I've got to say that I'm pretty shocked at the hubris here.

The NetBSD project has been busily porting our system to every
architecture on earth -- we've got over twenty separate architectures
working at this point, compiling out of one tree, sharing nearly every
line of code they can. Other than the HP PA (forthcoming), we support
every major architecture there is, and most of the ones that anyone
has cared about back to the Vax.

FreeBSD has ONE architecture up and fully supported. You have the
i386, period. You've never finished a single port to another
architecture.

How are you benchmarking your claimed performance issues on more than
one platform? (Where are your benchmarks even on just the i386?) How
are you checking the portability issues we mention out? On the basis
of what ports? Hell, you guys had to take OUR kernel as the basis for
the not-actually-working FreeBSD/Alpha port.

I'm not going to have the balls to claim that the stuff you guys do
better (like user interface issues) is stuff we know jack about. The
way you learn about things is by actually having experience, and we
don't have that experience.

Where, though, do you get the cojones to claim you know as much as
people who live, eat and breathe multiplatform portability, when
you've never done a single solitary port? How do you claim performance
benchmarks when you have no way to test this on more than one machine
type? How many architectures have you bootstrapped from scratch with
this design? Is this yet another sex manual written by a priest and a
nun?

Our design didn't spring from the air. It was slaved on over well over 
a year, with people applying the results to real machines, and then
altering those results based on the real problems they experienced --
not based on theory.

Where are your numbers? Where is your code? Where are the dozen
architectures you run on so you can test your theories?

I'm pretty flabbergasted. You've lost a lot of brownie points with me.


Perry