On Tue, 2005-03-01 at 09:06, Paul A Houle wrote:
I think Fedora should look at the CD crisis as an opportunity
rather
than the need to axe a lot of packages in a hurry.
First, the world is moving towards DVD. It's an absolute joy to
install Fedora off a DVD, although I've often found that process of
downloading an manipulating a disk image that big can be challenging.
(BitTorrent anyone?)
Moving to dvd and using CD's as secondary doesn't solve the problem.
You still have a huge download which can be annoying even on a fast
connection. YOu still have to engineer for CD's unless you get rid of
CD's all together. Lowest common denominator wins out.
My AMD64 has an optical drive that reads and writes every CD and
DVD
format (other than double-sided) that cost about $100; in the long
term, the majority of users are going to enjoy that convenience.
We have DVD iso's. We handed them out at Linux world so this is already
solved.
It may make sense for Fedora to see itself primarily as a DVD
product,
and to see the CDs as a secondary product. The CDs ought to be there
for people who don't have a DVD drive. Also, I like the idea of
being able to "target" Fedora for particular types of installation --
for instance, I can picture a single-disk distribution that brings
the system up with a minimal installation and does a network install
from there. I can also see a "KDE desktop" disk and a "Gnome
desktop"
disk.
I would like a base system that can then update from the net. As it is
you will most likely need to update anyway.
Part of Fedora's trouble is that it often packages five or six
packages
that almost work instead of one package that really works. (Having a
GUI and being written in Python are two big risk factors.) I'm typing
this on a Mac Mini that I just bought, and I've been quite impressed
with Mac OS X; Mac OS X standardizes on Postfix as an MTA -- it has
one program for organizing music, iTunes, which really works.
This is what we are getting away from. We want best of breed software
in the core and the rest in extras. What do you mean by "having a GUI
and being written in Python are two big risk factors"? As it is on the
desktop for the most part we only have one app for each task or at least
there is an app that is favored and the others are kept because of
various reasons. But, most of them are moving to extras anyway.
Of course, Fedora might not be able to shake that off the way that
a
new distribution can, say, Linspire. One some level it's really silly
that Fedora has both the Gnome and KDE desktops, yet there are two
reasons why that can't change:
Gnome is under heavy development and engineering at Red Hat. It may
just seem like we get it from upstream but that is just because all of
our work goes there. And yes we can move in other directions. Fedora
is leading the way in integrating new technology into Linux such as dbus
and hal. You might not see it because it is under the hood.
(1) Neither of those desktops are 100% satisfactory, and even if
you're
running one desktop, you're going to want some apps from the other
desktop and libraries to run the other desktop, and
(2) politics.
If Fedora remains a community project, it's likely to suffer
from this
"big tent" effect. It's going to be hard to make the decision to
marginalize a project which has made great contributions to the
Linux/Unix world, despite the fact that a real quality consumer OS is,
almost by definition, going to have one MTA and one desktop.
Um, we are already stripping out apps from the core. And a person won't
care about multiple MTA's as long as we have good defaults.
On the other hand, if we see it as an OS for enthusiasts that
other
entities can repackage, that might be a good place to be.
That is exactly what Fedora is.
--
J5