Fedora Board: Fedora Userbase Discussion

Last week the Fedora Board had an open, public meeting in IRC to discuss Fedora’s user base / target audience. Robyn announced the topic ahead of time and invited folks to join in. You can read the full meeting minutes, but I’ve gone through them and tried to pull out all of the interwoven threads of discussion and summarize it here for you as well.

Where we are now

The current Fedora user base is anyone who is:

Also worth mentioning are:

  • Fedora’s mission statement:

    The Fedora Project’s mission is to lead the advancement of free and open source software and content as a collaborative community.

  • Fedora’s vision statement:

    The Fedora Project creates a world where free culture is welcoming and widespread, collaboration is commonplace, and people control their content and devices.

  • Fedora’s core values:
    • Freedom
    • Features
    • Friends
    • First

Okay, now what?

With all of this in mind, Robyn opened the meeting with a few initial discussion points:

  • Does our user base, as currently defined, still hold true for us as a project in describing who we want our users to be?
  • Is what we are producing right now appropriate to meet that user base? If not what can we change to meet it?
  • Do our mission and vision statements still hold true?

Welcome to 2013 – what’s happened since 2010?

Starting with the first point, Robyn noted that it has been some years since we as a project re-examined the user base definition and determined whether or not it is something that still helps us achieve our mission and vision for the project. Our current user base definition and set of mission and vision statements were drafted in 2010. Robyn asked, “What has changed in the past three years that might make this list more or less relevant?” The changes meeting participants suggested were:

  1. Cloud (gholms) A potential set of contributors seem more interested in cloud images than desktops – they seem to be people who want to be involved in Fedora. (brunowolff)
  2. More desktops – there’s been an explosion of different desktop environments… Cinnamon, Mate, etc. (mattrose)
  3. Growth of Github – Github has grown from something like 100k accounts to 3 million users and 5 million repositories. Are those folks likely collaborators who care about open source? I would say yes. (rbergeron)
  4. ARM & maker culture – The rise of ARM, raspberry pi, and maker culture. Are those folks more likely collaborators more easily targeted? Possibly. (rbergeron)
  5. The rise of Android and ChromeOS (vwbusguy)
  6. Tablets – tablet devices have grown in popularity (vwbusguy)
  7. OS X – Mac OS X now seems to be the developer platform of choice, even amongst open source developers. “MacOSX still the easiest way to just get shit done, both as a user and a dev.” (mattrose) “Linux is staying at 2 percent of the desktop market.” (rbergeron)
  8. Rise of social networks (misc)
  9. Virtualization – Virtualization is much more prevalent (mattrose)
  10. The changing nature of open source contributions

    A bit further on, but relevant to the change that’s taken place in the technology and open source community since 2010, there was a discussion thread about how the nature of open source contributions itself has changed.
    In talking about the 5 million people on github producing open source software, Robyn pointed out that for them “packaging really sucks.” She also said, “I don’t know that they understand the value [of packaging]. I don’t know that swinging the pendulum back is that easy.”
    She continued, “Open source contribution has largely changed from ‘I work on the Linux operating system’ to ‘I get stuff done for my dayjob’ and ‘I scratch my own itch.’ To people whose experience is basically ‘github’ as far as open source contribution – which I would argue is far more likely to tap as a set of potential contributors than “ANYONE ON EARTH IN USER BASE” – packaging is a hell of a lot more work.”
    “So there is a new, gigantic community of people building cool stuff with open source for which Fedora as it stands is completely irrelevant,” said mattdm. “That gigantic, growing community isn’t a fluke – it’s just the result of the general trend in computing.”
    “I have to wonder:” replied Robyn, “Are we okay with knowing people are using Fedora, or are benefitting from Fedora, WITHOUT IT BEING the bright shiny thing on their desktop? Or saying Fedora in a big logo on it?”
    “If they mention it in their docs on setting up their app/tool/service, yes,” responded skvidal.
    “Our mission is to lead the advancement of free and open source software and content” mattdm said. “Increasingly, the movement around free software is going elsewhere. Not just from Fedora; from working on distros. Not because we suck; in fact, because we did such a good job that it’s boring. Interesting problems are now at a higher level.”
    Brunowolff brought up another point, “I am not sure developers really understand what free software is really about. I have had some interactions with upstream for some things where they really didn’t understand licensing and claimed their stuff was GPL when they were using non-free assets. They seemed to think if they could get something from a public web page, that it was free.”

    What good has the user base definition done for us?

    There was also a thread of discussion that centered around how having a user base definition for the past 3 years or so has affected the project. Several discussion points were brought up:

    • It only seems to apply to the default desktop – should it?

      inode0 also said, “Redefining Fedora to be something other than a live desktop is the correct place to start as what that is defined to be will guide the rest of this.”

      jreznik asked, “Do we know the user base of current gnome and target audience and as it’s our current default offering – does it match ours?”

    • Do we have any control over it?

      “Are we able to steer it, or are we completely in a position to follow our upstreams?” asked jreznik.

      mattdm answered, “We should have a user base statement of our own, and select upstreams that fit it.”

    • Why do we have a single defined character? Why not a target set of use cases?

      AndyP asked this question.

      “We do now, each spin has one even if it is implicit,” answered inode0.

    • We want contributors

      “The user base we want is contributors,” vwbusguy pointed out.

      “I think we *all* want Fedora to have more contributors,” Robyn added.

    • Many contributors ignore the user base

      “We have a fair number of contributors that don’t care about the user base as stated, and just ignore the wiki if they even know it exists,” pointed out mitr.

