Results-Oriented Work

["Jody" <jbruchon@nc.rr.com>]

I am seeing a lot of ideas relating to swapping and protected mode support
and far pointers and far pointer relocation and whatnot.

This is interesting discussion but I think we will make more headway if we
take a results-oriented approach to this stuff in the beginning.  Less
talking about whether to implement fancy 286pmode stuff and more coding to
make things work, one at a time.  We have some TODO lists going and I think
that we should start taking care of some of those things.

I got a luggable 256K PC-compatible.  Please code with the 256K machines in
mind :)  Because the 640K machines are not always the norm.

Please code a LILO-like program for ELKS so we can have our friggin'
software boot from a filesystem instead of a kernel floppy or something.
That's the next thing I want to see, because all this kernel work is a
little pointless to me if I can't run it off a filesystem of some sort.
Many 8086 machines have hard drives, so if someone would take the time to
do it, that would be awesome and lead to the possibility of an EDE with
*open source* boot-loader software instead of that binary-only "bootkit."

To review:  we need working and stable code for core stuff.  Let's try to
shift the focus there instead of spewing out huge lists of features to add.
Features are nice but quite useless if the core is unstable.

My opinion.  Please do not attack.

Jody
"w00t!"