Fedora philosophy (was ATI video comes out of the closet)

Les Mikesell lesmikesell at gmail.com
Mon Sep 10 03:16:25 UTC 2007


Ed Greshko wrote:

>>> I would disagree about bleeding edge - Fedora tries to be current Which
>>> is what people were asking for at the time and still do so: very loudly.
>> Are there _really_ that many people asking for major changes in the
>> kernel to be done in mid-rev of a distro?  Or are people actually just
>> asking for current userland apps and perhaps drivers for new sata
>> controllers and the like?
> 
> [Very large snip]
> 
> I'm having a bit of a difficult time following this particular topic.  Too
> many people responding with "sound bites" to other people's "sound bites".
> 
> I think, as I believe I heard you say, CentOS 3.x and Red Hat 7.3 work
> perfectly fine for you and that you've no particular reason to update.  Yet,
> you have a bone to pick with Fedora...which you have tried using but doesn't
> meet your needs.  Some may question why you would waste your time trying to
> do so it when what you are running fits your needs...but I suppose that is a
> different matter.

Quick recap: on the server side, RHEL (and thus Centos) is just fine, 
because the relevant server applications (apache, sendmail, named, 
dhcpd, etc.) were feature-complete ages ago and all they need to do is 
just keep running, following the same standards as when they were 
installed. The side effect of having old application versions as the 
price of keeping a reliable kernel isn't a big problem there.

On the desktop side the opposite is true because the applications are 
still evolving rapidly. So, when the distribution ties the same policies 
to the kernel and apps, the price of a machine that you can trust to 
keep working is applications that suck (RHEL,Centos) and to get current 
apps (fedora) you have to take a wildly experimental kernel.

> What I think would be helpful would be for you to list what you feel are the
> goals of the Fedora Project and why/how the Project has failed to meet up
> with those goals.

To put it bluntly, I think the goal of the Fedora Project is to make a 
system stable enough to use _only_ during the last few months before a 
release of RHEL is cut.  I haven't actually gone back through the 
archives to check, but from memory I'll bet you find virtually no 
instances of messages saying "my xxxx hardware doesn't work" or "my 
machine won't boot" after updates during the relevant times in the FC3 
and FC6 releases.  Everything was sweetness and light for those short 
periods of time.  Then the FC5 and F7 releases came, along with a return 
to the "this doesn't work any more" messages which match my own 
experience.   I realize that most of the breakage comes from the 
upstream kernel, but the distro packagers know how to deal with it when 
it matters to them.

> I think that would help, at least me, understand what point(s) you are
> trying to make.

What I'd like to see is a distribution suitable for replacing MS windows 
on most desktops and I think between RHEL and fedora, all of the parts
are available but no suitable product exists because to get the current 
applications in fedora you have to take the unstable kernel that comes 
with it (most of the time).   Anyone who has deployed fedora on desktops 
in a large enterprise, accepts the updates and has never had problems 
please feel free to jump in and contradict me - I'd love to be proven 
wrong here, but even FC6 recently pushed an update kernel that wouldn't 
boot on some pretty mainstream Dell and IBM machines with scsi controllers.

-- 
    Les Mikesell
      lesmikesell at gmail.com




More information about the fedora-list mailing list