Ronald G Minnich <[EMAIL PROTECTED]> writes: > On Thu, 29 Nov 2001, Adam Agnew wrote: > > > Wait, woah, hold up. What's going on? Could you please elaborate on what > > you're talking about here? > > sorry, adam, I misunderstood you too. So forget that, but ... > > > Where does efi want a fat-32 partition? as the first partition of your > > hard drive? Do they expect the unix boxen to conform to that requirement > > too? > > yessirree. Part of the "contributed technology" for EFI is microsoft's > FAT-32 drivers (binary, of course; why didn't they use an Open Source one) > and chkdsk (binary, of course ...). (Source: EFI spec and some slides at > intel.com). > > So, yes, Intel's plans for EFI including FAT-32 support in the BIOS and > running chkdsk on same. They either FAT-32 format FLASH or a disk > partition. Elegant, huh?
Wow! I'd forgotten about the required FAT32. > And you wonder why I thought motherboards would have enormous FLASH :-) > > EFI is such an amazingly bad idea ... Hmm. I don't think the idea of putting an OS in flash is bad. Nor do I think putting a filesystem in flash is bad. And not having to search for a dos bootdisk (DOS == EFI) appears o.k. as well. Not using a journalling filesystem, that is ugly. For ages x86 has been the expediant system that works instead of the well designed system that is too expensive. With Itanium intel is attempting to change that. My basic metric for telling how good a design is by comparing the requiredments list to the feature list. A nasty design will have more features then requirements. An uninspired design will have the same number of features as requirements. An elegant design will have many fewer features than requirements. Though perhaps I should say codepaths and not features. EFI appears to be a fairly uninspired design. And because it has so many requirements, and features. No one has been able to look at everything and make certain all of the pieces are nice and clean. Eric