      “But you don’t have to spec out every user, just every core feature; specing out users limits your userbase.” pointed out AndyP.

    • What about sub-communities?

      “Each sub community is in charge defining it’s target audience; of course, many sub-communities might be competing for the same target audience (like the desktop environments are doing)” VikingIce said.

    • If we want more contributors, we’re not doing a good job at attracting them.

      dan408 pointed out, “Contributors don’t appear out of nowhere. Is there a book on Fedora I can read? Do I install fedora and then start reading every wiki page and become a contributor? There is no clearly defined path on a) how to become a contributor and b) how to plan to attract contributors c) a good place for contributors to start to learn the Fedora bureaucracy.”

    How about the mission?

    andreasn asked, “Is it a important mission of Fedora to spread free software to the masses?”
    “Yes,” answered Robyn, “But does that mean that Fedora has to be their desktop? Can Fedora be *on* their desktop? Can they use Fedora to develop with? Are we okay with our technology being used to spread open source culture, even if we aren’t getting credit?”
    “In fact,” responded mattdm, “that IS the mission. Producing a Linux distro is just one of the projects we happen to undertake.”
    “This is about making *us happy*, and having something that we think people want to contribute to,” said Robyn.
    “I was mostly curious in the context of target of ‘likely collaborators,'” responded andreasn, “as that is sometimes used as a bat for making things less non-hacker-friendly.”

    Chicken or egg – do we define the potential contributors we want, or the contributors we already have?

    “Can we… step back a little,” asked mitr, “and make sure that the board agrees whether it is attempting to define a vision for (existing or potential) contributors to follow, or to codify a vision that reflects contributors we actually have?”
    Looking at the contributors we actually have Robyn said, “I think that Fedora has proven to be a valuable platform for people who want to take it an extend upon in HOWEVER THEY WISH to achieve their user ends. OLPC does this with the XOs and sugar.”
    “OLPC and XO is in a different position,” responded jreznik. “Their target is pretty clear, one hardware…”
    “It is absolutely true everywhere startup and cloud-focused that I look: Mac on the desktop, Linux in the cloud,” mattdm pointed out.
    Robyn then provided a few scenarios in which people are using Fedora as a platform for building upon:

    • Arista Networks uses Fedora for their EOS for their switches.
    • Red Hat uses Fedora as the basis for RHEL.
    • There are a lot of people who use Fedora to build things. ARM devices, etc., you name it.

    Then there was a bit of discussion as to what makes Fedora a good base to build on. “The reuse is tied to our insistance on using free software,” said brunowolff. “With Ubunutu you have to worry about what you can reuse.”

    Fedora could be a better base platform, though; mattrose pointed out, “If we want Fedora to be a base for other things, we have to be *way* more solid than we have been.”
    At this point, then, a lot of the discussion began to swirl around the concept of Fedora primarily serving as a platform to build other things on top of.

    Sub-communities & being a platform

    There was a long conversation thread interwoven throughout the meeting about Fedora’s subcommunities.
    “What I’ve always liked best about Fedora,” remarked gregdek, “is the ability to empower sub-communities to do awesome things. I’d like to extend that ability. And I’d like to see a structure that allows sub-communities to identify common problems and work together on them.”
    “Nobody said it was going to be easy,” responded VikingIce, “but before we can do that we need a stable release (synced core or base OS) which they themselves can then build upon.”
    “How does ‘Fedora as a platform’ relate to the ‘user base’ focus we have now? Does it replace it? Change it?” asked gholms.
    gregdek answered, “It puts user base focus on individual sub-communities. Where, in my opinion, it should be.”
    vwbusguy related gregdek’s idea to the Android community. “I like the idea of vanilla android vs vendor forks as a model,” he said. “Vanilla Android is both user-friendly and easily-customizable for forking.”
    mattdm warned, “As someone with experience in making derivative distributions, if we really want to be friendly to that, that has huge impact on how we do many things (our updates policy, for example.)”
    What exactly is meant by sub-community though? gregdek provided some examples: “Sub-communities will include more than just DEs. They’ll also include cloud platforms, web development platforms, mobile platforms… and so on,” he explained.
    “Sub-communities also define their own target use case by their own existing nature,” added VikingIce.
    mattdm had a very good question about conflicting and intersecting subcommnunity needs. “What happens when sub-communities’ needs clash?” he asked. “That’s kind of the point of having one default target originally.”
    “That’s the kind of problem that the Fedora board should be regularly dealing with, and making decisions,” answered gregdek.
    jreznik questioned this. “So you prefer subcommunities but wants one central board thing to make decisions?”
    “One central board to handle disputes and apply scarce resources to common problems, yes. That’s exactly what I want,” gregdek responded.
    “Yep, I agree in this way but I don’t see we hit such issues very often,” said jreznik.
    “I think we hit them all the time,” gregdek explained, “Like ‘better package build tooling,’ which could benefit everyone.”
    jreznik replied, “Better packaging tools are not a Board decision – someone has to do it. The Board can’t say ‘from today, you’re not Mate packager, but you’re working on tools…’ It even does not work in companies; I expect dan408 would leave such company.”
    So the points and suggestions that were made during the part of the discussion were:

    • Fedora’s subcommunities, in order to grow, need Fedora to be a more stable base for building upon than it is now.
    • Fedora may need to revisit many of its policies (the updates policy was given as an example) in order to be friendlier to folks using it as a base to build a derivative distro on top of.
    • If the Fedora project’s target / user base is technologists making derivative distros and/or using Fedora as a platform on top of which to build other technologies, then the subcommunities that come out of building on top of Fedora should determine their own respective user base targets.
    • When conflicts arise between sub-communities, it could be the Board’s job to resolve those disputes. However, the Board doesn’t actually have a whole lot of power to do this in reality.

