Matthew Miller wrote:
We do not have this luxury. If we "push" something broken
to the people,
the people will _completely understandably_ throw up their hands and say
"huh, guess that sucked" and try another distribution. Or stay on their
other distribution -- let Fedora be the playground for developers, and
serious distributions for real users can pick out the parts that worked
afterward.
So what? Isn't that part of what Fedora is about? There must be SOMEBODY
taking the risks or things will never move. One of the main goals of Fedora
is to be that somebody, even if it means other distros which parasite our
stabilization efforts will be able to provide a more reliable user
experience in some cases (but at the cost of shipping old software which is
often missing features the users need or having issues already resolved in
the latest upstream code, because those distributions, by relying on fast-
moving distros such as Fedora to test the stuff, have to wait for us to test
them out for them first). And to be blunt, why do you think Red Hat has
started and keeps funding Fedora? Is it not because it is a "playground for
developers" from which RHEL "can pick out the parts that worked
afterward"?
Of course there's a limit to what we can do while keeping our distribution
usable and desirable for people. But we have to take risks, or we will not
be able to ship the latest and greatest features our users choose us for!
And the whole GNU/Linux ecosystem would not work without distributions
taking those risks.
At the end, I think things always played out well, even Fedora 9 turned out
not to be the major disaster people were calling it out to be (and KDE 4,
starting at 4.0.3 and gradually upgraded to improved versions up to 4.2.4 in
updates, was not the only risky major change in it! There was also Upstart
(which for some reason didn't cause as much scandal as systemd now is, maybe
because all its improvements over SysVInit weren't actually used), a new
GCC, a new X.Org X11 drawing complaints from proprietary driver users etc.).
Kevin Kofler