Hacker News Reader

Comment by einpoklum | Hacker News Reader

I don't know what to think about this, as so little is said about the project. @namr2000 asked some very pertinent questions, but I'll ask more general ones:

1. Who is developing this? Is it a one-person project? A group of developers? An organization? If so, is it academic? Commercial?

2. What is the purpose of developing this OS + userland software system? Is it a personal hobby? A demonstration of certain principles lacking in existing OSes (and userland software)?

3. Why try developing everything altogether? i.e. all of a kernel, low-level services/userland code, a graphical desktop environment, and applications? Each one of these seems like a rather daunting endeavor.

4. What standards and what APIs do the different components adhere to? Is any of them intended to be interchangeable with existing code running on existing OSes?

View on HN

I very briefly stalked their osdev forum posts. Looks like a hobby project with a focus on writing (their own?) audio/video playing (and maybe even editing?) software on top of it.

Based on their docs, it looks like it's not POSIX compatible and they seem to have their own core libs for everything. Unlikely that it will ever go beyond a hobby project but its a very neat dive. Years of work. Solo osdev-ers are beasts, that's why they do it all.

They are indeed, but this kind of tragic... I mean, all that skill, that knowledge, the experience - working on something which, almost by definition, very few people will be able to use (because people require lots of things from their OS + apps, so that such a project is extremely unlikely to meet all of those requirements). And even as inspiration for other developers and designers - the solo OS-developer is unlikely to spend the time elucidating their design choices and architecture, in documentation, presentations, talks etc.

And this is in contrast with the solo developer of a library or app, whose work is immediately usable by many even it's an undocumented black box... OSes are a team sport.

> I mean, all that skill, that knowledge, the experience - working on something which, almost by definition, very few people will be able to use.

This sentiment makes me sad. You can do things for fun, y'know? And when someone's just having fun, not hurting anyone, and being creative - that's got to be more than just tragic. It's basically the definition of a hobby.

I would also assume that a lot of these hobbyists are excellent engineers at their day jobs. While tons of the osdev work is going to be specific to osdev/hardware, tons of it will make you a better engineer.

The operating systems course I took in college was foundational to my understanding of how programs work and the memory model of modern computers.

You can me things for fun, but the language on the website is ridiculously bombastic for that: https://www.getxeneva.com/

> OSes are a team sport

The comment you’re replying to mentions osdev. If you’ve ever been an osdev-er, it’s more about personal growth and understanding OS development for yourself…

We are still allowed to do things for fun, right?

Well, yes, fair enough, but - if a multi-person-year project is just for fun and almost unusable by others, then - I feel it's a shame.

Regarding point 3, having the entire base system developed as a single project has worked well for the BSDs. They are much more consistent than Linux distros and much better documented (although props to Arch on their how-to style docs, they're unmatched). Spend a bit of time with them and Linux starts feeling kinda janky.

> BSDs are [...] much better documented

This is an opinion that I've often read, but that does not match my limited experience. The difference in quality between BSDs and Linux is not clear. I've found outdated documentation on both sides, though BSDs mitigate this by being more stable (i.e. less innovative, if you prefer looking at the dark side).

For instance, a blog post was recently describing struggles with the NetBSD installer. It complained that the documentation chapter about installation was 7 years old, and about obsolete releases of the OS and its installer. https://eerielinux.wordpress.com/2025/05/31/installing-bsd-i...

Another example, this time with FreeBSD. The documentation still has a section about floppy disks, and the chapter about "Linux Binary Compatibility" is for Linux 3, about 10 years ago. Hard to tell if these pages are still valid.

I can't speak for NetBSD's install guide, other than I didn't have trouble with it when I last installed it. Also, floppies are still relevant for the BSDs, since a lot of people use BSDs in retrocomputing.

Regarding "less innovative," I suspect you're just unfamiliar with what goes on with FreeBSD and OpenBSD. There's lots of innovation there. NetBSD stays simple and traditional, but that's an intentional choice.

FreeBSD has a policy that they don't accept undocumented changes. If you add or modify a feature in a program, you have to update any relevant man pages and also the handbook if necessary. I assume NetBSD and OpenBSD have similar rules. The man pages are where the BSDs really shine - the difference is night and day.

> having the entire base system developed as a single project has worked well for the BSDs.

Well,

* I don't believe BSD develop their own GUI and apps. So, just the "base system" like you said.

* BSD is a 60-year project, started by a university, and taken up by many individuals and organizations, not a personal hobby project.

Reply on news.ycombinator.com