hi,
I think, one question we have to clarify is, what OS we want to run above our BIOS.
1. If it will be MS-DOS or Windows, we have to make the BIOS compatible to the standard IBM BIOS. There won't be much room for an own/better API. It would be a very tedious work to copy the whole interface.
2. If it will be Linux, we don't need the I and the O, a simple boot loader which initializes the hardware will do the job. You could think of the Linux Kernel as a very sophisticated BIOS, which adapts the hardware to a comfortable API (glibc) :)
Because the whole project is complex (think of all the chipsets, we had to care for and other hardware differences from an simple embedded 386EX to a Pentium II with PCI and AGP), I don't think, we can manage (1.) in half a year. If we concentrate on a simple, good, flexible boot loader, we could get a running system in finite time. For legacy OS's, I like the approach to load a legacy BIOS with this simple boot loader.
If we later want the I and the O, would it be possible to strip down a (already for many hardware platforms configurable) Linux Kernel? The aim would be then, to sepearate Linux in a hardware dependent and a hardware independent part. The hardware dependent part could migrate to a kind of BIOS, the hardware independent to a Linux Kernel, which don't have to be configured for a special platform any more. Big project anywhere, I think.
Opinions?
Corvin
--- Corvin Zahn zahn@zac.de