On 11/16/18 5:17 PM, Jason L Tibbitts III wrote:
>>>>> "MM" == Matthew Miller
<mattdm(a)fedoraproject.org> writes:
MM> It's the fundamental contradiction that all operating systems face:
MM> users complain "too fast and too slow!" at the same time.
Well, then lengthening the Fedora lifecycle does not seem to me to be
the real solution. Instead, I think, it's to piggyback onto the already
long RHEL lifecycle but provide isolated pieces of "instability"
(i.e. new stuff) to people who want them.
Or more simply, don't try to slow Fedora down. Let Fedora be Fedora and
instead leverage Fedora to speed RHEL up in selected areas where people
want it.
MM> As noted elsewhere in this thread, many packagers are already doing
MM> this: maintaining a slow stream for EPEL or for RHEL as part of
MM> their day job, and a fast stream in Fedora.
So let RHEL be the core bits, moving as fast or as slow as Red Hat wants
to support. And let someone who wants "the new stuff" layer things over
that. Fedora could provide the layers, which move as fast as Fedora
does.
Of course, the practicalities of that and the potential combinatorial
explosion of options and interdependencies might render the whole idea
pointless. But that's supposed to be what the whole module system
solves. Or is it the flatpak system? Maybe both. And I believe those
are both conveniently supported by RHEL8.
So basically something EPEL-like that
doesn't have the "don't mess
with the base OS" restriction? That would be really nice. The long
time between RHEL releases (which seems to be getting worse, too)
makes supporting it a real pain for my group.
--
Mátyás (Mat) Selmeci
Open Science Grid Software Team / Center for High-Throughput Computing
University of Wisconsin-Madison Department of Computer Sciences