Back to the osdev thread!

When i read documentations and tutorials on writing an operating system and the grub manual i came across a new thing called the multiboot standard. It said kernels are meant to do a lot of work and worrying about the bios and about how to boot the system should not add to the complexity of the kernel... quite right!

I agree that writing a bootstrap would help us better understand the operating systems, but making your kernel boot is not writing an OS. OS has lot more things to do, lot more algorithms to code and a huge number of decisions to make. Then why make developing an OS even more complicated by trying to implement a custom bootstrap - unless you want to prove to yourself and to the world that you can 'code' a bootstrap.

More about the multiboot standard is available at http://www.gnu.org/software/grub/manual/multiboot.

btw... there are more reasons why anyone would want to avoid a custom bootstrap! I prefer coding in 'c' to 'assembly'. The kernel that i am interested in is a very simple one meant for understanding the issues involved in implementing the algorithms. Above all, when i always wanted to avoid the E820 and E801 stuff - keeps my kernel easy to understand.

If E820 and E801 seemed strange to you, check out the Ralf Brown's Interrupts List - interrupt number 15.

Why would anyone want to write a new operating system?  Am not sure of why others want to write, but following are the reasons that interest me.

• I love systems programming
• Existing projects are already too mature for me to contribute anything.  The most appealing is the linux kernel, but there is hardly anything left for me to do
• Its always exciting to say "I use my own operating system".
• World Dominance :)

Need more reasons to write an OS?  Take a look at any of the famous osdev websites.  Google would help you better in getting the links to those sites!