    Changing the user base definition

    There was also some discussion about, if we were to change the language of the existing user base definition for Fedora, how would we change it?
    [We need to add something about ‘cloud’ and/or ‘derivative distro maker,’ in my opinion,” said rdieter, after the platform discussion. “Or heck, we can even explicitly say ‘maker of Fedora spins.”
    “So we’d be serving as a basis for derivatives, kinda like debian?” asked misc.
    StillBob pointed out, “that is what made it cool for us tinkerers and hobbyists.”
    “Absolutely,” said gregdek.
    The conversation drifted from there, however, so there were no other concrete and specific suggestions made for changing the current user base definition language after this point.

    Current perception of Fedora

    There was also a thread of conversation about how Fedora is currently perceived. This thread came up because it was thought that Fedora’s current perception might affect any future plans to change who the project is for and its priorities: If the project is known for being a specific thing and we decided to do something very different, for example, we’d face a lot more struggles in convincing folks interested in the new, different thing to join the community.
    There were two main current perceptions of Fedora that everyone discussed:

    1. Fedora as GNOME OS / as a desktop product

      DiscordianUK stated, “The market such as it is perceives Fedora as Gnome-OS.”

      “I think that’s probably largely fair to say,” Robyn replied. “I’m not saying it’s correct; I’m just saying that I think that is the perception.”

      vwbugguy added, “Yes. We have a strong KDE and XFCE community and it seems like they kind of get the back seat with our labeling.”

      “Yep,” said gregdek. “Have for years.”

      “We have many desktop environments,” DiscordianUK pointed out.

      inode0 suggested that one of the core problems might be that Fedora is a Linux desktop distribution.

      “That’s kinda paradoxal,” misc responded to inode0, “since people think that Fedora is ‘RHEL beta,’ but few people use Fedora on a server, and few people use RHEL as a desktop.”

      “Is it a problem?” jreznik asked inode0. “Do we want to be more [than a desktop distribution]? Do we have the contributors to we need to be more?”

      Robyn responded, “Yes. And we have many, many use cases that go beyond ‘as what’s on my laptop or desktop machine as the primary OS.'”

      “So… maybe instead of targeting a user-base, target use-cases?” suggested rdieter.

      “But we should attract more people who want to build other things too,” said inode0, “and we aren’t because of our desktop focus.”

      The conclusion of this thread of conversation appeared to be that Fedora is currently seen as a desktop-focused Linux distribution, and that this perception may be weakening our ability to serve, as discussed earlier, as an effective platform/base for building things on top of.

    2. Fedora as a RHEL testbed

      “The largest perception I see in the wild (after a decade!) [of Fedora]” said mattdm, “is'”that’s the RHEL beta test thing, right?'”

      A little further on, there was more in-depth discussion about this perception.

      “I’d like us to stop pretending Fedora is not build for RHEL,” said mitr.

      “Making fedora serve as a base for RHEL is an extremely important use case in terms of the sustainability of the project,” said mizmo. “That doesn’t mean it has to be the only use case. I also don’t see anybody pretending it’s not a base for RHEL.” (ed. note – that’s me. 🙂 )

      “If it is one of our strengths,” gholms pointed out, “we should play to it.”

      “Certainly!” responded mizmo. “Having a stable platform that is used for a widely popular enterprise Linux is absolutely a strength.”

      mitr pointed out, “The user base on the wiki and people who use RHEL have pretty much nothing in common.”

      “That’s actually not true at all,” responded mizmo.

      “Target use case: ‘make derivative spin/distro'” said rdieter.

      “Being seen as RHEL beta is harmful,” said misc, “as I am pretty sure there is less contribution from external company due to that ( when compared to debian, for example.)”

      “I don’t see RHEL as a problem,” DiscordianUK stated.

      “I don’t think RHEL beta is accurate,” said brunowolff. “It is more used for technology development, some of which succeeds and some of which fails.”

      “Well, as i said, RHEL targets mainly servers, and I think I am in the minority who run Fedora on server,” said misc. “That’s not a great way to test a beta.”

      The conclusion here, then, seemed to be that even though it is a very real perception that Fedora is merely a ‘RHEL testbed,’ it also directly conflicts with the other very real perception that Fedora is a ‘desktop distro.’ The conflict comes from the fact that RHEL is primarily a server, not a desktop, so if Fedora is a ‘RHEL testbed’ then how can it also simultaneously serve primarily as a desktop?

      It was also suggested that serving as the base that RHEL is build upon is a strength, and that strength could be used to expand Fedora’s utility as a base platform.

    3. Who determines Fedora’s direction? The project board’s futility

      A theme that came up a couple of times was how much the project board could actually do to change anything.
      “To some degree we don’t get to decide what we will be,” said inode0. “We start somewhere and the community takes us along for the ride.”
      “The board do not control people,” said misc.
      rdieter responded, “but we can say ‘no’ to things.”
      “Only if the board is asked,” StillBob pointed out.
      “The board can be proactive too,” said rdieter.
      StillBob questioned this. “How many things are happening without being asked?”
      “If no one notices or is hurt,” rdieter explained, “it’s not a big deal. When it causes problems or conflict, [the board] steps in.”
      “We can occasionally reset the starting point,” inode0 said, “but ultimately people contributing will again reframe the direction we go from there.”

