Re: Debian/BSD survey....
On Thu, May 31, 2001 at 10:02:57AM -0700, Michael Goetze wrote:
> Hi... I just read the archives for this list... sure hasn't been much
> happening. I'd like to conduct a quick and dirty survey to see whether
> it's worthwhile to get involved with this... I'd like to ask everyone
> subscribed to this list to take a minute or two to answer...
> Just so you know, my idea of a Debian/BSD system is a DEBIAN system
> with a BSD kernel... that's why Debian comes first.... (Whereby
> Debian's Not Linux, and GNU's Not Unix...)
> (1) Which BSD would you prefer to use as a base? (Free, Net, Open)
It seems pretty clear to me that NetBSD is a closer match to Debian goals;
e.g. its commitment to multi-architecture portability. We might also
discover less outright hostility to the project in the NetBSD community
than among the FreeBSDers.
> (2) What do you think is the best way to start?
The BSDs have a linux-compatibility subsystem on x86; I gather NetBSD's
is based on some version of Suse. Porting that to match Debian's ABI
would be a good place to start.
Once you have a NetBSD system that can run regular Debian x86 binaries,
you're on your way. Then, just begin replacing parts of NetBSD with
Debian components, probably starting with the init system. When you
run out of steam, package up what's left as the Debian/BSD core.
Besides the kernel, it might or might not include their libc and boot
loader. It probably would include their module, mount, ps, and
similar kernel-tied utility families.
> (3) Are you here to contribute, or to flame?
Post again with >80-character lines and I'll flame. :-)
As important as the technical details, we need to clarify goals.
(Of course different people have different reasons to participate,
so there's not just one goal.)
Is the point to have a BSD-kernel-based system that can run Debian
GNU/Linux binaries? Or just to have a BSD-based system that can be
configuration-managed with Apt tools? The former might be more useful
(at least for x86 users) but would require more deep magic, and cruft.
In practice, what we have to write down are the non-goals, so people
don't waste their time on things the project doesn't want.