Defining the Fedora Project [LWN.net]

By Rebecca Sobol
October 14, 2009

There has been a discussion on the Fedora advisory board mailing list recently on the subject: "What is the Fedora Project?". John Poelstra started the discussion in an attempt to get this item off the agenda before 2010.

We really need to resolve this topic that has been on the board's agenda since January 2009. For some of us, since we joined in July 2009. I'm proposing that we set a hard deadline of "the end of FUDCon." This means that by the time we leave FUDCon the first part of December 2009, this issue will be officially closed and off our agenda until there is a reason to revisit it and we can start 2010 with a clean slate.

Beside the fact that it's an old agenda item, the Fedora Project continues to grow, and, without direction, that growth could eventually lead to fragmentation and chaos. Some definition of the target audience for the Fedora distribution, and some goals for the project are useful for everyone involved.

Mike McGrath wrote:

I've said it on the board list so I'll say it here. I strongly believe that volunteers can be [led] and I believe volunteers can lead. Right now Fedora is a place for everyone to just come and do whatever they want which is harming us in the long term. There's plenty of room for everyone in the Linux universe. I understand that by narrowing our focus we might lose some contributors who disagree with our values and mission. But that's better [than] not having one and having volunteers work against each other because they joined The Fedora Project thinking it was one thing only to find it's something else.

While there was general agreement that some kind of focus was needed, Greg DeKoenigsberg wanted to make it clear that competing visions still have a place in the project:

I also believe, however, that the Board must guarantee the freedom for dissenting community members to move in their own directions. Fedora's governance was built to ensure precisely this freedom. The Board is empowered to bless the "Foo Project", but any Fedora contributor is free to form the "Anti-Foo SIG," even if the goal of that SIG is to prove, through their constructive actions, that the direction of the "Foo Project" is wrong.

But, Máirín Duffy sees it as more of a positioning and messaging problem, as, currently, there is no coherent story for Fedora:

I don't *think* folks here take issue with the ingredients we've got floating around in the kitchen, and I don't think anyone is looking to throw any of them out. I think the problem is more that we haven't decided on a recipe with which to present them in. In the end, we've got to offer a menu that makes sense. And to the outside world, the Fedora menu looks like a confused mess. Rather than try to interpret it, most folks head down to the street to the more-easily-grokked McDonald's.

The project is more than just a distribution, it includes the entire community of contributors and users. The Fedora distribution also includes many spins, each of which has its own target audience, so perhaps the definition of a target audience should only apply to the default spin. Does that default spin contain development tools to appeal to developers? Does it target the lowest common denominator user with software for email, web browsing and an office suite? The project would like the default spin to fit on a single CD, so that it is accessible to people with low bandwidth and older hardware. You can't make a decision about what goes into the default if you don't know your target audience.

Fedora does have a Mission statement: "The Fedora Project's mission is to lead the advancement of free and open source software and content as a collaborative community." It also has defined values, or Foundations.

"Freedom, Friends, Features, First"

The four foundations are the core values of the Fedora community. They sprung from work on the Fedora marketing plan, and have replaced the old "infinity, freedom, voice" slogan. That slogan originally emerged from the design of the Fedora logo. That design has become a very powerful and effective part of Fedora's brand and image, but does not sufficiently describe our core values in a clear and effective way.

Fedora needs to define goals that align with its mission and values. No decisions have been made yet, be we look forward to seeing some definition to Fedora's target audience and some goals for the future soon.

to post comments

Defining the Fedora Project

Posted Oct 15, 2009 12:53 UTC (Thu) by nix (subscriber, #2304) [Link] (2 responses)

I don't get it. What's wrong with simply producing the best OS you can? Why do you need a 'mission' and 'values' and all that corporate-management garbage? Does this sort of cod-philosophising do any good to anyone?

Defining the Fedora Project

Posted Oct 15, 2009 18:00 UTC (Thu) by ewan (guest, #5533) [Link]

The problem is in deciding what is 'better' or 'best' in any given situation without some clear idea of what you're trying to achieve. A clear case in point would be the existing difference between Fedora and RHEL; a decision that might make Fedora a better OS could make RHEL a worse one.

Defining the Fedora Project

Posted Oct 17, 2009 3:45 UTC (Sat) by smoogen (subscriber, #97) [Link]

These things show up when an organization grows past a certain size.. mainly because the Venn diagram of common items becomes too complicated for the human mind to see "whats most in common (all the circles)" and "whats somewhat in common (some of the circles)".

I think many people would say we are once again re-inventing the history of Debian :). There was a certain point in its history where the major discussion was not what the next release was but "What does Debian Free mean" and "Why do we need a debian-legal?" At that point the various statements became more firmed up.