      Use cases vs. userbase

      One very long thread of conversation started by AndyP was changing the user base definition document so instead of referring to a specific user we are targeting as a project, it referred to a set of use cases we had prioritized and committed to supporting.

      “Targeting the universe is the opposite of strategy 101,” Robyn said. “YOU CANNOT BE ALL THINGS TO ALL PEOPLE.”

      “If you focus on a user base, I think people look at it and think that if it’s not them, it’s a problem, and they disagree with it and ignore it,” mizmo said. “If you focus on a set of use cases we want to support, and they are cool things our community wants to do, it might be easier to build support for them.”

      “If the community wants to do them already,” asked mitr, “why would we need to define the use cases in the first place?”

      “If they aren’t defined,” answered mizmo, “then they are harder to get support for. For example, if we decide that building cool robots using Fedora is a priority use case, then we can get funding to go to robotics conferences and talk about Fedora there.”

      “If it’s just a small subcommunity doing robot stuff,” mizmo continued, “it’s hard to get other teams in Fedora like marketing and design and ambassadors and events all supporting it – because the priority is not clear.”

      “Or because there is more packagers than marketing/doc/etc team can handle,” misc pointed out.

      mizmo continued, “Without priorities, everything is equally able to be ignored / put to the bottom of the pile. With priorities, it’s easier to get resources to the things we think are important.”

      “That’s talking about some prospective community that doesn’t exist,” responded mitr. “OK, I was thinking in context of existing contributor base.”

      “No,” replied mizmo, “There is actually a robotics community within Fedora. I wasn’t making up an example.”

      VikingIce added, “But what’s important is what matters to the subcommunities, which are the people doing the work in the first place.”

      What can the board do, anyway?

      “Marketing is part of the problem,” said vwbusguy. “If people don’t know that a robotics community even exists, for example…”

      “Anyway,” mitr said, “the only thing the Board can directly affect is marketing money? How realistic is that it will significantly change what most of our contributors want to build on?”

      mizmo disagreed. “That isn’t the only thing the board can directly affect. It’s not just money. it’s how we talk about the project, what conversations we decide to enter into and which we abstain from.”

      “Perception matters a great deal,” gholms added.

      “It’s about building the relationships with those communities best served by the prioritized use cases to expand Fedora’s reach into the mainstream of those communities,” mizmo continued.

      “Realistically there are the dozens/hundreds of people building RHEL and they are not going away,” mitr responded.

      “Okay, and so what?” asked mizmo.

      “So they can and will drown out some other use cases,” said mitr.

      “How/why?” asked mizmo.

      “Yep. We have to build better relationships,” said Robyn, stepping in.

      No edicts from on high, please

      mizmo asked, “Does anybody disagree with the idea that we all want fedora to be used in building cool things? Because if we all agree with that, then the next step is to figure out what cool things.”

      “I agree completely with that,” answered inode0, “but I disagree that the board or any other small group should define the set of cool things.”

      “I think that the ‘cool things list’ may be ‘here’s what we’re doing the best at RIGHT NOW,'” Robyn responded. “We can’t tell people what to do. But we can certainly foster an environment that it easy for them to do really great things.”

      “Well, we can’t be everything – that’s what we are now, that’s not so successful is it?” asked mizmo.

      “We aren’t everything now,” inode0 responded.

      “We are a complete chaotic anything goes bazaar,” mizmo disagreed. “I consider that everything.”

      inode0 pointed out that he thinks we are pretty close to exactly one thing now.

      “By being everything we are nothing,” mizmo responded.

      “It isn’t about being everything,” said inode0. “It is about whether 10 people are smart enough to declare we are now X.”

      “I don’t think anyone is saying that anyone is going to do that,” mattdm said.

      mizmo agreed, saying, “I don’t think they should do that. Rather, I think we should 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.”

      “Well, if we say these are the 3 use cases we care about – I think that is what we are doing,” inode0 said.

      “If the board comes down with those use cases from the mountain, sure,” mattdm said.

      mizmo responded, “It depends at what level you’re defining the use case. A use case can be quite broad or very specific.”

      Building on, or building a base

      “On, or as a base?” asked andreasn.

      “Either,” answered mizmo.

      “Does it matter?” gholms questioned.

      “Well,” andreasn responded, “one is building a clear operating system that includes tools for making, say, blueprints for robots on, the other would be a base operating system for robot software.”

      Who doesn’t want to build cool things?

      “What would be the opposite of that statement?” asked mitr. “People who want to build uncool things? People who want to… watch youtube movies?”

      “Does there have to be a direct opposite of the statement for it to make sense to you?” asked mizmo. “The user base is a good alternative for comparison. E.g., Fedora being for people who build and make things, not people who consume or do ‘general productivity’ tasks.”

      “No,” mitr said, “but it needs to be more specific than “90% of breathing people” (I know the 90% is exaggerating.)”

      mizmo agreed, “Right, that’s why I’m saying the next step would be to define which type of cool things we want to focus on enabling people to build. Or think about what folks who aren’t currently using who fit our core values and would benefit from free software and try to engage them.”

      Big changes make a mess?

      “So, if we declare this the key use case, the release cycle changes *follow*. Making that the driver is *backwards*.” said mattdm.

      “The whole /usr move, gnome3, systemd and anaconda changes have made a mess of the last few releases,” said mattrose.

      “Well,” responded andreasn, “One has to evolve or die.”

      “Yeah,” agreed mattrose, “but it was still a mess.”

