Fedora

Amen, brother

A great blog post that Jon Masters made today pretty much sums up my frustrations with Fedora. I was really happy to see it! Jon basically lays out four things he’d like to see Fedora do:

  1. Slow the updates – they bring too much instability to what is supposed to be a stable release.
  2. Set a mandate – either ask the users or set it yourself, but take a stand. “Fedora can’t be all things to all people, that isn’t working,” OMG yes, preach it!
  3. Establish cross-functional workgroups.
  4. Set specific goals. “Have a strong process (stronger than now – before anything is allowed to be built and shipped out the door), with an overall vision that those ideas fit into.”

I wanted to point out some of the efforts the Fedora Board (current and over the years) has made thus far to rectify the situation (which, by the way, I think is a natural symptom of growing as a project and not too much cause for alarm given that we right it soon.):


“Panda on Crib” by Jan Alvin Dimla on Flickr, used under a CC-BY-SA 2.0 license.

Here’s how I see these interacting:

  • The mission statement sets out what we are doing. We’re focused on free and open source software and content. While running a panda nursery or developing new and exciting flavors of ice cream might have some positive value to society, those are out-of-scope for our project (damn! ;-) )
  • The target user base sets out who we are doing it for. Sometimes it can be hard or even impossible to please two groups of people at once, for example, how can you please both New York Rangers and New York Islanders hockey fans at the same game? The target user base defines the team we’ll be rooting for when conflict arises and a decision needs to be made (the Rangers, of course!)
  • The four foundations set out how we do it (from a values perspective.) There’s a lot of different ways you can get to the same end goal. For example, I could raise money by holding a bake sale with cookies and cakes, or I could raise money by selling tickets to the circus. If my organization is like Weight Watchers, I’ll probably opt out of using decadent baked goods since it conflicts with my organization’s core values towards healthy eating and healthy weight. I might sell tickets to the circus instead. If my organization works to protect animals, though, supporting a circus might be a questionable way to represent my organization’s values, and a bake sale might be a better choice.
  • The vision statement will set out why we do it and where we want to be. While driving around in a car without a destination or purpose in mind can be fun, it’s not necessarily going to be very productive. This isn’t to say that open exploration or innovation is a problem – not at all – but there are ways of doing that without being a total flake, I think. I think arguing that this kind of free meandering about will eventually lead to a great user experience does a great disservice to the time, effort, and discipline that is in reality required to achieve that.

So, what you may ask, and what I ask myself all the time, is – are these fancy documents really going to be of any use? I think Fedora has grown up enough that they really are necessary. It’s not a close-knit group of folks in North Carolina putting a Linux distro together. I think small, tight teams don’t need fancy documents because they work so closely together, they share a common culture and can work big-picture concerns out. When you’ve got a large global project with several thousands (if not more) contributors across the world working on it, it’s a bit harder to get everyone on the same page and resolve misdirections before they go too far. Thinking walking your dog vs. herding a hundred cats. Although I think maybe Smooge can speak more authoritatively on how Red Hat Linux was put together vs. how we put Fedora together today.

No, I don’t think finally having a vision statement as the final puzzle piece in place is going to magically fix everything. I think there’s a lot of work that is going to need to happen to put all of these things into working use. I think the documents and more importantly the ideas behind them are going to have to be socialized into the culture of Fedora to really be successful. That is why it is so important to get involved in the discussions around them if you are part of Fedora. So please, do :) Hit up the Fedora advisory-board mailing list to start, and maybe drop by a public Fedora Board meeting.

About Máirín Duffy

Máirín is a principal interaction designer at Red Hat. She is passionate about software freedom and free & open source tools, particularly in the creative domain: her favorite application is Inkscape. You can read more from Máirín on her blog at blog.linuxgrrl.com.

Discussion

74 thoughts on “Amen, brother

  1. One pitfall that is important for mission statements is that they do not become negative statements. Don’t define yourself by what you aren’t, define yourself by what you are.

    Posted by Corey Burger | August 28, 2010, 11:11 pm
  2. I think the cross-functional group area is probably the single biggest deficit that Fedora has. His point #3 in the original article was excellent, and if Fedora intends to continue focusing on actual use-cases, as opposed to commits, tickets, and updates, then it will have to change the way it operates.

    Posted by twilightomni | August 28, 2010, 11:42 pm
  3. Here’s a concern I have with the target user base definition.

    Because of the way Fedora has operated by default, with its very quick turnover, it has ceased to be a viable distribution for server use. It’s possible that in the emerging era of cloud computing (not as a buzzword — the real thing), there’s actually a place for it now. Or it’s possible that by including that in the target definition, Fedora could be made more relevant in that space. But as it is, those users are largely gone from the Fedora user community, except for those few of us who have been involved so long that we are still around because we really care about it. So a survey of current users is very likely to find that server use is not a major concern.

    One might say, okay, well, too bad for that, but here’s where we’re focusing now, and so much for server use. We’re aiming for the web/desktop productivity user. But the thing is, that’s a really narrow slice of real-world Linux. There’s only a few people (myself among them, but never mind) hard-core enough to make Linux their exclusive (or even primary) productivity environment. “Serious” Linux use still means on the server.

    That doesn’t mean that Fedora’s design direction needs to be for the server. There’s a lot of awesome desktop work coming from Fedora and Red Hat developers. That shouldn’t be discounted, and by all means let’s show off that strength.

