This is off topic from most of the discussion that has been occurring
on the list lately, but in a week or so, I would like to post a link to
the draft of my master's thesis proposal on the device driver framework
if that is fine with the list. I could use input from the
community, since it is something I would like to be useful to the
community.
Any constructive criticism would go a long way in helping me make a
good design for all of you. Marcus gave me a good tip on looking
at EROS' persistent design and to at least consider it if nothing
else. I haven't quite gotten around to reading the EROS document
on single-level stores yet, but I soon will.
I will make sure my master's thesis includes some background on how
Coyotos works, since that is a (potential) kernel we will be using
instead of the current L4. This way I can at least develop a
forward-thinking design plan that can adapt as the Hurd decides on what
to do with its kernel and other things.
Thanks, and these discussions and referenced papers have really enlightened my view of modern microkernel-based OS.
-- William M. Grim Student, Souther Illinois University at Edwardsville Unix Network Administrator, SIUE, CS. Dept.