      Users vs. Contributors

      “Looking on it – we maybe hit clash between contributors (who contributes to Fedora for some reasons) and users (who just wants to use it)…” said jreznik, “and there are not enough contributors to say, it’s the only audience.”

      “We should blur the divide between users and contributors as much as possible,” vwbusyguy said.

      mattrose added, “I think more users would contribute if it were easier. I don’t think anyone is ‘just a user.'”

      What do we want, anyway?

      “I think your point follows well too,” said herlo to inode0. “There’s no real value in discussing the user base until we know what we want.”

      “Well, I recognize myself in the userbase,” said misc, “but yeah, we are not using it to define what we do.”

      “What we are producing is a desktop distribution – see the website,” said inode0.

      mizmo suggested, “We can change the website. The messaging of the website is all F11/F12 era.”

      “It is the substance that is limiting,” responded inode0, “so we can think about how much of this can be accomplished via marketing/website stuff – some surely can.”

      “How does being a collection of desktop linux systems do any good for anybody?” mizmo asked. “Are there people in the world whose goal in life is to try multiple desktops? I can’t imagine there are enough for Fedora to then matter much in the world? Trying different desktops out isn’t really doing anything.”

      “Desktops should be one thing Fedora is used to create,” inode0 said.

      “What about servers then?” asked andreasn.

      “They’re there, but secondary,” said vwbusguy.

      VikingIce said, “If you want to make people contribute and participate it needs to be fun and fulfilling for them to do so; the more sub communities we have, the more they are likely to find something they feel is fun and rewarding to participate in.”

      “It’s going to be fun and fulfilling if what we’re doing MATTERS,” said mizmo. “Serving as a multidesktop distro doesn’t matter. How many multidesktop distros are out there?”

      Summary

      I think the overall conclusions you could draw from this discussion are:

      • We could replace the user base document with a definition of prioritized use cases Fedora is meant to support.
      • There is value in defining a set of prioritized use cases in that it helps the various teams and groups within Fedora determine more intelligently how to expend their limited resources in supporting any specific use case.
      • The board can help support a given use case not just with funding, but with direction, messaging, marketing, and other influence.
      • The project board shouldn’t just come up with a set of use cases and declare the community should focus on those; rather, they should review all of the use cases Fedora is being used for today and prioritize, promote, and find resources for those determined most important.
      • Whichever way we end up going moving forward, the Fedora website should reflect that direction. It currently is very desktop-centric.
      • Whatever we decide Fedora’s ultimate purpose is, we should make sure it’s something that matters so people will care about it. Perhaps being a ‘multi-desktop distro’ isn’t really something that matters.

      Other discussions

      There were a few other loosely-related discussions that took place during the meeting.

      Upstream

      jreznik brought up the tension Fedora experiences in terms of deciding its own destiny vs. being at the mercy of upstream.

      “As a distribution,” jreznik started to explain, “we really has to be more in touch with upstream, especially that upstreams that influences our offering and to know, where they lead…”

      “Upstream is defined again by subcommunities like each DE,)” said VikingIce, “So Fedora can never be any closer to upstream than the subcommunity makes it.”

      “Agreed,” said Robyn in response to VikingIce. “And when we have disconnects between communities, or we aren’t in trust of them, we have found different substitutes. Mariadb / mysql seems to be a reasonable extension of that idea, I think. It’s not about making us a great platform for mysql; it’s about us having a community that is happy about the communities it works with.”

      Content

      There was a brief discussion about what Fedora can or should do in advancing free content, since our mission statement mentions free content.

      “Our mission is to lead the advancement of free and open source software and content,” mattdm said.

      “I am not sure we are leading on the content part, to be honest,” said misc.

      “I’d really like to see that content part!” jreznik added.

      Mmsc replied, “Isn’t it the job of a already existing community (like wikipedia, etc.,) so maybe we should think how we can complement them?”

      “I mean it more in the way – propagate that content, show people free content, open data works great with our values etc.,” jreznik responded.

      An easy-to-modify platform

      The conversation then circled back around the concept of Fedora serving primarily as a platform for people to modify.

      The principle of freedom makes Fedora more compelling

      “I think that the best way to get more people to contribute is to have something that is easy-to-use / modify / reshape for their own ‘end user purposes’ – and to have an enduring promise and commitment to freedom that we yell out from the skies,” said Robyn. “There are 1 billion places for people to contribute. It is easier than ever. Standing for something and having a reason more than ‘easy-to-use’ affects people’s hearts, makes them have a relationship with us, etc. And frankly, it helps them overlook problems like ‘oh god updates just ruined my day’ too.”

      “Yup,” agreed vwbusguy. “The FOSS emphasis is a big draw for contributors, IMO.”

      “It lets them want to put up with ‘must upgrade once a year,'” added Robyn.

      “And we shouldn’t sacrifice that for the sake of being ‘user friendly,'” said vwbusguy.

      “Honestly: Fedora as a platform rather than “Fedora as the perceived gnome-os” is probably better as a base for RHEL anyway,” Robyn mentioned.

      “The reality is that Fedora *is* a platform,” said gregdek. “A great one.”

      Not perceived as a platform

      inode0 pointed out one issue, “The reality is no one really perceives it to be a platform and perception has a lot to do with who comes to look around.”

      VikingIce explained the issue is “because we are sending the message ‘we have a default’ and by doing that we ‘this is what we want you to do’ instead of encourage people to explore, try, and if they dont like what they see here are the tools for you to create what you like to see.”

      Building products?