    But, it _does_ mean the server can’t be forgotten. Think about how Microsoft got its OS into the data center… we even have some OS X servers around here for exactly the same reason: people like to have a common environment on their desk and on their servers. RHEL6/CentOS 6 is going to cause a major, inevitable drop in Fedora desktop use for the people who see it this way, and if Fedora ends up diverging in a server-unfriendly direction, we won’t get those people back. Meanwhile, users will naturally be drawn to use distributions which cover both desktop and server use in one general-purpose offering — at this point I’ll dare to mention Ubuntu by name.

    The original statement of the target user base was _very_ careful to equivocate on the scope of the target audience, and to limit it to defining the design goals for the “default offering”, which is the Live CD. I’m a little worried to see people now looking at that statement and skimming for excerpts, and inadvertently assuming the earlier work was meant to define a design target for all of Fedora. It’s not: it was _intentionally_ meant to be narrower than that, and it’d be a mistake now to reapply it more broadly. For example, at the end of the board meeting, Seth said “So we’ve precluded servers from the target. That’s fine, I just wanted to confirm it.” — but that’s not actually the original case, when taken in context.

    Posted by Matthew Miller | August 29, 2010, 12:27 am
    • I don’t think choices are mutually exclusive. I get so frustrated when people make that assumption.

      Posted by mairin | August 29, 2010, 12:44 am
      • That’s why I think any “target audience” statement is contra productive. We create an operating system and user A would use it for foo, B for bar and C for baz.

        Why should we tell B “screw you you aren’t in out target audience” ?

        The main points (that seem to result into people wanting a specified target audience) seem to be:

        1) Being easy to use for new users – Who on earth is hurt by that?
        2) Being stable as in stuff doesn’t break all the time – see above
        3) Want to use it on a server / desktop / laptop / kiosk / whatever … that has not to be mutually exclusive as you said.

        So the only thing a target audience statement gains us is telling people upfront “go somewhere else” without a reason for that.

        “The target user base defines the team we’ll be rooting for when conflict arises and a decision needs to be made”

        Which kind of conflicts do you had in mind when writing that? I might be missing the obvious but I can’t think of any decision that really has to be made on a target audience basis. Worst case is it comes down to common sense.

        Posted by drago01 | August 29, 2010, 4:08 am
        • Drago01, I don’t understand why picking a target user means all other users are told “screw you” – where is that in the target user definition again? See my other comment about mutual exclusion.

          Posted by mairin | August 29, 2010, 12:50 pm
          • “where is that in the target user definition again” .. err I didn’t say that the target user definition includes the words “screw you” nor did I imply that.

            My point was basically if you say “this product is for the users X, Y, Z” any user who reads that and does not fit to the noted description feels excluded. (whether this was the intend or not).

            “See my other comment about mutual exclusion.” See “3)” in my above comment … there is no disagreement regrading that ;)

            Posted by drago01 | August 29, 2010, 2:36 pm
            • If that is the case, those users are reading something that IS NOT THERE.

              Posted by mairin | August 29, 2010, 2:55 pm
              • But the problem is that they do read it like that. A statement can contain more information than it actually includes or even intends. There is a room for interpretation. In this case we have an inclusive statement titled target audience which is easy to read by people as “I don’t fit in this categories”.

                This is basically the same effect that happens when some women complain that some statements are sexist because they only talk about “men”, even though it meant to be gender neutral.

                And in our case I don’t think that we need any such statement at all IMO.

                Posted by drago01 | August 29, 2010, 3:13 pm
                • If I avoided doing something for fear someone would misinterpret it, I might as well be a vegetable or a rock.

                  Cute tie-in to sexist statements. Too bad it’s completely off-base.

                  Posted by mairin | August 29, 2010, 3:16 pm
                  • “Cute tie-in to sexist statements. Too bad it’s completely off-base.”

                    Ugh … sorry seems like I picked the worst possible example (noticed that shortly after submitting it).

                    But anyway I don’t understand why you feel the need to be overly defensive here (stop reading something that is NOT THERE ;) ).

                    What I am trying to say is

                    1) We have a statement that can easily be misinterpreted
                    2) That we actually don’t need and therefore can simply omit to avoid unnecessary confusion.

                    Posted by drago01 | August 29, 2010, 3:25 pm
                    • I think you are making a bit too many comments a bit too quickly and you should probably think about this a bit more and come back later before replying again.

                      I understand quite clearly that you are trying to say we don’t actually need a statement “and therefore can simply omit to avoid unnecessary confusion.”

                      What I have tried to tell you in response, over and over again in so many ways, is that *I* as a professional interaction designer involved with Fedora *need* that statement to help support me in my design work. The lack of such a statement in the past had led to me needing to make spur-of-the-moment decisions on my own that I should not have to just to get a design put together at all, and it has also resulted in me receiving a great deal of criticism from those folks who I decided not to support as strongly as others in the design work. The project should know towards whom its efforts are focused. When I am trying to create a simple design in order to support the project, I should NOT have that as a decision weighing down on my shoulders. It should be established. I want to create designs that help Fedora, and I want to be able to put my best foot forward as a designer when doing so. I do not want to have to wade through nasty politics in order to do so nor do I think I should have to.

                      Posted by mairin | August 29, 2010, 3:37 pm
              • I think this misreading can happen on both sides. People can feel left out, and other people can use it to justify decisions that are bad for use-cases outside of whatever the defined target happens to be. I’m not saying _you_ do it, but people clearly do.

                I’m not at all against defining a target, but I think it’s really important for the language around the target definition to be clear as to this point.

                I *would* also like to see any further official refinement of the desktop target for Fedora _also_ enshrine some core principles which are helpful for a broader Fedora audience.

