Discussion in 'Technical Discussion' started by Revival, Sep 15, 2008.
But will it run like shit like generic Linux graphics drivers on my nVidia card?
Great work so far . Keep it up! You've come farther than 95% other ppl who try to write OS (including myself)
Reminds me of my, kernel deving days. I too made couple of OSes (more of a bootable programs ). My last one called MKX, inspired by Mach 3 microkernel, had some support of SMP as well. My idea was making a multiserver microkernel OS in which, unlike other multiserver designs, microkernel was aware of some of the servers (network transparent IPC server for example, which is built within Mach) and they couldn't be replaced. I had some other ideas as well (like porting nanox but realizing I had to port GCC first then realizing I had to port binutils first!).
I have one advice for you, have patience :v: . Thats needed a lot while deving an OS.
I recently had a look at my kernel, and decided it was very useless. It was badly written, as it retained code from when I was an OSDev newbie. So rather than rewrite the whole kernel again, I've decided to move OnyxOS to being based on an older kernel made by a university in switzerland: HelenOS. This means OnyxOS is now a microkernel OS and also now includes a rudimentary GUI (Using VESA, it's basically text-mode with a mouse-pointer).
Hm, excuse me if I say this, but your decision wasn't good at all. You've got lost in the same wood most hobby OS developers get lost into, rewriting their kernel endlessly because they think it isn't good. Just rewrite (or fix) the parts you think are badly written.
There's no need to rewrite the ones which are working correctly.
You don't have to do an entirelly bulletproof and crystal clear implementation, it's not actually really needed, and it just has to work.
Don't be a perfectionist otherwise you'll never do anything usable or finished.
I was in the Exclaim project in the beginning of this year and at first they were really good at making progess, but they then decided it was better to write a new kernel (after three kernel rewrites), and that made me not check or otherwise return to the project at all, and they haven't made much progress since then.
Trap mindset! If it works and works well, then keep it. Hell, even MS-DOS and Windows have some ugly hacks that, unless you were shown the source code, you'd have never known existed.
Keep the kernel you made, or at least release it so that others can work and learn from it.
IMHO I thought this was awesome as a from-scratch OS instead of a distro of some other OS, or a fork.
Separate names with a comma.