      “The good thing about the desktop offering,” pointed out andreasn, “is that it is a clear product. The server should be a clear product, too.”

      jreznik said that he “would even not object with GNOME OS based on Fedora – even as a separate project for example same as RHEL is.”

      “rbergeron’s wording as ‘platform’ makes more sense to me…” said mitr. “Anything can be used to build a derivative, but that doesn’t mean the base does anything particularly well.”

      “Wording doesn’t matter to me,” said rdieter, “As long as it’s a clear (and important) documented use-case.”

      “I would like for us to be a kick ass guest,” said Robyn.

      The implementation of a platform will be sticky

      “I can stand behind the idea to build a platform,” said mitr, “but the implementation will be a minefield.
      Are we willing to walk through it?”

      “We must be,” gregedek responded.

      “A platform for what?” mizmo asked.

      “Everything,” VikingIce answered.

      “So now we are back to being everything again,” mizmo sighed.

      “For building $stuff on,” mitr explained. “Like, an API that actually makes sense instead of the 10 million libraries we have.”

      “Do we want as a project to be that one who ships the final product people can use, or do want to be a catalyst to allow other projects to build on top of it? GNOME OS, RHEL, cloud images…” asked jreznik.

      What our the next steps?

      “What are our steps at this point?” asked Robyn. “Do we agree that $userbase as defined is not matching up with what we are producing, and that we are not making great strides in reaching that userbase with a product or messaging?”

      Everyone basically agreed with this point.

      Agenda for the next meeting

      “It seems like common pattern here is platform – could we start next meeting with it? What such a platform could look like?” asked jreznik.

      “Maybe platform plus (use cases as they connect platform to subcommunities)” suggested gregdek.

      “There’s little brainstorming exercises; I have a book of them. Maybe I’ll go through it and pick out some I think might work for us and suggest them on the list?” suggested mizmo. “I think if the next meeting focuses on building an artifact, the discussion will be a bit less chaotic.” You can read those brainstorm exercise suggestions here.

      Material for the next meeting

      “I am going to write up thoughts on ‘what has changed,'” said Robyn. “I think a lot of times people don’t really look out the window at the rest of the world. And a few other thoughts as well. I have this enormous bunch of writing that is bordering on novel-size.”

      Research to gather

      “I think it would be useful to know why people contribute to Fedora *right now* – what are they passionate about?” Robyn said. “I wonder if a call for people to blog / write about their interests would be useful (and nice to see in general.)”

      “There’s a thread on devel list now about what people use Fedora for,” mizmo pointed out. “I think it would be cool to do a widespread blog thing on that.”

      “Yep,” said Robyn. “I think they’re different, slightly, though – contributing and using – though largely the devel list is going to be contributors.”

      “Maybe a meme,” herlo suggested. “Name 3 things for which you use Fedora.”

      “Fedora Haikus,” said Robyn. She then asked herlo, “Do you want to do a Fedora “3 reasons you use it” thing more in social-media land (Facebook, etc)? I think a lot of folks *don’t* read devel list – and it would be fun and concise.”

      “Maybe,” replied herlo. He said he will think of a “fun little gag that might catch on.”

      “What about 3 things you wish you could do with Fedora?” asked mizmo. “Or 3 dreams you have for Fedora?”

      “Why I use Fedora in 3 words (for twitter),” suggested AndyP.

      “We need a good way to find out a few bits of things: Why people care, why people use, why people contribute,” Robyn concluded.

      Packagers vs. other contributors

      “In fact,” said misc, “I think packagers are over-represented, in the sense people try to become packager from others team.”

      “While we need more packagers,” added brunowolff, “I think non-packager contributors are more valuable as they are harder to find.”

      “I don’t know that ‘make it easier’ converts people to contributors, though,” said Robyn.

      “‘Contributors’ is kinda vague,” DiscordianUK added.

      “But non-packagers probably find dealing with breakage harder. They aren’t in as good of a position to be able to directly fix it,” brunowolff pointed out.

      What’s next?

      I think the board is planning to have another open meeting to discuss this topic further. If you’d like to continue the discussion and join in on the next meeting, check out the Fedora advisory-board mailing list.

26 Comments

  1. Very nice writeup! And a very engaging discussion.
    As a GNOME contributor and former OPW intern, I personally feel that “The pre-eminent GNOME distro” would be a plus in Fedora’s marketing, if it embraced that. Instead of “Fedora as a desktop OS,” you could change the marketing message to say that “This is one of many cool things you can make with Fedora,” and play up our communities’ ties to show how Fedora works well with upstreams and empowers people to create new things.
    I personally use Fedora for content creation, i.e. writing for a living, and I think that — largely thanks to GNOME applications — it excels at that task. So if you wanted to emphasize the “content creation” angle, you could possibly do that through the GNOME lens as well. I feel it has a lot to recommend it vis-a-vis OS X, especially for people who can’t afford to buy a Mac.

  2. If Fedora is a testbed, statements like this doesn’t make sense: “Having a stable platform that is used for a widely popular enterprise Linux is absolutely a strength.”.
    I stopped using Fedora because of all the instability with systemd. It’s great if you want to move forward systemd, and test possible features for RHEL, but it’s atrocious for the general user who wants a distribution that Just Works.
    I switched to Arch Linux which has rolling releases, even newer packages, and somehow it’s much more stable than Fedora.

  3. floppy

    i am not sure what this last part means?
    “Whatever we decide Fedora’s ultimate purpose is, we should make sure it’s something that matters so people will care about it. Perhaps being a ‘multi-desktop distro’ isn’t really something that matters.”
    does this mean that fedora is not going to support xfce and kde now?

