I think we need to define a Standared Driver Interface for hobby asm OS's, so we do not need to reinvent the wheel.

I will post a example soon.

IMO "Hobby asm OS" and "not reinventing the wheel" don't mix very well.

That why things need to change, take fasm as a example, because the great foresight that Tomasz Grysztar put into it, its the most portable assembly theres is on x86.
Took me half a hour to port to my OS.

But, when Tomasz made FASM, he re-invented the wheel, making a better assembler. Just like the wheel, I know I prefer my cars rubber tires over round rocks.

However, I think there should be a driver standered! I think it should go as far as a plug and play driver, for example, I take a driver from DexOS and it works, and you take a driver from my OS and it works, with no modifications to the driver.

Actually, if it were up to me, the BIOS would contain drivers for your computer, for every mode you were in, then, any computer with this feature could run any OS at full that supported it.

Well, personally, I'd rather reinvent the wheel. That's kinda how assembly programming is these days...

If you wanted portability, try using C or BASIC...

I like the idea that a driver could inter-operate between OS but it might be difficult to accommodate the range of designs. I'd be interested in what you have in mind, as I'm coming close to thinking about serious driver implementation for my own OS.

EDIT:
The interface could be language independent anyway, which allows the OS / drivers to be implemented in their respective language(s) of choice.

Hi Coty,
We already have a standard interface - the BIOS. As crusty as it seems, many start out in RM16, it's just as easy
to use the BIOS in PM32 with a BIOS extender, see my port of MikeOS. 64bit long mode is also doable
(I just haven't gotten around to it yet).
There are plenty of examples of PM32 code which replaces BIOS functions which could be reworked to conform to
the BIOS standard. If enough hobby OS's used the BIOS it would be simple enough to test the new code in
working systems for development. That tested code could then be integrated into a persons OS without the BIOS
interface.
It's faster to write a small OS and then a larger one than it is to only write the larger. Many don't get much further
than the A20 line when it comes to PM32.

I agree so let do its.
To work with all OS's its needs to
1. be able to hook into int's
2. be relocatable
3. be self contained.

For 1. we need to agree on a int number that when called, lets us hook into the idt.

For 2. Need a simple format thats just a simple header, relocatable offsets and bin file.

for 3. For that, we just use simple reusable functions that are included in the drivers.

What you intend to do is somthing like a Cross OS DDK API?
somthing like FASM LIB but for Drivers?