                Posted by Matthew Miller | August 29, 2010, 7:50 pm
                • Hi Matthew, I think I understand your points a lot better now and I do agree. There are decisions that might be made in the best interest of desktop users but could completely (and unnecessarily) make life for folks using Fedora as a server of some type really difficult.

                  I think since Fedora’s code base is eventually branched to make a commercial server OS that decisions that would make life harder for sysadmins tend to not happen since folks in Fedora don’t want to hurt RHEL & CentOS, but, maybe that does need to be codified in these statements somehow. Maybe by keeping the desktop target statement but having an additional target user statement for all of Fedora that mentions ability to be a server (within some defined scope) as important?

                  Posted by mairin | August 29, 2010, 8:04 pm
        • “Which kind of conflicts do you had in mind when writing that? I might be missing the obvious but I can’t think of any decision that really has to be made on a target audience basis. Worst case is it comes down to common sense.”

          I could spout off 20-such questions that have arisen during the http://www.fedoraproject.org redesign. Which spin is the default? Should we have a single download button, 6, or 20? Is this tutorial appropriate for the site or is it aimed at the wrong audience? Is this the right color scheme / layout / feel or does it miss the mark?

          Having a single download button obviously doesn’t prevent people who want their Fedora how they like it, hold the mayo, extra tomatoes, and no sesame seeds on the bun, obviously, But it does make a world of difference to folks who don’t know and don’t care about the 50+ combinations of Fedora you could possibly have.

          Posted by mairin | August 29, 2010, 12:52 pm
          • Again you missed what I tryed to say … picking a default and having an easy to use and clear UI does not require any target audience its just common sense.

            So it is made easy for those who would be confused by lots of options, while still making it possible for others to find what they want (the ones with specific needs know how to get what they want).

            My point is we don’t need a (perceived) “if you don’t fit into category X, go somewhere else” statement.

            Posted by drago01 | August 29, 2010, 2:41 pm
            • “picking a default and having an easy to use and clear UI does not require any target audience its just common sense.”

              I really wish it were simple, but it’s not. If it was, the warfare that exploded over the new get.fedoraproject.org design would not have happened.

              Posted by mairin | August 29, 2010, 2:50 pm
              • Well no. The “warfare exploded” because random people wanted to make user interface decisions even thought they aren’t user interface designers or aren’t at least experienced in this area.

                User interface design is perceived as “can be done by anyone” which is clearly not.

                I am not a user interface designer myself but that is basically what I learned while working on projects such as gnome-shell. Everyone acknowledges and respects expertise in the coding/technical area but feels like user interface design does not require any.

                Posted by drago01 | August 29, 2010, 3:18 pm
                • “Well no. The “warfare exploded” because random people wanted to make user interface decisions even thought they aren’t user interface designers or aren’t at least experienced in this area.”

                  Wrong. If you read through the threads you’ll see quite clearly the loudest complainers explicitly stated they had no desire to be involved in the design process at all; they felt that the UI design was not targeted for them explicitly and were upset. It wasn’t. The default download page was targeted for people who have no idea what all the 50 flavors of Fedora are because we want them to be able to try it. For folks who know a bit more about their options it was felt they could ‘opt out’ of the simple interface and click through to a more ala carte style interface. Many of the complaints about the redesign centered around the more advanced users being forced into that single extra click.

                  “User interface design is perceived as “can be done by anyone” which is clearly not. ”

                  Earlier you said that “picking a default and having an easy to use and clear UI does not require any target audience its just common sense.” UI design cannot be both “just common sense” and something that cannot be done by anyone.

                  “I am not a user interface designer myself ” – yep! If you crack open any of the widely-lauded manuals on interaction design such as Alan Cooper’s About Face you will see that defining the audience for a particular interface is one of the first steps you need to do when following a proper design process. In some cases the user is implicit, e.g., it’s a systems management product – the target audience is system administrators. But even general-purpose products had a target user in mind. The example Alan Cooper uses is airline flight crew members – the target user for today’s rollboard suitcases that everyone uses. The inventor specifically worked to create a product specifically for airline crew, and ended up with a product that helped everyone. For simple products it may be possible to just create the product and see where it sticks, so defining a target audience isn’t always a necessary step to success, but not defining it certainly can lead to a lot of wasted time and effort should it not just ‘stick’ with anyone.

                  Posted by mairin | August 29, 2010, 3:30 pm
        • Because by not being clear about an audience and instead cling to mythical Libre notions of “everyone”, you are inadvertently already saying, as you state, “screw you” to everyone.

          And while you might think it is counter productive, the alternate viewpoint is summarily disregards a vast body of design thinking and theory on the whim of an incorrect guess.

          Design ultimately can be distilled down to needs.

          Needs are rooted in historical context, age demographics, cultural demographics, religious demographics, body type / physical contexts, and many other complex areas.

          While we may deeply desire to create this mythical and Utopian system for “everyone”, I’d encourage you to do consider the ability to write a book, create food, write music, create a movie, fabricate a piece of furniture, build a house, etc. for _everyone_. Sound foolish?

          Further reading:

          The fallacy of “intuitive” and “easy” and other meaningless words:

          http://www.uigarden.net/english/easy-intuitive-and-metaphor

          The fallacy of a global culture as needs driver. Pay close attention to the cell phone designed around religious belief. If your belief system is strong enough, it creates a need. That need may very well have a priority that supersedes other needs:

          http://www.uigarden.net/english/global-market-global-emotion-global-design

          And finally some illuminating quotes from Havoc Pennington from 2006. This is the earliest serious discussion in Libre circles that I have seen regarding practical considerations for audience.

          http://www.mail-archive.com/desktop-devel-list@gnome.org/msg04377.html

          http://www.mail-archive.com/desktop-devel-list@gnome.org/msg04387.html

          http://www.mail-archive.com/desktop-devel-list@gnome.org/msg04403.html

          Without an audience, it is all empty words such as “easy”, “simple”, “intuitive” and other worthless terms.