  4. drago01

    “Does our user base, as currently defined, still hold true for us as a project in describing who we want our users to be?”
    Depends on how you define “we” … I’d think we should try to not exclude users because the needs of users are not that different as you might thing.
    Some examples:
    Having a stable platform helps server users that don’t have to worry about stuff randomly breaking, as well as desktop users that don’t always have the knowledge to fix things and developers that want to focus on the application / component they work on and not have to worry about everything.
    Optimizing the system to better utilize resources helps embedded users as they try to run on resource constrained devices, laptop users that will benefit from longer battery life as well as cloud / virtualization users who will be able to run more instances on the same physical hardware.
    Having reasonable defaults and optionally allow customization helps user who just want to “get shit done” as well as not excluding the ones that want to “pick their own adventure”.
    I think you get the pattern there is no need in saying “we care about user type X” so we don’t care about “type Y”. Not everything is either black or white.
    So is there any problem with the user definition? Yes that would be the fact that it exists at all.
    Do we want more contributors? Sure every project aims to do that … that’s rather obvious. How do we do that? Is a discussion worth having. Increasing the user base as a whole will get us more contributors. Having lower barriers for contributing will as well. Less policy and more freedom (heck that is one of our foundations after all) will get us more contributors. That is especially true for people that contribute in their free time and not as part of their job.

      • drago01

        I tried to explain that in the above comment … basically it is 1) not required (i.e does not solve any real issue) 2) causes some user groups to fell excluded.
        I do consider 2) harmful.

        • mairin

          1) It is required at least for a standard design process. One of the first things you do in designing UX is to define who you’re targeting your design towards. The reason the user base document exists at all was because at the time we were overhauling the Fedora website and needed to understand who we were targeting it towards, so we asked the Board to please tell us.
          2) I think it’s nonsensical to conclude from saying ‘we prioritize our work based on this extremely broad target’ that we’re excluding anybody. You need to have priorities to get everything done. You cannot do everything 100% for 100% of the people 100% of the time – it’s just not possible. Time and resources are limited for one. For another, some decisions just fair one group of users more than another, and to be able to make intelligent decisions consistently across the project we need to know which group(s) we should prioritize for.

          • drago01

            I know we had this very discussion back then as well. Lets look at what others do. Whom does Microsoft target with windows if not “everyone” ? Windows is used by pretty much all types of user. What is Google’s target for Android? Same thing here …
            Trying to narrow the user base (ok our current definition is indeed very broad but that does not mean that the newly defined one has/will be as well given the discussion seems to focus on “people that want to build stuff”).

          • drago01

            Err rest of the sentence should read:
            “just results into less users, which means less general interests and in the end as a consequence of that less contributors.”

          • mairin

            microsoft windows is used by everyone but that doesn’t mean it was initially targeted by anyone. Look up any book or article by a reliable source on the design process. You’ll find one of the very first steps to any design process is defining your target audience.

  5. hector louzao

    we have to break with the follow myth about fedora:
    Fedora is for intermediate,
    Fedora is only for developer,
    Fedora is difficult to use,
    Fedora is not for server.

  6. floppy

    i was thinking about this post. i have some ideas for you. because that is what the comments are designed for i guess.
    1) to vague.
    the meeting reads like a dialogue between the existential philosophers contemplating the meaning of existence. that is all the fashion now but not very effective if you want your work to be efficient.
    2) total inclusiveness = excludes everyone
    there is a problem when you make strong gestures about including everyone and every thing in a happy fake tv commercial way. people know its fake. everyone has biases. everyone know that everyone else has biases. to say you have no bias is a lie. so people know you lie and also feel they will have no space to express themselves because of the “idea police”. that is not a fun environment. you can keep the door open and welcome them in but if the sign on the door says “everybody only” people just don’t want to go in. they are not everybody.
    3) gnome problems hurt fedora.
    does fedora need a default desktop? probably but the current gnome problems make some people avoid fedora because its “gnome land” all the hackers i know roll there eyes at gnome even if they think fedora is ok distro. does not seem like how hackers want to work. to much smoke and mirrors. shell is bad idea etc.
    4) redhat is fedoras biggest strength
    you want to get githup on fedora? good hackers know redhat. they know redhat is solid. running servers all over the world with solid review and strong engineering that can beat microsoft and other competition in the real world. they want you to show them how good you are. show them your code and what you offer and they can make the judgment. no one want to hear puff. everything is puff now. you have to prove you really are in touch and doing the job well. people see that in redhat but not so much in fedora.
    there are two types. “political” and “hacker”. political does not care so much about github, uses debian. those people will not use fedora anyways because they hate the “big corporation” redhat. you can’t fool anybody. hackers are all over github,.likes redhat, likes success, likes open source, likes google, gsoc etc.
    5) website looks bad.
    no ideas here. ovr all to much puff.
    very nice article. quite a lot you covered. hope these help the shadow council and the invisible hand etc.

