beautiful code, Django, testing frameworks, docker/serverless, tutorials on CI, and so forth. All EXTREMELY useful, but not clear to me how those things make Python a better lanuage.
I’m not as interested in donating $500 to make Pycon or community events happen as I am interested in core developer summits happen.
I think that @piotr-rarus is on to something when he says there are those willing to donate to CPython development. And I agree that the hurdles of accountability could be lower than Mark Shannon articulated.
I’d be super-interested in PSF giving Mark Shannon some help, and pyston 2.0, and Pyjion, and Hpy. Perhaps PSF is already doing all they can for such projects. If so, I’d be glad to donate to PSF; I’d be Delighted to be assured that PSF is helping to assist the heroic isolated projects resolve these hard problems.
To say in another way, the only “accountability” request I would have of PSF is the assurance that donations are going toward real Python development, rather than community events.