          Audience governs all.

          Posted by Troy James Sobotka | August 29, 2010, 4:02 pm
      • That’s a really good point. The requirements of server use are basically: reduced churn within a stable release, the ability to install and run a minimal system, straightforward config files, and good command-line tools and documentation. (And of course a longer lifecycle would be nice.) None of those things are harmful for the desktop/productivity case, and in fact can benefit it.

        But I think it’s a legitimate concern that if the server case is left out of the definition, decisions will be made that go against those needs. Basically, it’s a minority voice that needs to be “constitutionally protected”, so to speak.

        Posted by Matthew Miller | August 29, 2010, 8:49 am
    • (I need to write a proper reply but it’s too late, btw)

      Posted by mairin | August 29, 2010, 1:13 am
    • Who do you improve that situation?

      IMHO Fedora isn’t a good choice as server OS because:

      – short support time for every release (a server needs a longer support cycle than a desktop)
      – there’s not a stable/reliable/secure/fast way to upgrade to the next release (that’s not a must if the support time it’s long enough)
      – sometimes (rarely, but happens once in a while), a update break something

      Those are THREE points, and I don’t think Fedora can address both three right now.

      Actually, I don’t use Fedora in a server just because of that three points, but because other Linux distributions provide 2 or all 3 points.

      Posted by Juanjo | August 29, 2010, 3:53 am
  4. “Slow the updates”

    Very good point. About 5 months ago I read about “update policy” proposal, but I haven’t seen anything more about this. I think I remember a post about it in the planet, and this:

    https://fedoraproject.org/wiki/User:Adamwill/Proposal:Package_update_policy

    Currently Fedora updates bring me some kind of fixes that I really appreciate, although sometimes you know that something can get screwed (back in April a kernel update that was unable to boot my laptop was pushed into F12).

    In that topic, I think Fedora has to improve the upgrade process (from Fn -> Fn+1), but that’s more likely a wish that something that can be improved just by a policy :)

    My two cents!

    Posted by Juanjo | August 29, 2010, 3:47 am
  5. Bravo! (both your and JonM’s post)

    As a user, these are the sorts of topics I want to see being discussed.

    Posted by anon | August 29, 2010, 4:07 am
  6. Switch to Debian or Ubuntu? Cause we have all of those 4 points. Debian has larger cycles, but testing is generally good so your average upgrade happens ~12-18 months. Ubuntu has much stricter updates policy than Fedora with a support lasting 18months for regular release or 3-5years support (Long Term Support) releases every 2 years. LTS do get slightly more updates, simply due to longer support period (so on average it’s the same). I find myself upgrading Ubuntu every 7-9 months depending on where I’m in the University year. I don’t recall updates breaking anything for me since 2007.

    Posted by Dmitrijs Ledkovs | August 29, 2010, 9:26 am
    • Sorry, Ubuntu doesn’t meet the cultural & values requirements I have in an OS. Bzzzzt

      Posted by mairin | August 29, 2010, 12:54 pm
      • On the other hand, if you just focus on the quality of the software itself, then they’re neck-and-neck, and Fedora might be getting edged out. Ubuntu formally places priority on user experience, as far as setting out infrastructure, priorities, and authority for those efforts. Mairin, you seem to consider Ubuntu anathema, but I wonder with curiosity what really cool things you could get done if in some alternate reality you were a member of the other.

        I also imagine that proponents of Ubuntu community culture don’t think of their values as “beneath” Fedora culture.

        Posted by twilightomni | August 29, 2010, 1:18 pm
        • “Mairin, you seem to consider Ubuntu anathema, but I wonder with curiosity what really cool things you could get done if in some alternate reality you were a member of the other.”

          Wishful thinking, sorry. I could never condone the compromises they make wrt freedom especially in terms how closed-off their design team is. Treating design as some sort of ivory tower and considering community design contributors as somehow not capable just makes me sick. Never mind the use of OS X.

          Posted by mairin | August 29, 2010, 1:29 pm
          • I don’t think they see it that way. I know this whole article isn’t about Fedora’s design initiatives, but Ubuntu’s compromise seems to be not about where the designers come from, community or not — but that they must actually be invested with some power over the direction of development to enforce their design. And I think that’s practical, not closed-off. If you mean closed-off by “they don’t listen to anyone”, then that’s obviously false. If you mean closed-off by “other people can’t do what they want”, that’s debatable, as long as the code is open. If you mean closed-off by “people can’t make *Ubuntu* do what they want”, that’s called not being the programmer. I can’t make Lennart focus on my own personal points in Pulseaudio. It’s his project.

            The whole point of *these* discussion points seems to be that part of Fedora’s unnecessary churn is because developers push updates and code willy-nilly on their own time. The only way to change that is to change control. There is no self-policing way to make Fedora spontaneously craft itself into a better operating system — that is, better than the process is already working itself – and as Scott mentions below, it hasn’t done too bad. It just has some built-in limitations, and the lack of design mandate power is one of them.

            I’m not picking on your use of OS X. I…think that’s what you mean. Kind of an obtuse sentence.

            Posted by twilightomni | August 29, 2010, 1:49 pm
            • “The only way to change that is to change control.”

