←back to thread

Writing an operating system kernel from scratch

(popovicu.com)
304 points Bogdanp | 4 comments | 14 Sep 25 15:44 UTC | HN request time: 0s | source
Show context
lordleft ◴[14 Sep 25 16:28 UTC] No.45241113[source]▶
I will never not find this kind of project incredibly impressive. It’s interesting to think that Linux, after all, is really just the kernel — and yet getting that work done paved the way to getting an open source version of Unix installed on billions of machines. Great stuff!
replies(4): >>45241149 #>>45241727 #>>45241785 #>>45241926 #
hollerith ◴[14 Sep 25 16:34 UTC] No.45241149[source]▶
It is equally valid to say that Stallman's starting to write a C compiler and Unix utilities (in 1984 whereas the Linux project started in late 1991) paved the way to getting an open source version of Unix installed on billions of machines.
replies(2): >>45241551 #>>45242627 #
1. kimixa ◴[14 Sep 25 17:24 UTC] No.45241551[source]▶
I agree - there's a number of kernels that were "open source" and released at a similar time enough time to linux (e.g. 386BSD in '92) that I could see any of those winning the "community battle" and taking that space instead, but no real credible "development toolchain" equivalent until decades later.

Though I'm unsure how differing licenses might have affected this - I suspect that really early in it's development the "copyleft" nature of the GPL Linux didn't make as much of a difference, as from what I remember most commercial uses of Linux didn't come until it had already gained significant momentum.

replies(2): >>45241601 #>>45241723 #
2. ◴[14 Sep 25 17:30 UTC] No.45241601[source]▶
3. treyd ◴[14 Sep 25 17:44 UTC] No.45241723[source]▶
The copyleft nature was essential to good driver support. It set it up such that for corporations making drivers the easiest path was to get the driver upstreamed. There was a bunch of hoops they could have gone through to avoid that (as many did, like Nvidia) but that became a sorta-default.

Copyleft encourages a collaborative relationship between entities because it makes trying to play it close to the chest with IP involve more legal effort (if it's possible at all).

replies(1): >>45241813 #
4. kimixa ◴[14 Sep 25 17:54 UTC] No.45241813[source]▶
Yes, I can see that stalling development as (at best) it turns into a pile of private forks rather than a cohesive project, but from what I remember that was already after Linux had "won" the "Open Source Kernel" race.

Commercial support for Linux was... Sparse... before the early 2000s.