On 07/23/2013 06:07 PM, Jiri Eischmann wrote:
Matthew Miller píše v Po 22. 07. 2013 v 09:38 -0400:
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
So there we have it. Comments and discussion, please!
The proposal looks frankly very cloud-centric. I have no problem with
that. What else should a Fedora cloud architect propose? But I'd like to
know a few things:
Is the proposal based at least a bit on some kinda of analysis of our
more successful competitors in the cloud area? Yeah, I'm speaking about
Ubuntu which currently holds 50 percent of the market. Ubuntu has been
very successful in the cloud and in the proposal I really don't see a
lot of things that Ubuntu has/does better and Fedora doesn't have/does
worse.
I just want to make sure that we won't turn the whole Fedora upside down
to make us more successful in the cloud and then find out that something
completely different was making us unsuccessful and competitors
successful. IMHO closings gaps between the competitors and us and
staying excellent in our strong areas would probably be probably a safer
strategy than turning everything upside down.
BTW speaking of Ubuntu, I think they've got quite different strategy -
one tightly integrated product across all uses (server, cloud, desktop,
and now maybe even tablets and phones). To solve the problem of newer
versions, special interests etc., they've got the ecosystem of PPAs.
That's where third-party entities can deliver software the way they
want. And AFAIK it has been widely popular with upstream projects
because they've got free hands with PPAs. And Ubuntu still has one
defined product and doesn't have to lower standards for software
inclusion.
IMHO it's a better solution than breaking the distribution into several
parts with different speed of development and different quality
standards from which you can build all kinds of fragmented products. At
least from the marketing point of view. As a user, I'd rather use a
well-defined distribution with one set of quality standards (and if I
wanted something special, I'd easily enable a third-party repo for that)
than a distro with well-defined core, but not so well-defined layers of
grey zone above it.
Just my 2c,
Jiri
I'm not cloud person at all and I like the Rings proposal. Server can be
still based on Ring0 and Ring1, so I don't see how it harm other use-cases.
Same standards for all packages simply didn't work. It can be seen
during every (mass, language) rebuild, which brings many problems for
those running the rebuild. Different people tend to package things
differently, even if there are guidelines. Lowering standards in some
areas and creating packages automatically might give people time to work
on their projects based above these packages. I guess the example with
Hadoop and Jetty is a good one, which can be often seen.
I also believe we need something like PPA. If koji needs new features,
or if new build system would be used, that's another part of the discussion.
Marcela
--
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/devel
