I'm not convinced that putting most of an OS into the bootloader is a good idea. Particularly when that OS is closed source and written by the same geniuses that write BIOSes.
(a) TianoCore is only a small part of the code (the clue is in the name "Core")
(b) The bits of TianoCore source that I've read are overcomplex and ugly. There's no reason for a bootloader to even have all this stuff.
(c) That's not the UEFI implementation I have on one of my servers which AFAIK is completely closed source, and does all sorts of weird stuff when I boot. I reverted the machine back to plain BIOS "boot the first sector" booting.
a) As far as I know, EDK compiles to a working bootloader on supported platforms.
b) Not knowing what exactly you've read, I don't know if I'd agree, but anyway, it just implements the UEFI spec.
c) As far as I'm aware, all UEFI implementations are built on top of TianoCore. But as it is BSD-licensed, the OEMs don't have any obligation to release their source code.
I really think that UEFI is superior to BIOS in every way. It could also be open, except that hardware manufacturers choose to keep it closed. Not allowing user keys for Secure Boot is, also, a decision of Microsoft and manufacturers, not a problem with UEFI itself.
"supported platforms" being a couple of emulators. That is, without any need to do hardware initialization.
Tianocore "just implements" the UEFI spec, but that's what is overcomplex and ugly. Also, Tianocore picks the ugliest possible way to implement things (or at least, they really try to).
