GitHub stars won’t pay your rent

> But I just. can't. stop. [...] I've tried suppressing the urge to create, but I just can't do it. Ideas, challenges, problems, unmet potential, are everywhere, and I can't unsee them. On many days I despair that I won't live long enough to build even a tiny fraction of the things I see in my mind. [...] my idea list just keeps on growing faster than I can keep up.

If you really want to stop, well, I can tell you what worked for me. :/ Basically, burn out on it. Leverage that despair to turn aside. Increase it by focusing on things whose absence badly harm people, but where external constraints make progress largely intractable. Unripe fruit; unproductive research career planning. This degrades the reinforcement of little successes. Disengage from the list. Make it something that hurts to look at; don't review it; touch it less often; set it aside; forget it. Keep two or three open-ended and intractable projects around to serve as distraction from the rest. Incrementally pull back on seeing opportunities. Sort of a twisted mindfulness: see them swarm by; acknowledge them; let them go; forget them. With time, they become less in-your-face noticeable. Let's see... Make supporting past work a resented burden. Have stuff you enjoy doing and "care about" to absorb the displaced attention and effort. And so forth. Attack the system. So the hose can be turned off. Or at least down to dribbles. Just perhaps be clear on what you hope to accomplish by doing so. "This hurts so I want it to stop" is perhaps an insufficiently nuanced objective. FWIW.

Well for this problem I have since a long time an idea I am sure will works, I just need to kick-start it in the right way.

It is the "fair publishing edition", a system of publishing a work in a way that is fair both for the author and for the users. The author get paid an amount that is proportionate to the amount and value of its work and the users gets a fair value in return from their money.

The principle is simple, the author publish a work and decide a fixed amount of money that he judge correct as a fair total reward for his work. In turn the people that begin to buy the product pay only a small quota of this amount and the more people buy the product the smaller gets the amount. If the product becomes successful eventually the entire amount of money will be paid and the product will become free for everyone.

A detail, the first people will pay a larger amount of the latest people and this is a little bit unfair. Yet this is a minor problem that can be fixed by not making the payment immediate and when other people buy the product the payment of the first people will be reduced. In this way the problem will not disappear but it can be smoothed so that it will become more acceptable.

The faire publishing edition can be applied to any product that is trivially copiable, so for things like books, software, games, movies etc.

Each product launch will be caraterized by the amount requested and by a paying time which is roughly the time the author estimate to complete the payment of the product.

Note how this is much better than the current situation: free software authors doesn't get any reward for their work. It is unfair for the authors! At the other side traditional editor get indefinitely paid for a work, doesn't matter how much money they have already collected. It is unfair for the users!

If some people want to help me to kick-start this project, hey, just make me a message! :-)

The dot-com bubble crashed in 2001, which left the entire tech sector in a recession throughout the first 2000s decade. The foreclosure crisis started a recession in 2008 which really didn’t end until 2015 or 2016.

This is why all my projects are published under a non-commercial-use-only license. Hobbyists are free to use it, but companies need to contact me to license and pay for using my code. Makes me very good money, and my code is still out there for free for hobbyists to use.

I've less repositories, ~30 repositories. 5 or 6 of them I would consider significant projects at least from my point of view (aka a lone developer who is actually more of a sysadmin/devops in his day job, so not it's not really my main trade). I've yet to reach a project with 100 stars (even if two of them are getting close \o/).

And indeed, I never got or even tried to make money out of it. I already fell kind of a moral obligation to answer to tickets, sometimes emails, and exceptionally phone calls (actually happened only once), if someone gave me money for a project I would feel even more morally obligated, so that's why I don't even try. Also, my projects don't exactly have a success in the thousands of starts and users, so it's kind of hopeless.

However, contrary to you, I'm far less active in the last few years (contrib per years, 2014: 491, 2015: 1228, 2016: 1095, 2017: 646, 2018: 225, 2019: 288). Basically, I lost a good chunk of my motivation, not sure why however. Either I grew tired of it, or my current job (since 2016) drain me of my motivation, or seeing tickets accumulating in my half abandoned projects discourage me, half thinking, "let not create another ticket generator" and "I should really fix this project before starting that new project... (and in the end doing none of those)".

It's not the lack of ideas, I have actually a few from small to quite big:

* web ui in JS to calculate production in Hearts of Iron IV (actually started that one, and also a good opportunity to learn frontend development)

* a dead simple and as transparent as possible artifact manager

* a small daemon to name your NIC because it's on a given network (think servers with several NICs and dedicated prod/admin/backup networks, it would far nicer to have them named eth-prod/eth-admin/eth-backup instead of eth0/eth1/eth2 (and sometimes eth1/enp0s3/eth0 because this special duck is wired differently).

* a C clipping library (I need one for a current project).

* a tsdb taking ideas from rrdtool, like fixe size DB, values computed at insert time, reducing resolution as time passes, round robin, with meta data like the unit attached to the metric, with a network endpoint, and some support for a multi-node setup, all that written in Rust (because I don't know the language yet).