              I disagree completely. Right now the culture is an open bazaar where almost anything goes. I think if the culture can be changed to follow better practices many of the problems can be righted. I don’t believe it’s a power issue, I believe it’s a “not everyone is on the same page” and “different groups are pursuing goals that are at odds with each other” problem.

              Re: the OS X comment, OS X and other proprietary software (e.g., Basalmiq) is used in the creation of many design artifacts in Ubuntu. I have heard this from Ubuntu developers and witnessed it with my own eyes.

              Posted by mairin | August 29, 2010, 2:54 pm
              • “the culture can be changed to follow better practices”.

                This is the part I’m not sure about when you say you disagree with control. How can this possibly happen without it? Are we talking about the same “control”?

                What is control? Manipulating motivation is one. Dis-incentives to push your packages or something similar, is a form of control as well. I’m not sure what definition of control you are reacting so negatively against, since Fedora contributions will have to be channeled or monitored in some way for anything proposed in these articles to actually take place.

                Posted by twilightomni | August 29, 2010, 3:27 pm
                • “Control” as you are using it is authoritarian and involves enforcement. I’m speaking about cultural change because I feel folks should not be forced into doing something; they should understand why and want to.

                  Posted by mairin | August 29, 2010, 3:40 pm
                  • As far as I can tell, the ability to “enforce” policy was actually a central point of the transcripts in your last article on the Fedora Board Meeting minutes. If a policy can’t be enforced and no one can be held responsible for its implementation, what good is it?

                    I say this rhetorically, as I know obviously FESCO has purpose in its existence. I just don’t see how Fedora can actually get some of its desired large-scale changes accomplished without some type of authority invested in something, anything. It can’t merely be people doing what they feel, or there will always be anarchists contributing to your OS with no regard for the vision or direction of the others who contribute to it as well.

                    Quite frankly, that’s not how I want an OS that I use to be built, but I’ve realized since then [and I think you'd agree] that merely wanting a good OS doesn’t make me Fedora’s target audience.

                    Still, if the goal of the Fedora contributors is to deliver a product as a whole, rather than a bundle of packages that requires people like you [and I mean this as a compliment to your hard work here] to come in *post-hoc* and provide all of the design that actually makes it a presentable product to distribute to real people, then they should be willing to put their personal motivations aside to cooperate toward that, and also be willing to hold each other to account for it – via an authoritarian intermediary if “best behavior” guidelines don’t work.

                    What’s that Einstein quote? You can’t solve a problem with the same level of intelligence that created it. Fedora is not “problematic”, but its processes seem to be, and no one in the community can seem to transcend their understanding of the whole system to realize what must be done. I’d say you might be against the idea of any one person “doing” anything to the “system”, but can a community ever really become self-aware of its own flaws?

                    Posted by twilightomni | August 29, 2010, 4:42 pm
                    • “It can’t merely be people doing what they feel, or there will always be anarchists contributing to your OS with no regard for the vision or direction of the others who contribute to it as well.”

                      Perhaps by defining what we are about more clearly that can be broadcasted such that the ‘anarchists’ can at least be defined as ‘anarchists’ because right now we have warring camps and no single camp can be named as the one to follow.

                      Posted by mairin | August 29, 2010, 7:56 pm
  7. While it’s well and good that some people want more of a set stable release, keep in mind that there are some people that like Fedora in part precisely *because*, while it may not be rolling, it’s not as conservative as most other big-name distros and actually has some more major package updates during the lifecycle of the stable version. Do we really need all big-name distros to be clones of each other?

    And no, Rawhide is not the answer. It goes too far the other direction in terms of stability, and doesn’t even have signed packages.

    Posted by marvin | August 29, 2010, 10:32 am
    • Where are these people?

      Posted by mairin | August 29, 2010, 12:58 pm
      • Apparently not on Planet. They’re on the forums though, if you remember Adam Williamson’s informal poll last year, where more adventurous updates trumped conservative updates by quite a large margin. Again, not saying your opinion is invalid, but don’t think that most people agree with it just because it’s a hot topic on the Planet right now.

        Posted by marvin | August 29, 2010, 1:06 pm
        • I don’t actually, reference?

          Posted by mairin | August 29, 2010, 1:08 pm
          • He’s referring to this forum thread:

            http://forums.fedoraforum.org/showthread.php?t=241710

            Btw, there already is a yum plugin which lets you restrict the updating of packages to those with security issues. While it may not be sufficient to make people happy again (and as far as I know lacks integration in the graphical package management tools), it could be extended to give the user the ability to choose between adventurous and more conservative update schemes. Telling the adventurous folks to use Rawhide because Fedora n is a stable release isn’t any other than telling the conservative people to switch to RHEL/CentOS because Fedora isn’t an enterprise distribution.

            Posted by mindfaq | August 29, 2010, 5:07 pm
            • Unless that plugin is install and set to security only by default, it’s kind of useless :( The very users who need it most will be the users who will not want to or be capable of opening up a terminal and modifying the config files as needed.

              Posted by mairin | August 29, 2010, 7:53 pm
              • It’s also problematic because you never know when an update which is nominally for security will also pull in a new version of something it was linked against. It works best if preferring to only update for security fixes and serious bugs is a coherent policy. (That said, the plugin is a great thing.)

                But no matter what, the core issue with this kind of stability is that backporting fixes takes a lot of knowledge and development resources. It’s totally reasonable of Fedora to decide that that’s not where we want to spend those resources, when there is already RHEL available.

                So, of all the things I’ve outlined, I think update stability is actually the least of my concerns.