But to better answer your question: http://xkcd.com/592/

Defining the Fedora Project

Posted Oct 15, 2009 13:46 UTC (Thu) by miguelzinho (guest, #40535) [Link] (9 responses)

No mater what they say, to me Fedora will always be the beta version of the next RHEL. Why they just don't admit that? When Fedora gets 100% detached from Red Hat, then one can think about a mission or something like that to the project.

Defining the Fedora Project

Posted Oct 15, 2009 19:21 UTC (Thu) by mmcgrath (guest, #44906) [Link] (1 responses)

> No mater what they say, to me Fedora will always be the beta version of the next RHEL. Why they just don't admit that? When Fedora gets 100% detached from Red Hat, then one can think about a mission or something like that to the project.

Ahhh yes, to make up ones mind and not change it no matter what. Imagine if I copied everything you did. Now imagine if people thought less of you for it. This attitude towards Fedora has always amused me. Especially since no one considers Debian a beta of Ubuntu.

Defining the Fedora Project

Posted Oct 15, 2009 20:10 UTC (Thu) by nix (subscriber, #2304) [Link]

Ah, that's the thing, you see. Debian is a beta of Ubuntu, but Ubuntu is
surely a beta of Debian! ;}

Defining the Fedora Project

Posted Oct 16, 2009 4:39 UTC (Fri) by bojan (subscriber, #14302) [Link] (5 responses)

> No mater what they say, to me Fedora will always be the beta version of the next RHEL. Why they just don't admit that?

Because that's rubbish and you don't know what you are talking about. Fedora cannot possibly be a beta for RHEL, because the last several releases of Fedora did not result in anything resembling a release in RHEL world.

Repeating nonsense all the time does not make it sense.

> When Fedora gets 100% detached from Red Hat, then one can think about a mission or something like that to the project.

Oh, please! Red Hat is kind enough to pay for the infrastructure, bandwidth and most importantly developers in order to get Fedora out the door.

I'll give you a concrete example of how Fedora is a distro all by itself. Where I work, we are decked out in RHEL. We have a subscription agreement with RH that would cover all and every machine I ever wanted to have RHEL on. But I want and do run Fedora on my own (work and home) machines, because I know I'm going to get the latest Linux software, all packaged up the way I want it. So, I do not want to switch to any "final release" (what you think RHEL is) - I already am running one. Other Fedora users and contributors do the same.

Defining the Fedora Project

Posted Oct 16, 2009 18:14 UTC (Fri) by dlang (guest, #313) [Link] (4 responses)

since when is a company required to only have one beta release?

the multiple releases of fedora that did not directly produce a RHEL release all tested things so that redhat could decide if they belong in RHEL or not.

Defining the Fedora Project

Posted Oct 17, 2009 2:53 UTC (Sat) by rahulsundaram (subscriber, #21946) [Link] (1 responses)

Let me put it more simply: From a RHEL development perspective, individual Fedora releases might be viewed as development milestones leading to a RHEL release but for the large majority of Fedora contributors, Fedora stands on its own and is used for other purposes like say running the #1 supercomputer in the world or powering all those OLPC systems.Fedora includes around 12000 software packages but typically includes only about 2500 or so. It is easy to miss the rather vibrant and diverse community with a wide variety of goals if you focus only on the RHEL relationship. A lot of activity in Fedora has absolutely nothing to do with RHEL whatsoever.

Since the large majority of software packages in Fedora is maintained by volunteers who have a independent interest in Fedora outside of RHEL, they are now discussing what the future of the project should be along with Red Hat employees working exclusively on Fedora or other leading edge projects. Fedora's technical development is led by FESCo which is a completely elected body of representatives. Things like release schedule, update policies etc can be revised independently by this group along with other leaders in the Fedora community.

Defining the Fedora Project

Posted Oct 17, 2009 2:55 UTC (Sat) by rahulsundaram (subscriber, #21946) [Link]

I meant to say:

Fedora includes around 12000 software packages but RHEL typically includes only about 2500 or so.

Defining the Fedora Project

Posted Oct 17, 2009 8:56 UTC (Sat) by bojan (subscriber, #14302) [Link] (1 responses)

> since when is a company required to only have one beta release?

If you are claiming that Fedora 7 to 12 are beta releases for RHEL, you are clearly out of touch with what Fedora project does. And yeah, that's 6 "beta" releases. Please be serious.

> the multiple releases of fedora that did not directly produce a RHEL release all tested things so that redhat could decide if they belong in RHEL or not.

Fedora releases in fact never produce a RHEL release. People that release RHEL produce its release based on factors that most outside contributors to Fedora neither understand nor care about.

