World Domination Plan

[Robert Millan]

Hi there.

I think this is time to reveal my World Domination Plan [tm].

  - Goal #1 is to release Debian GNU/kFreeBSD with sarge+1.
  - Goal #2 is to get the rest of the world to use it.
  - There's a third goal, but it's secret yet.  Be scared, really scared!

Guess I'll focus on goal #1 for now =).

I switched to GNU/kFreeBSD as my main production system, and I run it on my
workstation now.  On the last few weeks I've been using it to check mail,
maintain some of my packages, etc.  The only annoying limitation I found has
been the nasty DNS resolver bug.  Other than this, the system is usable in
production environments.

I've setup a buildd.  Guillem wants to do the same so expect that, together
with the almighty sanvila (and more to come), we'll have enough CPU power to
take over the archive soonish.

As for {wo,}manpower, we used to be short on that.  Certainly we need
collaboration from lots of people if we want to conquer the Debian archive.
My idea is:

  - Get the buildd infrastructure properly setup and working at good
    performance.
  - Integrate friki (Hector's box) into the Debian machine park.  This way
    maintainers can port their packages to GNU/kFreeBSD easily.
  - After sarge releases, announce (d-d-a) that we intend to take over the
    world^W^W^W^W release with sarge+1, and that we'll be filing a wishlist
    bug report for EVERY FTBFS.  This includes even libtool and trivial
    config.{guess,sub} stuff.  It also includes bugs for lame "Error: unknown
    system" -style FTBFS bugs.  Maintainers are pretty capable to fix all of
    the above.
  - Start filing bugs and offer assistance (and patches) for the non-trivial
    FTBFS bugs (These are very few, actualy).

In the meantime:

  - Get a working d-i.  We need to fix GRUB properly as a pre-requisite for
    that.
  - Go hunt the milestones.  The biggest showstopper is the toolchain (I'm
    working on gcc-3.5 and gdb, and Guillem said he'd fix binutils) but there
    are a few others.

I think all of this is fairly doable.  Please send me your suggestions etc.

-- 
Robert Millan

(Debra and Ian) (Gnu's Not (UNiplexed Information and Computing System))/\
(kernel of *(Berkeley Software Distribution))