At least at this point, none of those options will work for us.
Any option that involves changing the licensing or creating proprietary extensions simply isn’t an option because that’s not something we want to do (and given most things we do are based on other projects’ work, we couldn’t change the licensing even if we wanted to).
Selling services obviously works only if you have users who are both able and willing to pay for them. This is viable if you have tens of thousands of users, some of which with sufficient money at hand – but it doesn’t work with a very small user base especially if almost all of those users are home users who couldn’t afford a service contract even if they needed and wanted it.
Selling merchandise similarly requires a bigger user base. If we sold 5 T-Shirts and a mug or two, that would be more than I think we would - but nowhere near enough to hire someone even for a day.
Nobody gives a **** about certification from a project nobody uses, so that one is out.
SaaS doesn’t work for our type of project - I presume we could offer some online storage or an email account or something with the OS, but I don’t see anyone going for that if they can just get the same stuff free of charge in hundreds of other places (and I certainly don’t want to make the OS worse by restricting users’ options by making competing services harder to access or something).
Voluntary donations/Partnerships with funding organizations: We’re doing that. We’re not even getting enough to pay for our infrastructure.
Bounties: Well, of course anyone is free to put out a bounty on adding a feature/fixing a bug/whatever and we’ll gladly accept either the money to do it ourselves or the code submitted by someone wanting to claim the bounty – but so far nobody has put out a bounty on anything.
Crowdsourcing: Tried that and it failed badly. You need way more visibility than we have in order to do that successfully.
There’s one thing I could see potentially working (though nowhere near a level where it would pay for even a part-time employee - if we’re lucky, this would cover our infrastructure cost), and that’s selling hardware with the OS preloaded. That’s actually one reason why I think the ARM/AArch64 ports may be even more interesting than the traditional x86 builds – people can get an x86 box anywhere, but there are few (if any) places where you can get a ready-to-run off-the-shelf ARM or RISC-V box.
I have some nice prototypes right here, but of course they’ll essentially need the 4.0 release given 3.x is x86-only.