                Posted by Matthew Miller | August 29, 2010, 8:12 pm
      • Hi! This is one of the reasons that Fedora interested me; I’m currently dabbling in various flavors of Linux, and I’m currently giving Fedora a shot. I have only been using it for about a few weeks, though, so I haven’t had the misfortune of having an update break anything on me.

        Dunno if this is a good sounding off board for ideas, but there are a few problems with Fedora that bug the heck out of me, and may prevent me from being a long-time user:
        * The update manager’s GUI is… weird and unintuitive. For example, it does not tell you the download rate of the packages or any sort of estimate for how long the download will take. Also, the only way to tell what is downloading/installing is by looking at the status column. If you have a lot of updates, the GUI will scroll to the last package that had a status change, which is terribly unintuitive. This would only work if the updates were listed in the order that they are installed in. At the least, tell me how many packages are installed and how many are left to go; when I had over 100 updates, I had absolutely no idea how much time I had left to wait. I’ve resorted to just using yum on the command line, which provides much more useful metrics and output, but I’d like to have a nicer GUI.
        * There’s currently a bug that prevents yum’s update manager and yum itself from installing updates from third party sources, like RPMFusion. I found the actual bug page before, but I can’t find it now. The workaround is to run yum with –nogpgcheck on the command like. This is a huge deal for me, since I had to use RPMFusion’s third party repository for mplayer. Since I’m completely new to Fedora, I have no idea why mplayer is not in the main repository, but I’m sure there’s some reason. However, if I didn’t know how to use the terminal, I would just think that my Fedora install is broken and try some other distro, since no updates would install.

        Previously, I’ve only used Debian-based distros, and I miss apt a bit. I’ve been reading Planet a lot lately (just found the site a few months ago), and I really like reading about what’s going on in the Linux scene. It would be really cool to be a contributor someday as a programmer (I’m currently a university student, graduating next year with a Bachelor of Science in CS). It seems sortof difficult to jump into, though, and it’s hard for me to figure out what I would like to contribute to. Plus, I’m not sure what sort of development environment and tools are commonly used for the larger C and C++ projects… I wonder if there’s any information about that anywhere…

        Well, I went off on a tangent there! I suppose I’m done with this comment now…

        Posted by Nintendud | August 29, 2010, 8:38 pm
      • I happen to be one. I have appreciated the KDE updates to the next major version (i.e. from 4.3 to 4.4, and the hopefully soon-to-be 4.4 to 4.5 update) greatly.

        Posted by Matt Rogers | August 30, 2010, 11:31 am
  8. Argh… we hear these same arguments every single year. The only thing that changes is the wording. Fedora isn’t going to change so forget about it. There was a time when I wanted it to change too… but I ran into the “I can’t do that Dave” Hal voice over and over… so I gave up… got used to it… and have been at peace with how Fedora is.

    The properties that you guys and gals want… are already available in a handful of other distros. The properties that Fedora has now… you don’t find in any other distro… so the chaos that is Fedora… is what makes it Fedora.

    The only way you could possibly change Fedora is if it became a commercial entity and there was a boss and employees… and the employees had to do what the boss says. That ain’t going to happen. Developers make Fedora what it is… and they work on whatever they want to… except for those that are Red Hat employees being paid to do Fedora work. To maintain the “First” of the four F’s you have to keep pushing the bleeding edge. Fedora has been responsible for so many technologies getting into Linux faster, getting refined faster, and then eventually being adopted by everyone else. If you want to find someone else to do all of that innovator stuff and just be a follower now so be it… but I don’t see how that is going to work.

    For all of Fedora’s faults… oddly it has worked out quite well… at least from where I’m sitting. I do realize that it took a lot of hard work from a lot of individuals and I do appreciate it. Whatever happens, I’ll still be following Fedora so don’t take this post as some sort of ultimatum.

    Posted by Scott Dowdle | August 29, 2010, 12:17 pm
    • “For all of Fedora’s faults… oddly it has worked out quite well… at least from where I’m sitting. ”

      Not from where I am sitting, sorry. It’s a MESS.

      Posted by mairin | August 29, 2010, 12:59 pm
      • I think his point is that the problems Fedora faces that gain the most attention, are in fact inherent side-effects of the Fedora developer community power structure, and unlikely to change as long as most Fedora structure remains in place unchanged.

        Posted by twilightomni | August 29, 2010, 1:19 pm
      • It might be a mess but it is OUR mess… and a mess many of us have grown to love. The things that definitely can’t be changed though are the 6 month release cycle and the short support cycle. Those two things define many of the dynamics you have to work with… but I guess you could slow down the updates. Some of the changes can’t hurt so go for it. I wish you luck.

        Posted by Scott Dowdle | August 29, 2010, 1:41 pm
        • Sorry, I don’t love messes. I love functional operating systems that I can share with my friends and family to bring software freedom to their lives. Wish we had that.

          Posted by mairin | August 29, 2010, 2:57 pm
          • Wow. I guess you decided Fedora was a non-functional OS that you can’t share with your friends and family to bring to them software freedom. Sorry to hear that. It works for me. I don’t want to sound too snide because I do appreciate your hard work. You are a fantastic artist and I have no complaints with the work you’ve been doing. If you can help change Fedora to make it better, go for it. I’m mostly just a user and advocate… and don’t do much in the way of real contribution so my opinion doesn’t matter.

