"Writer's" design block

[farcas82]

Hello, I have already developed a small project, now I would like to develop something bigger.

If possible, something with an actual design. But I am suffering from a massive "writer's block". I am familiar with designs such as microkernels, exokernels, SASOS, plan9 where everything is a file, capabilities, etc. I cannot decide on what to pick because I want "the best" design out there.

Could you guys be kind and provide me with some cool design ideas of yours? Something modern that would help me gain traction for my project once I complete the MVP. Something you personally wish you would see in a new OS.

I dream my OS to be useful and I worry about investing too much effort into something that will not gain traction, and the effort would be a waste. The OS must be "modern", have both a usable minimal GUI and a powerful terminal, something like Linux with GNOME combo, but smaller in size and 2D-accelerated only.

I plan on a SATA and a NVME driver for storage, AMD video as accelerated 2D graphics, virtualized drivers as alternative (for Ethernet and sound).

I dream big and I want to dedicate most of my spare time to this project.

PS: I am interested in both cool kernel and cool userland ideas. The language will be C or Zig. The userland will not be GNU.

[iansjack]

If you want to dream big, then go for your own dreams rather than asking others to dream for you.

But, TBH, I would put aside ideas of making a better OS that the world will flock to. That way you are almost certainly going to face disappointment. If you are after fame and/or fortune there are easier ways of achieving that aim.

[nullplan]

I cannot decide on what to pick because I want "the best" design out there.

There is no such thing, there are only tradeoffs. The best you can do is pick the best design for the thing you want to build. But I can't tell you what that is, you have to figure that out yourself.

I dream my OS to be useful and I worry about investing too much effort into something that will not gain traction, and the effort would be a waste.

Let me pull that tooth right now: You will not create a mainstream OS. None of us will. Only a very small number of operating systems get used to any significant degree, and for the most part they have all been invented already. Unless you happen to be one of the few people to actually have a brand new idea for a killer feature that cannot be ported to one of the existing choices, it is just not going to happen.

The point of this hobby is more self-actualization than helping others.

[farcas82]

Thank you for your feedback. I do not intend to write a "mainstream" OS in the sense of popular so far. More something like SerenityOS (this project seem to be stalled).

[farcas82]

I'm coming back; I'm thinking now that one reason I'm asking for other people's opinions is the need for design/concept validation. It is difficult for someone who is a lone worker to get feedback on his/her design.