Hurd and microkernels (was: Re: A simple question)

[olafBuddenhagen]

Hi again,

On Thu, Apr 13, 2006 at 10:48:50AM -0400, Jonathan S. Shapiro wrote:

> Building a new microkernel is certainly possible, but it's a difficult
> thing to do. Given that both Coyotos and L4.sec are coming, I do not
> think that this is the best focus of effort for Hurd-NG.

I'm not so sure about this. Throughout its history, Hurd again and again
was waiting for some microkernels to become available. That didn't do it
any good.

Also throughout its history, Hurd's design has always been influenced by
choice of microkernels. That didn't do it any good either.

I think building an own kernel instead, while it might be more work
(though I'm not really sure about that), would allow concentrating on
how *we* want the system to work, instead of following some microkernel
designers' ideas about how systems should be build... Might very well be
worth trying for a change.

(As an interesting side note, I can't think of a *single* example of a
system built on a third party microkernel, that ever gained any
commercial, academic or other relevance -- besides of serving as bad
examples, of course ;-) )

-antrik-