What was your most challenging accomplishment in osdev?

[VolTeK]

Hello, very soon i will be done (for now) with writing my fat12/16 driver. After 150-350 lines a night (assembler), It at the moment is a massive portion of my operating systems code, hopefully soon to change (i have yet to redo my memory manager, i have better ideas).

However, i would consider it to be my biggest (current) accomplishment so far in osdev, and with much work and effort, it is almost done. (So big to me, that i backed everything up, and do so every night).

So i ask, what was/is your biggest accomplishment in Operating system development

[bluemoon]

My biggest accomplishment is an "Hello World!".

The kernel loads disk, file and display drivers, then scheduled two ring3 processes to run two program printing "Hello World" by calling newlib then invoking INT80.

It was way lot more of work for an hello world than I thought when I started.

[piranha]

Getting my OS to be self-hosting. Well, mostly so. I had to write a script that my shell could run that would compile and link the latest sources of my kernel, and then copy the kernel to the right place on the drive manually because make doesn't work yet on my OS. gcc did, and ld, so I figured I'd give compiling the kernel within itself a shot. It worked flawlessly. I mean, it took about 10 times longer to compile, and I had to manually update the hard drive with the new kernel, but once I had the HD image with the self-compiled kernel, it worked exactly as the kernel that I compiled on linux did. I literally jumped for joy when it booted.

I'd say that was the biggest accomplishment because it took the entire life of the kernel to get to that point.

-JL

[Jezze]

Writing a OS with quite alot of features without using any type of malloc.

[thepowersgang]

I would say mine is getting ASCII Star Wars streaming properly (telnet towel.blinkenlights.nl).
That took the full TCP/IP stack, and ended up exposing bugs all over the shop.

[turdus]

Designing and imlementing a new object format and ABI (including stack-overflow free calling convention). Writing the compiler and dynamic loader for it was a hard task too.

[Combuster]

The most interesting one is probably running custom firmware on a 3D graphics card. I would add UDI as well but it's not technically complete enough to label as an accomplishment.

[shikhin]

Hey,

...new object format and ABI (including stack-overflow free calling convention)... the compiler and dynamic loader for it...

Any link to the specifications?

Mine would be perhaps remaining motivated enough to continue to work on my OS.

Regards,
Shikhin

[OSwhatever]

The challenging part is not that any individual part stands out being the most challenging. What's challenging is the scale of an OS project. There are so many things to be done and so little time.