On Wed, 17 Dec 1997, Jason Evans wrote:

: > The natural conclusion (to me, at least) is that we (NetBSD users and
: > developers, if not core) should offer assistance to Jason, in return
: > for inclusion in the porting process and information pipeline.
: 
: There is another person (Ian Logan) who is now working on supporting
: sun4m.  He has been digging into NetBSD's code, an he probably has
: questions about it.  I'll also be pulling from NetBSD, so if you all are
: willing to answer questions from time to time, that would help out a
: bunch.
: 
: Chances are that enough of NetBSD's code is going to end up in FreeBSD
: that you won't have much problem with moving our changes back into NetBSD. 

Then that begs the question:  why don't you actually help us on NetBSD, and
*then* see if porting FreeBSD is worth the effort?  Read on:

: My guess is that there are three main reasons people are asking for
: FreeBSD particulary:
: 
: 1) Some of them have already implemented systems using FreeBSD.
: 
: 2) FreeBSD has more press exposure (that I've seen), so it's better known.
: 
: 3) FreeBSD's VM subsystem is quite excellent.
: 
: Reason 2) is a poor reason to choose FreeBSD.  1) and 3) make some sense
: though.  I've been told a number of times that NetBSD will be merging in
: FreeBSD's VM within the next year (likely with some improvements even), so
: 3) could go away as well. Little can be done about 1), which is SME's
: primary motivation for porting FreeBSD. 

Isn't 1) just another manifestation of 2)?  If people have FreeBSD installed
on their i386 boxes, what's to stop them from using NetBSD on their sparcs,
since the base hardware is already radically different?  The _name_.

FreeBSD and NetBSD are two different projects, with a rather similar (and in
many cases identical) interface as seen by your typical user and system
administrator.  Only the name is keeping people from accepting NetBSD as an
OS for their SPARCs; I'd bet $100 per customer that these customers don't
even know the fact that NetBSD exists, or don't know that NetBSD is that
close to FreeBSD that they probably wouldn't tell the difference.

But when it comes to the nitty-gritty, inside guts of the code, they are
very different.  NetBSD and FreeBSD both started around the time of
4.3BSD-Net/2, and not long after that point there were two "releases"
available:  FreeBSD 1.1.5.1 and NetBSD 0.8.  Both were very much alike,
albeit different in different ways.  Back then, I was only first playing
with NetBSD on my Amiga, and FreeBSD on a friend's PC.  Neither were all
that fun to use at the time.  :) 

After that point, and in the 5(!) years intervening, NetBSD and FreeBSD have
developed dramatically, and diverged.  FreeBSD took the approach of faster
feature integration, at the expense of supporting only one platform.  Now,
as you're about to embark on a porting project, you'll see how difficult
that web will be to untangle.  Just ask the RiscBSD folk. 

NetBSD took the approach to a slightly slower feature integration, in the
name of machine independence.  The result was a 64-bit-clean (Alpha port),
endian-independent, well user supported OS where only a handful of code is
dependent on the hardware or processor. 

Don't get me wrong; I'm not bashing FreeBSD or trying to start an OS war. 
FreeBSD is a great platform, and has many things NetBSD still doesn't have
(easy boot-time kernel configuration, better system installer, a little
better VM...); however, we are getting there.  NetBSD does have something
that FreeBSD doesn't, that helps you significantly:  a stable, fully
functional 32-bit SPARC port.  (And some 64-bit framework; perhaps the
people involved in this can talk to you privately about it.)

: You're right that NetBSD would be an easier starting point.  Even a number
: of FreeBSD users pointed this out to me at an early stage.  Simply put, I
: don't have a choice of which BSD I start with, so I haven't given it much
: more thought.

The FreeBSD guys pointed you at NetBSD for a reason.  (There should be a
reason this many intelligent people all agree on this point, shouldn't
there?)  If your management says "Do FreeBSD," looking only at the name, why
not point out the cost/return ratios on doing NetBSD versus FreeBSD:  NetBSD
would probably be portable to the Ultra in a matter of months, if not less. 
I suspect porting FreeBSD, and porting the devices and other necessary code
to FreeBSD, would take a year and a half at the least. 

: I'm willing to provide whatever information I can to help out with NetBSD
: as well (my time is limited though). 

Hope you have a big vacation coming in about a year... you'll need it then.
<grin>

Anyway, enough of my rant.  If you want someone from NetBSD to explain to
your management why it's a major win to choose NetBSD, I know there are
willing and able volunteers here.  I am one, but I have not been involved in
NetBSD long enough to be purely confident as a spokesperson.

=====
===== Todd Vierling (Personal tv@pobox.com; Business tv@lucent.com) =====
== "There's a myth that there is a scarcity of justice to go around, so
== that if we extend justice to 'those people,' it will somehow erode the
== quality of justice everyone else receives."  -- Maria Price