Fedora.next in 2014 -- Big Picture and Themes
Matthew Miller
mattdm at fedoraproject.org
Tue Jan 28 14:42:54 UTC 2014
On Sat, Jan 25, 2014 at 12:39:53PM +0100, Thorsten Leemhuis wrote:
> Understood, but OTOH it makes me wonder if Fedora.next is a step to
> big and needs to be split or something.
Well, in practical implementation, it probably _will_ be done as incremental
steps. For example, there's the possibility of decoupling the schedules of
the base release from the products, but one of the few solid decisions we've
already made is that that won't happen for F21.
> point is needed obviously. But sometimes I miss a few introductions
> words on the "why" we want all of that and how it's supposed to make
> Fedora better. But that's obviously meta/abstract again, which I
So, here's some things I see.
* Fedora's drift towards being primarily a desktop OS (with other use areas
considered secondarily if at all) ends up practically restricting uses
which people really do want Fedora for. That's bad for people who want to
use Fedora in innovative ways in server and cloud environments. Even
though we have a lot of sysadmin users and there are many examples of real
Fedora in production server environments, every time over the past decade
that someone has tried to figure out what Fedora Server might actually
mean, it's gotten stalled. This has left many sysadmins feeling like
either Fedora isn't a place that they can meaningfully contribute, or else
that their job is to be the Voice of No. Even when one doesn't want to
just be the project's "stop energy", it sometimes felt like there was no
other option. Fedora.next should *give* that option for postive
contribution.
* Although it's certainly not the only reason, Fedora as _solely_ a hobbyist
desktop is not ideal for an upstream for RHEL server and cloud products.
That doesn't mean that there isn't room for one, but if we would focus
Fedora on just that, we'd become increasingly irrelevant to our biggest
downstream -- and to the project sponsor. (And make no mistake -- that
concern of mine comes from a point of view of caring about Fedora first,
not just my employer, because we benefit from taking part in that whole
ecosystem cycle in ways beyond just Red Hat's direct employee and monetary
contributions.) One particularly underserved area for Fedora is RHEL
customers who would benefit from Fedora in some key areas (possibly even
in production!), as well as those who are interested in experimenting with
and possibily even shaping the future of the downstream.
* General trend in Linux towards the base distribution being "boring" and
not mattering. I asked several dozen different people at a gigantic Amazon
conference why everyone was using the distribution they chose instead of
Fedora, and the answer was almost universally "oh, I don't care; that's
not really an interesting question because there's nothing important at
that level". Now, that might not be really _true_, but it's definitely an
increasing perception. How can we either fight that perception, or make
sure that Fedora expands to also do work in the "interesting" space?
* Difficulty building things on top of Fedora. This means both end-users
with their own applications and bigger more complicated applications like
OpenStack which it would be nice to have in the Fedora universe. There is
value in the perfect-crystal-castle-of-all-packages approach, but if we
build that in one place while everyone is going by on the highway, we're
not really fulfilling our mission of leading free and open source software
-- we're doing cleanup from behind and not making the impact we could be.
I strongly believe that we can make room for this in the Fedora Project
overall, even for things that are not appropriate in the Fedora
*distribution* in its traditional sense.
* As you mentioned in an earlier thread, it would be nice if we could make
it easier to contribute minor changes, including big changes to more
obscure packages. We can't just be Wikipedia and allow anonymous editing
of packages (for one thing, we don't have an easy rollback mechanism and
consequences could be much more severe), but maybe we can find a way of
loosening things up for the fringes -- maybe even a really large fringe --
while keeping the core distribution under closer care. (Going way back to
last summer's discussion, the problem with the former distinction between
Core and Extras was that Core wasn't really open and the distinction was
based on who you worked for; I'm quite certain we won't make that mistake
again. Also, the standards for Extras were higher than those for Core,
which was of course backwards and an example of community showing the way.
But all that doesn't mean that we can't take the lessons and do it right.)
* Wanting to eat our cake and have it too when it comes to balancing
innovation and change management. In order to move fast while not breaking
everybody all the time, it helps to break things up into larger logical
segments than just packages. This helps draw out exactly what we really
care about not breaking, what all might need to change together, and
allows us be clear about what amount of breakage is an acceptable
exchange.
So that's some of my thoughts. More later -- gotta take the kids to the
dentist now. :)
--
Matthew Miller -- Fedora Project -- <mattdm at fedoraproject.org>
More information about the devel
mailing list