It is no big secret that open source projects often wrestle with
defining themselves—who their target users are, what roles the
project should play, and so forth. Fedora recently examined those
questions head-on in a Fedora Advisory Board IRC meeting.
Specifically, the meeting addressed the Fedora user base
definition, which is meant to describe the people that Fedora
seeks to serve, as well as the project's related mission
statements. The current user base definition was written in
2010, which might not sound like that long ago, but the computing
landscape has in fact changed quite a bit in the intervening years.
The board meeting took place on April 4. Prior to the meeting,
Fedora project leader Robyn Bergeron posted a blog
entry outlining the topic. In that post, she explained the
rationale for starting the discussion:
While you want to have a mission and vision that is more long-lasting,
as a technology company or project, you have to recognize that the
*roadmap* to how you deliver on that mission and vision is subject to
being affected by many trends, market forces, and the like. The
mission may be the same, but how it’s achieved needs to be examined
from time to time (I would argue almost on a yearly basis) to ensure
that the assumptions you’ve made continue to be true, that you are
reacting properly to market influences, user trends, etc.
of the IRC meeting is available, but Máirín Duffy also posted a more
of the session on her personal blog. The crux of the discussion was
whether the current user base definition is too limiting, either
because it is out of date (and therefore out of step with current
computing trends), because it is too narrow (thus leaving out
important segments of the potential-user population), or because it
does not adequately guide Fedora development (and thus does not
advance the mission and vision statements).
2010, we hardly knew ye
The current definition is "a set of four characteristics that
describe the minimum level of consumer for whom we'll design the
default offering." The characteristics are being a "voluntary
Linux consumer," being "computer friendly," being a "likely
collaborator" with the Fedora project, and being a "general
productivity" software user.
Bergeron kicked off the discussion asking those present what
had changed since 2010 that might make the definition more or less
relevant. Among the replies were the rise of cloud computing and
virtualization, the rise of alternative Linux platforms like Android
and Chrome OS, the popularity of GitHub as a software development
site, and the growing importance of non-desktop hardware (including
tablets and inexpensive ARM products like the Raspberry Pi).
From those suggestions, Bergeron highlighted the explosive growth
of GitHub, which constitutes a major shift in how open source software
development is done. GitHub has five million users, she noted, most
of whom encounter open source software solely on GitHub, where they
either do open source development for their day jobs or to scratch a
personal itch. Consequently, they are not drawn toward participating
in a desktop distribution like Fedora. Bergeron noted that these
developers are often uninterested in packaging their software, and
Bruno Wolff commented that they often do not understand open source
That leads to a fundamental question: whether Fedora should help
spread free software in a community like GitHub which does not find a
Linux desktop distribution particularly relevant. Matthew Miller
argued that the mission statement is clear on that point; leading the
advancement of free software and free culture is the mission, and
"producing a Linux distro is just one of the projects we happen
to undertake." The trickier question, of course, is what if
anything Fedora can do to advance free software in an external
community like GitHub.
All user base are belong to Fedora
As to whether the user base definition establishes too narrow of a
target user, the discussion addressed both cloud computing and Fedora
sub-communities that do not work with the default GNOME-based
desktop. Fedora is still perceived primarily as a desktop
distribution, which may inhibit its use as cloud platform or in the
server room. This is a bit of a paradox, as Miller pointed
out, since Fedora is also widely perceived as a testbed for Red Hat
Enterprise Linux (RHEL)—which is predominantly a server
Nevertheless, cloud computing users constitute a distinct
sub-community within Fedora, as are the sub-communities that prefer
KDE or Xfce over GNOME. Duffy pointed out that there are also
sub-communities centered around specific computing tasks, such as
robot-building, and that those tasks are not well-served by "general
As a step in the right direction, Andy Price suggested that
Fedora's current, single-person user base definition be replaced with
a set of several "use cases" that cover a range of possible scenarios
that the project would commit to supporting. Drawing up such a list
of use cases would not be trivial, but the project could at least
prioritize the use cases to focus development.
Naturally, development work is not the only facet of "supporting" a
use case; Duffy commented that the distribution's messaging and
marketing materials (starting with the web site), can help
too—and the current messaging dates back to the Fedora 11
release cycle. The board can assist in messaging efforts, but John
Rose argued that having the board select the use cases was a
bad idea. Duffy agreed, suggesting instead that the project
"take a look at what people are doing with Fedora now that is
cool and elevate those things that are the most promising, call them
out, and get them resources."
If Fedora does decide to target several use cases (albeit
prioritized ones), as so many in the discussion seemed to feel it
should, then the project will need to focus more on being a stable
platform. That is a contrast with Fedora's current "bleeding edge"
approach, but the fact that Fedora already serves as the base for RHEL
shows that it is capable, and there are already several Fedora "spins"
that target specific user groups. But neither RHEL nor a Fedora spin
is quite a derivative distribution in the sense that (for example)
Ubuntu is derived from Debian. A Fedora spin shares the same core as
Fedora, and differs mainly in package and configuration changes.
Scott Williams suggested that the Android model might be a better
fit, where the "vanilla" Android is perfectly usable as-is, but the
project is still friendly toward the forks produced by device
vendors. "Vanilla Android is both user friendly and easily
customizable for forking." Others, however, were less
enthralled by the prospect of encouraging the parallel development of
several flavors of Fedora. Bob Jensen said: "The brand dilution
that Ubuntu has experienced is our future if we as a community do not
protect the Core and Foundations."
Of course, there is quite a big gap between replacing the
current user base definition with several use cases and supporting the
development of multiple Fedora derivatives driven by sub-communities.
The latter may be a long-term possibility, but at the moment all that
the IRC meeting seems to have established is that many in the Fedora
project think a single definition is insufficient to completely
describe the target audience for Fedora. But the conversation is
ongoing. Duffy posted a follow-up message to the
advisory board mailing list, suggesting brainstorming activities to
revisit the topic in further meetings.
Where further meetings will go is anybody's guess at this point.
One thing is for certain, though: by addressing the value and
viability of the user base definition in the open, whatever Fedora
ends up doing with it will more accurately reflect the will of the
community. A telling side-note in the IRC discussion was Miloslav
Trmac's observation that many of Fedora's current contributors simply
ignore the user base definition. Perhaps it does not work for them,
or perhaps they do not know it exists—an open and frank
discussion will hopefully alleviate either problem.
to post comments)