long term support release
dcbw at redhat.com
Wed Jan 23 17:05:48 UTC 2008
On Wed, 2008-01-23 at 11:41 -0300, Horst H. von Brand wrote:
> David Mansfield <fedora at dm.cobite.com> wrote:
> > On Tue, 2008-01-22 at 19:25 -0800, Sean Bruno wrote:
> > > To be honest, that's more or less what RHEL and the free rebuild CentOS
> > > are.
> > > Fedora is a sandbox of sorts. It's a place where applications come
> > > together and sometimes, where they come to die. :)
> > I use RHEL/CentOS extensively at work (versions 3, 4 and 5), and I'd
> > have to disagree about that. Tons of the 'cool' stuff that's in fedora
> > gets left out of RHEL/CentOS.
> Have you looked at EPEL? Much cool Fedora stuff ends up there. And if not,
> grabbing the SRPM and rebuilding is not /that/ hard either...
> > I don't know who decides what 'makes the
> > cut' for RHEL, but it certainly isn't the Fedora team.
> Small wonder... that is in the hands of the RHEL team ;-)
> > For example, gnumeric and git, both 'everyday' tools, are missing from
> > CentOS 5, AFAIK, but I'm talking about tons of other goodies. The RHEL
> > package selection process is too restrictive it would seem.
> RHEL is for production use, has to be rock-solid. Implies long-time
> commitment to the (minimalistic) package set shipped (even longer than the
> lifetime of a particular version). Fedora is bleeding edge, for adventurous
> users. "The kitchen sink and then some" is acceptable, as is "Oops, this
> didn't work out too well, let's try another approach/package in
> paralell/next time".
Right; having the latest "cool stuff" and support for the latest
hardware _by_definition_ is not going to be as stable as something that
moves more slowly and waits for the bugs to get shaken out.
You can't have it both ways. Either you pick "latest and greatest" or
you pick "rock solid". If you try to compromise too much, you loose
most of what makes both of those desirable. So you must essentially
choose to focus on one of those two, and the other one will
More information about the devel