On 07/22/2013 07:38 AM, Matthew Miller wrote:
Conclusion
---
* Refocus Core to provide a better platform for building on
* Make room for innovation at the "Ring 2" level
* Empower SIGs to create solutions that fit
* Won't break what we have
* And we can start right now
I'm afraid this is all a bit too hand wavy for me to get too much out of
it. I guess first off I'd like to know in specific terms what would
change for the ring 0+1 items from what we are doing now. Anything?
As for the higher stuff, I think this has brought some good discussion
around some of the issues of app deployment and managing competing
dependency requirements - and that dialog needs to continue and expand
to include the broader development communities. But I really see it as
impractical for Fedora to host much more than what it already does. And
the focus Fedora brings on keeping development up to date I believe is a
vital one for open source in general.
And it's not "rings". It's a rats nest of interconnected dependencies.
And until there is a system for completely isolating the different
chunks that would added on (if that is even desirable for Fedora to
develop/support) it's an intractable problem.
One the anecdotal side - I manage a small office of scientists. We use
ScientificLinux for most of our servers and computational nodes. We use
Fedora on many of our laptops and desktops. I do so because EL just
doesn't handle some of the new laptop hardware. I generally have half
of the users on one release and the other half on the previous and
upgrade to Fn+2 when the time comes to limit upgrades to once per year.
And I use Fedora because it is closely tied to RHEL.
However, most of my users complain about the constant changes and
breakages, both in the UI and in the underlying system. For that reason
I'm going to give it a shot with running SL6 + Rex's KDE 4.10 repo on
some user's machines. While the UI will still change, hopefully the
breakages will slow. Also, it will sync development with the compute
machines.
So to some extent the idea of layered systems is appealing. But I
realize that supporting the development of repos like Rex's would be a
larger burden on our infrastructure staff. I would like to see the
development of a FreeIPA + koji + ? VM appliance that would allow
communities to come together on their own to provide their own repos.
Many of our users have shifted to Macs because of the necessity to work
with Microsoft Office documents and LibreOffice just doesn't cut it.
I've been having reasonable success running an increasing number of
Fedora servers. By using VMs and targeting single services it is fairly
easy to maintain them (with the help of tools like puppet and cobbler)
and upgrade as needed. For such systems yum upgrades work reasonably
well if the re-install options is too painful. A minimal kickstart
package set plus the few specific servers and I'm up and running. So
again, I feel like we are already producing a good 0+1 now - what
exactly changes?
- Orion
PS - I didn't use the word platform (damn)
PPS - I think our most useful efforts would be to reach out to the
development communities each of us are involved in and urge them to join
Fedora; to get the OS distribution perspective, and to give us more
resources. Because that I think is what we're really lacking.
--
Orion Poplawski
Technical Manager 303-415-9701 x222
NWRA/CoRA Division FAX: 303-415-9702
3380 Mitchell Lane orion@xxxxxxxxxxxxx
Boulder, CO 80301 http://www.cora.nwra.com
--
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
