[wplug-bsd] Fw: OT: Beastie makes a cameo appearance on apple.com.

[Bill Moran]

"Jonathan S. Billings" <billings at negate.org> wrote:

> On Mon, 2004-06-28 at 19:40, Bill Moran wrote:
> > I think the problem is that a microkernel is too much effort to develop,
> > debug, and maintain, and you really don't gain a lot by using one, unless
> > you're a developer.
> 
> Actually, I believe that the Mach microkernel API doesn't change very
> much, and porting it to a new architecture is probably simpler than
> porting a monolithic kernel.

That's what I've heard ... once the API works.  But I've also heard that
a microkernel API is very difficult to debug when things go wrong.

(I'm no kernel hacker, and this is all second hand rumor, so if I'm way
off base, it's because I only have a fuzzy idea of what I'm talking
about)

> Also, the separation of microkernel from
> BSD interface can produce a cleaner upgrade path.  I believe the biggest
> problem has always been speed.  IIRC, Mach always suffered from too much
> overhead of passing messages between it's subunits.

I've heard that as well.  It appears as though Apple has worked around
that, somehow.  Additionally, the hardware is getting fast enough that
it's hard to care if your kernel isn't quite as fast as the next guy's.

> (for what it's worth, we used to run an OS called Mach which was a
> BSD-like OS with a Mach microkernel, and it ran on several different
> platforms, including x86.)

It's amazing the things that have come and gone.

-- 
Bill Moran
Potential Technologies
http://www.potentialtech.com