  7. Gary Scarborough

    I read your page a few days ago and after letting the ideas stew for a bit, here are my thoughts. Personally I don’t think that the people behind Fedora have been doing too bad of a job. You continue to be in the top 5 distros on pretty much every list. The fact that you are a test bed for RHEL is an awesome strength. I think you have suffered recently due to some design decisions, but overall, Fedora is doing well. As a user and a system admin, I use Fedora so that I can keep up with what is coming up in RHEL. By the time the technology makes its way into RHEL, it will be old hat for me hopefully. As far as RHEL goes, I push it everywhere I can at work (RIT.edu) in our IT courses as it will probably be what the students will be working with when they get jobs. Here are a few criticisms I have about the distro:
    1. I think Gnome has hurt Fedora greatly. As a distro that has focused on Gnome for the main DE, Gnome’s problems are equated to Fedora problems. Gnome’s extreme change in design has left a lot of people unhappy. While I think the technology being used in Gnome 3 is awesome, I am less than hapy with the poor attitude I have seen from Gnome devs regarding user concerns. Gnome’s developers have alienated a large part of the Gnome community, hence the rise of Mate and Cinnamon. While some think that forking/competition is a good thing, keep in mind that any knocks Gnome takes along the way will be knocks that Fedora also has to endure. As a platform for future RHEL distros, I really have to ask: Is Gnome going to be appropriate for RHEL 7?
    2. The last several Fedora releases have had time slips and several awful design problems. Keep in mind that your users are trying to do just that, use Fedora as a full time desktop. When Gnome 3 first shipped with Fedora, it was atrocious. Why would you expect users to stay loyal when the main DE they are used to becomes something no where near what it used to be. The same can be said about firewalld. The tools for easily manipulating it were not ready when it first shipped. Even now in a Fedora 18 box, I ended having to rip out firewalld and add iptables back in because I couldn’t find the info on doing what I needed. I expect some things to be a little bumpy. After all, Fedora is cutting edge. But if adding iptables back was my easiest solution, should firewalld really have been the default choice yet? Why not leave the buggy unfinished software as an alternative people can install and try if they wish.
    3. Part of the issue with 2) may be your short turn over. You try for 6 month cycles, yet you clearly fail to properly anticipate the amount of time needed for some tasks. Otherwise you wouldn’t keep slipping. How about going to a 1 year or 9 month cycle with a beta release half way through. Many people don’t want to upgrade more than once a year and many skip every other release. I have to be honest, I have considered moving to CentOS or Scientific Linux as my desktop OS. The main issues preventing me from doing it are some missing libraries that don’t seem to get packaged for RHEL and a few things that I really need up to date packages of, like stuff dealing with KVM. Also the need for hardware support on newer laptops.
    Overall I am happy with Fedora and its community/developers, but I do see room for improvement. If you would like me to post these ideas somewhere else, I would be happy to do so.

  8. floppy

    @Gary Scarborough
    i have an interesting situation because at my job we use redhat/linux and windows. and its for desktop users with professional software. so they are serious power users and there working 8-10 hours a day on these systems on the desktop. they have to preform and get there work done. most of the applications are using qt. the desktop they can use gnome2 or kde. most of them use kde. they all use windows7. and there have been discussions about the next desktop roll out. i can tell you that no one wants to to roll out gnome3 or windows8. there is really no debate about that any more it is a fact that we are going to skip windos8. there is still debate if they want to force everyone to use kde or to let people use one of the gnome forks but gnome3 is not going to be an option.

  9. MrMagoo

    I totally agree that Gnome’s problems directly become Fedora’s problems. In the area where I work, we have a little over 5,000 desktop systems, and 30 z/OS systems in a parallel sysplex. All of the z/OS boxes have RHEL on them, and a most of the desktops as well. Some of the desktops have Fedora if they aren’t tied into one particular network. Word has come down from the top that under no circumstances is Gnome 3 to be installed on any machine in the place, and it’s pretty much immediate termination for anyone that does install it. They are still dissecting systemd and have yet to make a determination on it.
    So this mean that anyone that does install Fedora must make certain to get the full DVD install, or one of the other desktop spins, and not even install Gnome 3.
    So my thoughts are, if you keep defaults (and I agree there should be defaults), they must be able to be changed, and changed before the install. In my case, a default Gnome desktop install can get someone fired.

  10. I found this summary very useful. I found it easy to follow.
    As a voluntary user who is not a “power user”, I find myself drawn to Fedora’s strengths for my intended uses. Fedora works well “out of the box”. Fedora features lots of spins, including the lightweight LXDE spin best for the fairly low CPU used computer on which I installed Fedora 18. Fedora recognized the wi-fi on two drivers easily. Despite the concerns raised in some tech press that it is “not for newbies” due to its “free only” default, I found the rpmfusion libraries easy to install and to pose little real challenge. The large amount of available software is a plus for the distro. Fedora starts with a lot of positives for new users, which contrasts sharply with some of its public perception.
    I want to see Fedora remixes and spins continue to adjust to the rapid proliferation of new low-power
    CPUs. For example, I have a Pengpod. I use Android and Linaro on it now, because it looks as if the Fedora-based efforts are still not very mature, and fail to get the wi-fi to work. I’d love to see more support by Fedora as an institution for those working on the liliputian end of the spectrum. I notice some folks are working on ways to get Fedora to work on chips like the Allwinner. It would be fun if Fedora could be a player as a go-to O/S in the lightweight CPU movement.
    To me, Fedora’s variety is one of its strengths. I find the idea of it being seen as a monolithic GNOME
    O/S to be unfortunate. My feeling arises from more than my preference for lightweight DEs like
    LXDE or Razor-QT or E17. I think that one thing that’s great about Fedora is that it is simply too “big” to be locked down to one DE.
    But my thoughts are like anyone’s–tied to and limited by my own limited interests and my own limited knowledge base. The thing I liked about the meeting notes is that the issues being addressed looked lik the “right” issues. I’ve been very impressed with Fedora, and it’s good to see that its planning makes sense.

  11. Pingback: Fedora for short-lifespan server instances | Groveronline

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.