kernel (was: Re: Translucent storage: design, pros, and cons)

[olafBuddenhagen]

Hi,

On Wed, Jan 10, 2007 at 04:09:27PM -0500, Jonathan S. Shapiro wrote:

>    "We'll build our own kernel" hasn't worked for HURD in the past.]

Actually, quite the contrary is true: "We will use an existing kernel"
hasn't worked for the Hurd in the past. To the best of my knowledge,
nobody ever attempted building an own kernel for the Hurd.

I'm more and more convinced that the original Hurd would have worked out
*much* better if there hadn't been this desire to build on an existing
kernel. It could have started much earlier without waiting for Mach to
become free; and most likely it could have been finished much sooner as
well, not being bound to a kernel not formed according to the Hurd's
actual requirements.

I just can hope ngHurd will have more luck if choosing an existing
kernel again...

-antrik-