OneAndOneIs2

A recent post on the LinuxBIOS mailing list highlighted that some recent developments in the world are looking very promising in the BIOS world.

The BIOS is the thing that gets your computer up & running when you turn on the power: You might think that the OS does that, but how do you think your PC knows that it's supposed to boot an OS?

It's an important function, and the BIOS has served well for a number of years. But it's not without its problems. Most BIOSes are proprietary, they're fairly slow, and all-too-frequently they're buggy - due in a large part to being written in hard-to-maintain assembly.

The LinuxBIOS hopes to solve all these problems by creating an open-source, 32-bit, C-coded BIOS replacement, based on Linux. The biggest problem they have is getting the necessary hardware support: Linux has many many hardware drivers, but they only know how to use the hardware once it's up & running: Getting them that far is the BIOS' job. Spot the problem.

So the first thing of note is that Intel is getting involved with the OLPC. Intel's been doing pretty well with open-source drivers etc. lately, but not so much when it comes to the BIOS area, where they're pushing their own replacement, EFI.

If Intel truly wants to benefit from the OLPC, they'd need to get their hardware into it. To do that, by the OLPC's rules, they'd have to release source code for it.

And if they release the source code, the LinuxBIOS devs can of course use it in their own project. So from that point of view, the more Intel puts in, the better - an open-source EFI could be very useful...

The second is Dell. Having recently started selling Linux machines, they already see enough enthusiasm that they're intending to expand and thus they have two things to think about: Firstly, proprietary BIOSes often present problems for Linux; and secondly proprietary BIOSes are actually quite expensive and bump up the price of the hardware.

Both problems are solved with an open-source BIOS. So we see that replacing the proprietary BIOS with LinuxBIOS has actually made it into Dell's top 10 in a list of priorities.

If somebody like Dell throws their weight behind LinuxBIOS, hardware manufacturers would start falling over themselves to get their hardware supported: Lack of support means lack of Dell selling your stuff.

So there we are, the tipping point where the clapped-out, primitive BIOS gets replaced with a superior open-source alternative is getting ever-closer.

And the main advantage here is that if the hardware is supported well enough for LinuxBIOS to initialize it, then it's supported well enough for Linux to run it. Which means that the FOSS-only machines we're all waiting for are also getting ever more likely to be made at last.