What is success for Fedora?

[Matthew Miller]

On Wed, Jul 02, 2014 at 05:01:33AM -0400, Christian Schaller wrote:
> To me any definition of success needs to be tied to popular adoption of
> the products we make, any measure of success that doesn't take that into
> account becomes to me a version of congratulating yourself for having
> achieved freedom of speech by putting yourself in a situation where there
> is nobody around to listen to what you have to say.

I personally agree that this is at least a large part of what I want for
Fedora, but I think it's important to note that there are possible answers
to the question which *aren't* necessarily related to "popular" use.

For example, one version of success could simply be "provides a good
integration point and test platform for future RHEL releases". Another might
be "attracts a packager/developer talent pool for Red Hat to hire from". Or,
a litte more nuanced version of the first: "Allows community experimentation
with distro creation, from which lessons can be learned for RHEL".

These are very limited definitions that I don't think reflect what we want
as a commmunity (or represent all of the value Red Hat gets from its
investment in Fedora, by a long shot), but I hope serve as examples of
_possible_ alternate views of success. Another not-RH-focused approach to
the same thing might be "provides the top-level packages and best practices
used by popular user-focused downstreams and remixes".

For all of these things, having a strong and growing user community might be
a factor which contributes to success (more users -> more developers, more
testing, more feedback), but isn't necessarily success itself.

And, I think that if we made "grow the user numbers" the _primary_
definition of success, there are other aspects of Fedora which would need to
be significantly adjusted. If we look at the "diffusion of innovation" curve
<http://en.wikipedia.org/wiki/Diffusion_of_innovations>.... let's see if
I can ascii-art it...


                                  *****
                                **  |  **
                               **   |   **
                             **     |     **
                             *      |      *
                           **       |       **
                          **        |        **
                         *          |          *
                        **          |          **
                       * |          |          | *
                     **  |          |          |  **
                  ****   |          |          |   ****
              *****      |          |          |      *****
  ************* Early    |  Early   |  Late    |          *************
  Innovators | Adopters  | Majority | Majority | Laggards
    2.5%     |  13.5%    |   34%    |   34%    |   16%


So.... I think mostly Fedora has aimed at that "Early Adopters" segment, and
a lot of our tension is from people who either want to pull more to the
"Innovators" side (the cries that Fedora isn't "first enough") or to the
next segment to the right ("stop breaking stuff so much").

I think we could in fact, with some investment, move more towards that Early
Majority segment. Even more so with greater investment -- like a longer
release cycle. But this might come at the expense of being less of a fit for
the early parts of the curve. 

I think that with Fedora.next we can create spaces within the project that
appeal to the innovators and (the earliest part of) the early majority --
but mostly, work at growing within early adopters.

This has an important consequence for the user numbers metric, though: by
definition, there are many fewer potential users there. Without
significantly refocusing the project, we're never going to be as popular as
distributions that target the majority. And that's completely okay, if we
define our success in the appropriate way -- and can find a good way to
measure it.

-- 
Matthew Miller
<mattdm at fedoraproject.org>
Fedora Project Leader