            This whole argument could also be applied to the mainline Linux kernel. There is a ton of change with it each release… and a major new release comes out every 2.75 – 3 months. Only a few kernels actually get picked by major distributions and used. There is no real master plan for the Linux kernel. It is moving in all kinds of directions… and the chaos factor is huge… but yet it works. Does that mean that it can’t be improved? Sure it can but I realize I’m not the one to tell those who are doing the work and understand the system in place much better than I do… how they need to improve things. Fedora is a lot like the Linux kernel.

            Perhaps someone in the Red Hat leadership has decided a more centralized direction is needed and you are just part of the effort to make that happen. Or perhaps it is totally antonymous. Either way is fine. Make it happen.

            Two years ago I was saying many of the same things being said now but they fell on deaf ears. My suggestion on the massive amount of updates was to refresh the install media once a month. I was told that was impossible. There are some vocal folks telling you that you can’t make the changes you think are necessary. If you can… go for it. I won’t stand in your way. Good luck. I hope you prove me wrong… but don’t be crushed if you can’t. Just keep on keeping on.

            Posted by Scott Dowdle | August 29, 2010, 4:10 pm
            • Hi Scott, I never said Fedora was non-functional so please don’t put words in my mouth. I said I wanted an OS that would be functional for friends and family. When you have to drop to a terminal, for many people, that means non-functional. Obviously it doesn’t mean non-functional to everyone since we have plenty of users.

              Anyway no hard feelings, I’m just trying to make Fedora better. :) Thanks for the discussion!

              Posted by mairin | August 29, 2010, 7:47 pm
          • A key question which I haven’t seen answered is how many other people are prepared to do the work necessary to make that goal statement a successful reality.

            It’s a nice thing to want, a nice thing to wish for, but do we have enough of a critical mass of a team to overcome inertia and nudge the ball more firmly in that direction?

            I’ll tell you right now, I don’t think I’m one of those people. I won’t get in the way of that goal but I also don’t think I can be relied on to significantly adjust how I’m doing packaging contributions to make significant headway towards that goal. Looking at my current packaging activity, my contributions are primarily self-serving and work oriented. Not to hold up myself as typical, but I do have a concern that there a lot of people like me in the long tail of the active contributor pool who are at best going to see a change in primary focus as a non-event on their day-to-day contribution activity.

            What I’d to see, to gain some confidence, is a group of people standing up in key areas who are going to form the spine and central nervous system of the team who are going to work towards the goal you just expressed before the project as a whole commits to do that work.

            Posted by Jef Spaleta | August 30, 2010, 12:24 pm
  9. Slow the updates – they bring too much instability to what is supposed to be a stable release.

    I don’t know about other people, but I like using the latest stable software. Why should I wait a couple of months until, let’s say, the latest stable mono release is available in Fedora’s repositories? Or if you think that package is too critical, replace it with MonoDevelop, Eclipse, vim, git or tryton. If the developers of a program say that the release is stable, then it should work. I don’t see too much value in delaying so much its release to the general public. If it doesn’t work, bugs are reported, they get fixed and everyone is happy.
    Let’s not forget the fact that people running Windows are able to run the latest software the day it because available upstream and I haven’t heard anyone complain about this.

    If my organization is like Weight Watchers, I’ll probably opt out of using decadent baked goods since it conflicts with my organization’s core values towards healthy eating and healthy weight.

    No, you sell cookies and cakes to get new members for your organization :-)

    Posted by Cristian | August 29, 2010, 4:21 pm
    • “If the developers of a program say that the release is stable, then it should work.”

      If only :)

      There has been one ‘stable’ Inkscape release in recent history that I can recall that really obliterated my ability to get work done until I downgraded. The bug involved the font selection dropdown embedded in the toolbar, rendering it completely unusable. Having to go to the full-blown font selection dialog slowed down my workflow frightfully. I believe a Summer of Code student had worked on rewriting some part of the font system, and that bug was known about and released since there was a ‘workaround’ to go through the menus to pull up the full font dialog.

      Posted by mairin | August 29, 2010, 7:51 pm
      • I agree that releases aren’t always great, but I still think that problems should be fixed at their root. Also, if the old version would have been used, I think that other users would have been complained about the lack of features that were available in the new version (nicu comes to mind :-D).

        Posted by Cristian | August 30, 2010, 1:21 pm
    • I don’t know about other people, but I like using the latest stable software. Why should I wait a couple of months until, let’s say, the latest stable mono release is available in Fedora’s repositories?

      New Fedora releases come every six months! It’s not like you have to wait years. And if you want some specific thing sooner, you can always pull that individual package from Rawhide. That way, you can have the newer version of your specific thing, without the uncertainty about components that you don’t care so much about the freshness of yet still depend on

      Posted by Matthew Miller | August 29, 2010, 7:54 pm
      • I don’t think that rawhide packages can be mixed with regular ones.

        Posted by Cristian | August 30, 2010, 10:49 am
        • Sure they can. Usually. :)

          yum –enablerepo=rawhide install bleeding-edge-package

          Posted by Matthew Miller | August 30, 2010, 2:12 pm
          • I thought installing a package from rawhide was inadvisable because it could bring in many rawhide dependencies turning your system into a hybrid of stable and rawhide.

            Wouldn’t a good solution be to make greater use of the testing repository?

            Those, such as myself, wanting the “latest stable software” would enable the testing repository. I am willing to file bugs and downgrade to previous version if necessary.

            My friends and family, using the default install, would not have the testing repo enabled. They are not the type to be filing bug reports or missing the latest point release of inkscape. Still, I’d like them to get the “latest stable software” after it survives being in testing for two weeks without any issues.

            Posted by Allen Halsey | August 30, 2010, 4:57 pm
  10. Hi
    I agree with the issues identifiedn in your blog Mairin and if threre’s something I can add i’d propose to use the features Fedora already have.

