I think a lot of the people who have problems with Fedora would be
happy to have an alternative ("fork" distribution) to work on.
It looks easy to spin a custom distribution -- basically you can
unpack the fedora DVD on your web site, remove any RPMs you don't want
from the RPMS directory, add any ones that you want, run a few
scripts, and then cut your ISOs. All you need is 15 GB or so of free
disk space.
It would be trivial to develop a distribution that consists of a
subset of Fedora Core packages plus packages from Fedora Extras: the
advantage here is that users can plug into the existing system for yum,
so there's no need to work on security updates, plus users get 100%
compatibility with Fedora. It would be easy to address common
complaints about Fedora such as "bloat" (Two desktop environments, who
knows how many GB of internationalization files), packages that make it
difficult to install your own software (OpenOffice). It might be a bit
silly, but I'm still missing 'fortune' and the games package that came
with Slackware.
In the age of BitTorrent, the task of distributing the new
distribution would be easy as well: set up a tracker and a few seeds
with good connectivity, and the problem is solved.
(Actually, that scares me a little -- what if somebody spins a
'black hat linux' with a bad ssh that misrepresents itself as FC and
spreads the .torrent file around the net?)
This kind of project would make Extras more relevant: people who
want to put packages in alternative distributions would have a
motivation to get packages into Extras so they can benefit from the
update network.
Longer-term, it might be interesting to do more of a fork: but the
further you diverge from Fedora the more problems you have. For
instance, people who want to play mp3's might like a distribution that
has no media players and no dependencies on media players -- they can
install what they like the way they like it. Trouble is that they might
use yum to install some other packages that have dependencies on media
players and all hell breaks loose. Open Office creates similar
problems: removing OO gives us more flexibility with the web browser,
but if someone tries to install OO, it either aborts or the web browser
gets borked.
One answer is to maintain a parallel tree of rpm files (lots of
work, lots of resources), another answer is to make yum (perhaps a
forked yum) smart enough to 'overlay' one repository on another: 'Hard
Core' might balk at an attempt to install OO from Fedora Core, or
overlay it with an OO that has fewer dependencies on it's environment.
------
Many of us can imagine our own personal perfect Fedora-derived
distribution, but really the challenge is to think of a coherent
mission for an alternative distribution (or series of alternative
distributions) that would be compelling to enough people that it could
get some momentum. Any ideas?
--
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
http://www.redhat.com/mailman/listinfo/fedora-devel-list