It would be really handy to have VM software just "built in" to the BIOS and chipset. You could do some things at this level that are just impossible at higher levels, because you know some of the hardware you're sitting on at dev-time.

I don't know about the rest of the features of this thing, but the "virtualization at the firmware level" thing has potential.

So much potential in fact that mainframes have worked that way since the 1960s ;-)

Now coming to a laptop near you ...

Microsoft also seems to be heading that way, now. Perhaps too late - they have an extremely heavy 'backward compatibility' load, carried so far (maybe too far) on a single code base.

Yay. Now we have an OS hard-coded into the BIOS. Why is this a good thing?

What does this buy over simply putting a flash chip on the motherboard and letting the OS treat it like another peripheral, perhaps with a pre-installed slimmed linux distro?

This has the potential to be very cool. Basing it on Linux would dovetail nicely with the FSF's Campaign for Free BIOS as well:

http://www.fsf.org/campaigns/free-bios.html

I'm sure there would be something proprietary in it that would ruffle the feathers of rms, though.

The article doesn't address it but what effect does this have on hardware visibility to the guest? (How) can Windows still get unfettered access to e.g. the video chipset or other expansion devices?

It's like going back to the days when ROM BASIC was built into every "microcomputer." Only this time they're going for a full OS instead of just BASIC.

I guess the question I have for this is very similar to the question Bill Gate's mom asked him when he said he was working on personal computers.

Why would anyone want this?

Title borrowed from here:

http://linux.slashdot.org/comments.pl?sid=1233739&cid=27...

which summarizes well the potential of this "OS in ROM, with hypervisor for VMs" approach.