Subject: Re: The future of NetBSD
To: Marc G. Fournier <scrappy@freebsd.org>
From: Daniel Ouellet <daniel@presscom.net>
List: netbsd-users
Date: 08/31/2006 22:24:34
Marc G. Fournier wrote:
> I'm curious here, but why did the *kernel* diverge for each project? 
> Like, I understand (or think I do) the philosophy of the OpenBSD 
> project, and that is high security ... but, wouldn't the security 
> improvements that go into the OpenBSD kernel not be applicable to NetBSD 
> / FreeBSD?  At this point, I couldn't imagine merging, but when OpenBSD 
> first branched off, one would think it would have been fairly easy to 
> keep the *kernel* itself relatively in sync, no?  Especially the code 
> audit that I imagine went into securing the OpenBSD kernel itself ...

Because they all have different goal and way to go at it. Just think 
treads as an example. NONE of them are doing it the same anymore, or 
will anyway. And some wanted more features oppose to make it stable and 
secure before any features were thought of. Some prefer read clean code 
and do the clean up before moving forward, others just takes what's 
offer by vendor and patch around it to make it work as blob.

I can think of a very long lists and I am not even a developer, so just 
imagine how different it is now.

Many see it differently. Some buy hardware, the latest, etc and then 
bitch that it is not working for them and expect to have it working. 
Others pick an OS because of it's goal and the fact that all along it's 
history, it stick to it like HELL and never compromise on it or it;s 
GOAL! Then the same users that have security and stability at hart, will 
look at what's supported and then buy accordingly and that also include 
boycotting the blob vendor until the provide documentations for the 
hardware they try to sale.

Just more example and a simple quote as recent as 4 days ago "The 
interface to the hardware (the PCI layer) is different, along with the 
SCSI midlayers, and the way DMA memory is dealt with." from 
http://www.undeadly.org/cgi?action=article&sid=20060827154223 and that's 
just one of many.

Or even the secure use of "strlcpy, strlcat" as an example to be more 
secure all around. Even the maintainer of libc still refuse to include 
them if I am not mistaken. Not all project have the same goal and go at 
it the same way.

The list is just way to long to put here, way to long and I don't even 
know much compare to many of the OpenBSD devs! So, just imagine the 
differences.

Also, Theo said many times that OpenBSD evolve by evolutions, not 
revolutions oppose to some project that will sale their mothers to get a 
few more users at the price of stability and even more gross, security.

I say, no thanks.

I am sure there would/could be some benefit for all the *BSD to work 
together, but for a same kernel, I don't think so. They can't even agree 
to the danger of BLOB and you expect them to do one *kernel*.

Plus BSD IS NOT a KERNEL, but a FULL fledge system and pretty soon a 
FULL router as well that will even give Cisco and Juniper a run for 
their money, or SmartNet! (;>

End of a to long post anyway.

Best,

Daniel

PS: The only one that survive, in any situations, are always the one 
sticking to their goal and they know what they are suppose to do and 
where they should go! The comities are left to the endless waisting time 
of the governments and manager that can only talks and do no good! But 
always think the know better and will tell you how to do it, but hey, if 
they knew, they would have done it already! Like we eared many times, 
shutup and hack...

PSS: So, one *BSD can claim the most different platform it runs on, one 
was able in history to claim to be the fastest of the *BSD, one was and 
still is the most secure of not only the *BSD, but of the OS at large 
and it sure didn't come over night either, but by sticking to the goal. 
And finally one can claim to be run by the most amount of users, that's 
the *Linux what ever flavor of the month you want. No wonder that need 
the most amount of users(tester) and peer helpers to keep it running and 
somewhat secure for a few hours now, well may be more at this time, but 
still a very long away off. Even IBM saw the light and they said they 
would run it, but they realize they can't, however, they sure see the 
$$$ coming in at supporting it and may be keeping it in the dark as to 
not loose their new source of income.