    - IIRC the target audience was already defined somwhere around launch F13, it had to be some more tech-savvy users, maybe not devs but peapole who doesn’t mind read few lines to learn how to solve their issue or spare a few minutes to fil bug report.

    - regarding new features, I believe Fedora needs something like technology goals, things community wants achieve in next few releases. So each new release brings Fedora closer to the desired state. Of course some random feature here and there won’t hurt, but we should identify areas where Fedora and Linux in general is lacking.

    - as far as updates go, I’d propose to advertise testing repository and Fedora n-1 release more. Right now the fast pace development is something I pariculary like about Fedora. As it sometimes might breat things but also can bring some fixes faster too. To catch serious regressions testing repos should be advertised to the volunteers more. It’s kinda like light involvement, where you duty is just to let know if the new update breaks something for you. Especially if there was a plavce to post comments about updates other than bugzilla. If there’s any critical regression people will start to shout and it’ll be better if those are testers who accepts the risk and can know how to revert update than regular users. You know there’s no such things as too many testers ;)
    Also I think, n-1 release is a way for those who want stable Fedora. There’s nothing wrong with not using all thelatest and greatest toys. I believe in 3 months after release all the major issues should be known and team should decide if it’s fixable within this particular version of sotware, if there’s a fix that could be backported frm next release or some commits needs to be reverted so when next release will be made, this one can be safely put in kinda maintenance mode with most of the identified major issues solved this or other way.
    Actually, looking at state of past Feodra releases n-1 release is exactly what other distros ship so users won’t get anything worse that available anywhere else.

    Am sorry for quite long “comment”, but these things affects me as Fedora user in very direct way, so I felt it’s important to provide some opinion on that matter
    Take care
    /Stan

    Posted by Stan | August 29, 2010, 5:28 pm
  11. I for one appreciated a lot more the times when I was able for example to find and install a package for the newly released Inkscape 0.48 on the stable F13 and not having to jump to F14 pre-Alpha. Those were the times when we had “features” and “first”.

    Posted by nicu | August 30, 2010, 2:17 am
    • Um, really? I spent some time looking at doing that myself last Friday and could only find F14+ packages….

      Posted by mairin | August 30, 2010, 8:47 am
      • Yes, we have in Koji only F14+ packages, in older times we used to have them also for the current releases, in updates. Now we have to wait until the next version, so we are not the first to get the feature, as we used to be.

        Posted by nicu | August 30, 2010, 8:57 am
  12. Thanks Mo. I think it’s all fixable, but I also think I might stand for election sometime with the aim of addressing some of these, especially lack of cohesion. btw, my blog is down due to the server being moved this afternoon.

    Posted by Jon Masters | August 30, 2010, 1:52 pm
  13. Good to see that someone who is active in the project is becoming dissatisfied with the result. As I have noted on the lists (too often) in the name of progress regressions are introduced and treated dismissively. My pet peeve is video, all of the machines which ran FC4..9 fine are now using VESA mode or vendor drivers. Go get them, and next time you are at RPI I’ll try to drop by, I’m neither a student nor alum, but I am a hockey and Linux fan.

    Posted by Bill Davidsen | September 5, 2010, 5:47 pm
  14. IMO Fedora is a very closed project in the sense that decisions are made by a small (relatively) group of individuals. Feedback is generally not welcome, or ignored.

    The reason I use it is because the technology choices are the same I would make if I was building my own distro, and because it’s stable.

    However, it’s clear that all the exciting stuff happens on rawhide, then there’s a period of many updates when the release happens, and then the release is forgotten.

    This makes chasing bugs very difficult, and proof of that is the 27000 open bugs, many of which will be ignored, and automatically closed when a deadline happens.

    Clearly, Fedora is still too tied to RedHat and Fedora’s bugzilla only serves as a thermometer to get a feeling of the stability of certain components.

    If the Fedora leadership actually cared about their user-base, they would setup an ideastorm to see what people actually care about.

    Posted by FelipeC | September 5, 2010, 9:09 pm
    • Wow, you haven’t spent very much time in Fedora’s community have you?

      Try showing up at the public board meetings for Fedora, and bring up your concern there. Yes, we have public board meetings. They take place in IRC on a bi-weekly basis. I look forward to seeing you there.

      Posted by mairin | September 5, 2010, 10:15 pm
      • I have been involved since Fedora 1.

        My bugs reports gather dust, my feedback is ignored, even my package in need for review has been untouched for 8 months even after I addressed all the hostile comments and I’ve repeatedly asked for comments on the bug, and on #fedora-devel.

        Also, each time I ask something in #fedora I’m welcomed by an array of questions directed to undermine my intellect and suggest that I shouldn’t do what I need, and not to address it.

        Anyway, you didn’t address my proposal of an ideastorm; I can already imagine all the Fedora “leadership” being against it; “Why have ideas nobody is going to implement?”, well, exactly.

        Posted by FelipeC | September 6, 2010, 4:01 am
    • Really? I’ve been involved with Fedora for a long time, and I really can’t find any agreement with a single one of your statements. I don’t mean that there are no problems with Fedora (no project can be perfect, especially not one this large), but often, the problem is the *opposite* of what you’re saying!

      Posted by Matthew Miller | September 5, 2010, 10:42 pm

Trackbacks/Pingbacks

  1. Pingback: Beyond dogfood « Alex's Blog - August 29, 2010

Follow

Get every new post delivered to your Inbox.

Join 63 other followers

%d bloggers like this: