OSDev.org

My guess is that 99.99% of the folks on here are building an OS from scratch "just for fun" or because they have a unique idea they want to test out.

Yet a complete OS is a mammoth effort - few of us have the time and resources to build something that could compete with Linux, for example.

But my question is, have any group of OS designers on the forum ever tried to get together on a serious project? The downside is, of course, a lot of things will already have been done, but the upside is a lot more progress.

Several posters have requested people to join their "team". I'm not sure that many have had any success.

My take is that I play with this sort of thing for fun; I've no interest in diluting that fun by becoming a cog in a machine. If I did want to join a multi-person project I'd choose one of the mature ones - such as Linux or FreeBSD - rather than an unproven new project.

Collaborative projects will mean you need to compromise on how stuff is implemented, and unless you have mainstream ideas in that area, this will become a burden. And if you have relatively mainstream ideas, then you are better off joining one of the more mainstream projects like Linux as those are built on such ideas.

rdos wrote:Collaborative projects will mean you need to compromise on how stuff is implemented, and unless you have mainstream ideas in that area, this will become a burden. And if you have relatively mainstream ideas, then you are better off joining one of the more mainstream projects like Linux as those are built on such ideas.

You're exactly right.

0b1 wrote:My guess is that 99.99% of the folks on here are building an OS from scratch "just for fun" or because they have a unique idea they want to test out.

Yet a complete OS is a mammoth effort - few of us have the time and resources to build something that could compete with Linux, for example.

But my question is, have any group of OS designers on the forum ever tried to get together on a serious project? The downside is, of course, a lot of things will already have been done, but the upside is a lot more progress.

A very few projects here had contributors, unfortunately. Rdos explained well why. Also, I have the perception that everybody here wants to create their OWN project, and very few would agree to just learn kernel development by contributing to other people's project. In other words, the goal often seems to be more about creating a project, than "just" learning. I don't know, that's my perception.

For example, can I ask what's your goal? Mostly learning or mostly developing something?

Agree with every word! I feel the same way, but I wish there was more than one of me, or more time, or something.

I guess we do learn from one another to a degree - in 2018 the BareMetal source code taught me a lot about keeping things simple.

The goal of my OS is to be a compact application server. No desktop, no UI, and specifically written for VMs, with web and data services and not much else. I think there is a market for an application server that will launch 20 instances in the same resources where only two or three windows guests, and maybe 8 Linux guests, would run. Whether that will still be true in 20 years when I finally finish it, we'll see

0b1 wrote:Agree with every word! I feel the same way, but I wish there was more than one of me, or more time, or something.

I guess we do learn from one another to a degree - in 2018 the BareMetal source code taught me a lot about keeping things simple.

The goal of my OS is to be a compact application server. No desktop, no UI, and specifically written for VMs, with web and data services and not much else. I think there is a market for an application server that will launch 20 instances in the same resources where only two or three windows guests, and maybe 8 Linux guests, would run. Whether that will still be true in 20 years when I finally finish it, we'll see

That's an interesting idea, actually. I've talked about it with other people in the past. While it's true that today services run in containers more often than directly on VMs and that the Linux kernel could run, per se, with a just a few MBs of ram, a whole Linux system able to run an HTTP server is pretty bloated. A stack of kernel + userspace designed for that purpose could theoretically do a better job, but it would require an incredible amount of work to achieve that.

Anyway, I just didn't understand if you would make some compromises to achieve the "compact application server" goal faster by joining/forking another project and pushing it in that direction or it's part of your goal to design and write everything from scratch.

EDIT: is that related with the "cooperative multitasking OS" you talked about in another post or you have ideas for two different operating systems?

0b1 wrote:The goal of my OS is to be a compact application server. No desktop, no UI, and specifically written for VMs, with web and data services and not much else. I think there is a market for an application server that will launch 20 instances in the same resources where only two or three windows guests, and maybe 8 Linux guests, would run. Whether that will still be true in 20 years when I finally finish it, we'll see :lol:

Take a look at includeos.