Summary: experiences with NanoBSD, successes and nits on a Soekris 4801

[Poul-Henning Kamp]

In message <20050619155228.Y6413 at fledge.watson.org>, Robert Watson writes:

>I general, I was quite pleased with the experience.  NanoBSD is fairly 
>straight forward to configre and adapt.

I'm still not satisfied with the nanobsd config/customize process,
ideally I would want to have only a single file with a sensible
format control the nanobsd build process.

The major obstacle is the "cutting things down to size" process
using NO_FOO options.

In order to get down a 31MB partition size things have to be cut
very extensively and not even the NO_FOO options is enough at that
level but sniper rm(1) commands are necessary.

I think the NO_FOO options is the best compromize, but we need them
to be more aligned to user concepts, "I don't need a compiler and
all that", rather than "Don't build the C++ compiler and hobble
the build because of this".

I am painfully aware that nanobsd is a case of moving from a
"program" to "program product" in the sense of The Mythical
Man-Month, and therefore I have no (or at best little) hope
that nanobsd will ever get "make world" like ease of use,
but we can certainly make it better than it is now.

For anyone intested I should point out that there is a webpage,
first draft at this point:
	http://www.freebsd.org/projects/nanobsd

and that any and all help is welcome!

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.