Re: vote

[Richard Tibbetts <tibbetts@MIT.EDU>]

On Tue, Jun 26, 2001 at 09:56:13AM -0600, ray p wrote:
> What would be the point? If you want the userland that the flavor of
> BSD that you like comes with why not just use it? After all they all
> tend to be at least as "free" as Debian in any case. I would think the
> whole point of this project would be to get the GNU toolset with a BSD
> kernel and the Debian Way of doing things. I just don't see the point
> of copying the userland from one of the BSDs. Please tell me if I'm
> wrong on this.

What if the point is that one likes (for example) NetBSD, but finds it
useless as a login head due to the difficulty of maintaining current
software. Thus, one would want a decent package system, like the
Debian one.

One fundamental question about this project is whether it is creating
a second distribution of one of the BSDs, or if it is attempting to
become the primary distribution of whichever one it chooses.
Personally, I am interested in the latter, for NetBSD specifically.
But coming at this task from the Debian developer side doesn't seem to
me like it will work. So this project is probably constrained to being
a distribution of a BSD.

If you are going to be a distribution of a BSD, then throwing away
everything interesting about that BSD except the kernel is reasonable.
Otherwise you will be in a position of maintaining an unloved set of
base packages, and demanding that other developers constantly be
fixing interaction bugs. That way lies madness I fear.

So the question of which BSD becomes one of which kernel. I like
NetBSD best, because it gets you something the debian project doesn't
have, portability. FreeBSD gets you speed, but I don't see a real
demand for that. Cycles are cheap. OpenBSD gets you security, in
theory, but security is really a property of a whole system. By
throwing away huge pieces of OpenBSD and introducing GNU/Debian
pieces, you will destroy most of thet benefits. And as long as you are
throwing out everything except the kernel, supporting all three (four)
kernels wouldn't be that difficult.

Personally, I want to hack on NetBSD and steal Debian tools to do it.
But this is the wrong mailing list for that. So if you re going to do
Debian/BSD, then start with NetBSD and throw out everything except
maybe the libc.

tibbetts

PS: Unless someone is going to go off and do a big chunk of the work
with no reward, this is going to have to be a
release-early-release-often sort of project, just to keep people's
interest. I suggest you remember that when building a project roadmap.