If the question is whether RH folks do heavy lifting of Fedora releases, the answer is of course yes. If the question is whether RHEL release team take software out of Fedora to build RHEL, the answer is also yes. But, Fedora is most definitely not a beta version of RHEL. RHEL has its own beta version, before its own release. Fedora, in the meantime, continues on its own path.

It seems fashionable these days to bash Red Hat, Fedora and confess loyalty to Debian and Ubuntu. Whatever. Personally, I find it more juvenile than Apple fanboyism.

Defining the Fedora Project

Posted Oct 17, 2009 21:47 UTC (Sat) by Los__D (guest, #15263) [Link]

It seems fashionable these days to bash Red Hat, Fedora and confess loyalty to Debian and Ubuntu. Whatever. Personally, I find it more juvenile than Apple fanboyism.

Hmmmm... I see more of it going the other way... Anyway, we agree on the level of sillyness of it all.

#define Fedora Project

Posted Oct 27, 2009 11:39 UTC (Tue) by gvy (guest, #11981) [Link]

> No mater what they say, to me Fedora will always be
> the beta version of the next RHEL.
It's not so. What you might have wanted to say is "part of fedora package base, including basesystem, is developed as beta stage for testing, stabilization and later reuse in rhel".

> Why they just don't admit that?
'cause that's not true. But not willing to admit that significant interest in the project is getting common package base hammered out with those who didn't buy ubuntu is scarey! :)

Defining the Fedora Project

Posted Oct 16, 2009 23:06 UTC (Fri) by Tet (subscriber, #5433) [Link] (2 responses)

Mike McGrath says:

I understand that by narrowing our focus we might lose some contributors who disagree with our values and mission

The Fedora project is already failing to live up to its stated mission (particularly the goal to produce a general purpose operating system). And yes, it's alienating me. Despite having dabbled with the other options at various times, I've always felt more comfortable in the Red Hat world, and it's been my primary Linux distribution since 3.0.3. But lately, Fedora releases have been of such low quality, and of such a specific nature that they're becoming problematic to use on anything other than a home desktop class machine running GNOME or KDE. That's fine. But it's not what the project was set up to do, it's not what I want from a distribution, and it bodes ill for future RHEL releases too. Is the solution to further narrow the focus? Personally, I don't think so. That's going to doom you to obsolescence, because there will always be those that want to use it for something outside your chosen focus, and I think the number of people in that category will only increase.

After 13 or so years, I find myself having to look for a new distribution, and it's a depressing thought. I don't much like Debian/Ubuntu, OpenSUSE isn't great either. Arch is looking promising, but hell, Fedora still gets a lot of things right that the others don't. I like having SELinux and SystemTap and other goodies in all sorts of areas where Fedora has blazed the trail and the other distributions are still lacking. But it really is getting to the point where I'm going to have to switch :-(

Defining the Fedora Project

Posted Oct 17, 2009 1:16 UTC (Sat) by mmcgrath (guest, #44906) [Link] (1 responses)

> The Fedora project is already failing to live up to its stated mission (particularly the goal to produce a general purpose operating system).

Actually that's what I'm advocating the goal to be. To produce a general purpose operating system. At present we don't have that goal. So what do we produce? You've got me but if you're not satisfied with it I can't say I'm surprised. After all, it's over 2,000 people working on a product with little oversight or direction.

Defining the Fedora Project

Posted Oct 17, 2009 21:35 UTC (Sat) by Tet (subscriber, #5433) [Link]

To produce a general purpose operating system. At present we don't have that goal.

Sure you do. It's listed as the first of your objectives, and has been since the release of Fedora Core 1: http://fedoraproject.org/wiki/Objectives.

What do you produce now? Mostly a desktop oriented OS. Try doing even basic things that would make it appropriate to a server build, like a headless install with the base packages only, and you'll see just how far from general purpose it actually is. If it doesn't work over a serial console, it's not suitable for a server (several of mine are in datacentres in the Far East, and you simply can't rely on having anything more). Now it's not that I actually want to use it as a server OS. I'd rather have something with a little less churn and a larger support window for that. Nor am I advocating Fedora LTS or similar. But if Fedora can't get that right, that's a bad sign for RHEL 6 being any use, and my servers are currently a mix of RHEL and CentOS, so that thought makes me very uneasy. It's not just issues related to servers, though. Horrendous dependency creep makes it virtually impossible to install a minimal desktop system, which some of us like to do for security and maintainability reasons. But the assumption is that you'll have the whole of GNOME, and if you try and install just the bits you need, yum pulls the rest in anyway.

Plus, of course, whenever I try and report these things as problems, they're usually either ignored until that release is no longer supported, or closed as NOTABUG. I know manpower is limited, and everything needs prioritizing. But it does seem like I'm fighting a losing battle even trying to get people to acknowledge the problems.