Two Major KDE Developers Weigh In On Mir, Wayland


To the average user, the announcement of Mir is simply news of something that’s going to ‘happen’. To the seasoned user, Mir raises a lot of questions. Those questions have been asked and speculated upon back & forth across the Internet for days now, and finally we’ve got some real meat from two KDE developers; Aaron Seigo of Plasma, and Martin Graesslin of Kwin. These two projects are closely related and are major components of KDE.

I reached out to Aaron on Google+ for his thoughts about how Mir might affect a project like Kubuntu. I asked for his thoughts about Kubuntu specifically because it is the most closely related to Ubuntu. He offered some excellent speculation and valuable thoughts. Here was his response.

You’d best ask the Kubuntu devs .. but here are my expectations:

Kubuntu is developed separately from Ubuntu these days. Canonical does not pay for development there anymore, and it becomes quite clear why. :)

I would honestly expect Kubuntu and other *buntu derivatives to continue using and eventually move to Wayland. This is because they are more likely to work with their upstreams, who are either staying on X11 or working on Wayland migrations, rather than jump to the Mir Unity shell.

Where it will get uncomfortable is that these derivatives will likely end up having to maintain, Wayland, etc on their platforms and will no longer be able to rely on Canonical for this. This is likely to push the derivatives closer to each other and further from Canonical. Perhaps we’ll even see some simply rebase on Debian itself.

We’re still at least one year from that becoming a concern, however, and I honestly expect Ubuntu to keep around for far longer than that on the desktop for practical reasons.

This also assumes that Mir will work acceptably on the desktop at all; that too remains to be seen.

Definitely a fair and balanced response.  The one thing that stands out above all else is the idea that projects like Xubuntu, Lubuntu & others might be tasked with having to maintain Xorg, a duty currently provided by Ubuntu since they rely on it in the same way as everyone else.  Will these distributions be able to handle the task?  Will they join forces and share ownership of the duties?

I’d say that having to maintain Xorg would be a little ways in the future, only because when Mir debuts in Ubuntu 14.04, there will still be several applications that users will need that won’t be compatible with it, so Xorg will still be required.

And then today, Martin Graesslin of Kwin offered up his thoughts on Mir and the effects it will have in a very frustrated post.  His first point doesn’t really offer anything technical, but does shed the initial light on his overall annoyance with the topic.  He writes:

Does Mir affect us? Yes, obviously. Because of Mir I have to write this blog post, Wayland developers have to get the FUD out of the Mir documentation, it’s creating tension and it harms the development. We will have to face again and again the question whether Wayland is better or not. So yes it affects us and I’m not happy about it.

In his second point, he demonstrates some of the same sensibilities as Aaron Seigo simply stating that there is too much that remains to be seen to make any real judgments.  He then proceeds to explain the future of KWin and Mir and whether or not they will be compatible.

Will KWin support Mir? No! Mir is currently a one distribution only solution and any adjustments would be distro specific. We do not accept patches to support one downstream. If there are downstream specific patches they should be applied downstream. This means at the current time there is no way to add support and even if someone would implement support for KWin on Ubuntu I would veto the patches as we don’t accept distro-specific code. If Mir becomes available on more distributions one can consider the second question. Given the extreme success of Unity on non-Ubuntu distributions I’m positively optimistic that we will never have to do the evaluation of the second question.

What started as a hobby project in 2008 is now becoming something very real for Linux.  But when will Wayland be finished incubating?  And will Nvidia and ATI ever support it?  With Mir on the horizon, and Nvidia & ATI still providing questionable support for their hardware in Linux, they might choose which side of the fence they want to be on and leave the other behind.  But as Martin & Aaron have said many times already, much has yet to be seen.

Source | Google+, Martin’s Blog

About Dean Howell

Aside from being a huge Sega fan, Dean is an LPIC certified Linux professional with over a decade experience. In addition to spending his free time burning through the classics from Sega and evangelizing open source, he's also the editor-in-cheif of The Powerbase.
  • niagr

    I’m not really familiar with the internal workings of X but would it not be possible to develop graphic card drivers independent of the display server being used?

    • Aaron Seigo requires an “x driver” which is something pretty specific to X.

      systems like Wayland use the openGL stack for the GPU directly. the catch is that the openGL stack needs to expose the features that Wayland requires (e.g. sharable buffers), which not all of the binary blobs currently do.

      but with the Wayland/SurfaceFlinger/Mir style of approach, it is indeed possible to have one set of openGL drivers that works with all of these systems.

      ( would still be out in the cold there, though, requiring its own drivers on top of the openGL stack :)

      • niagr

        Sounds to me like not adopting one of the modern solutions would be detrimental from the code reuse point of view. Would most graphical applications relying on popular toolkits like Gtk and Qt be affected if these toolkits provide support for Mir/Wayland directly without changing their APIs? Is that even possible?

        • decora

          qt5 (qt4 is current) already has some wayland backend code, you can google for it.

      • sola

        Which practically means that if Canonical can force these OpenGL driver modifications out from the GPU vendors, that will benefit Wayland as well.

  • Jai Luthra

    Some links in this post aren’t appended with “http://” thus firefox thinks you are refering to a local (on powerbase’s host) file, and it opens up a link like –

    • Dean Howell

      Fixed the link to Martin’s blog. Sorry about that.

  • Dean Howell

    I just realized that the featured art for this article makes it look like Aaron is angry with Martin. Kind of funny looking, but not the desired effect… :)

  • Sergiu Bivol

    This article presents the opinions of two people that I have lots of respect for – they write code that I’ve been using daily for the last three years. While I agree that, at the moment, Mir is not even close to delivering what it promises to deliver, we should keep in mind that choice is a good thing for Linux.
    If Canonical succeeds and makes Mir a great display server with great driver support (both free and proprietary), other distributions will jump on the Mir train. It’s too early to tell. One thing is for sure – X has to go.

    • MarcusKlaasDeVries

      Choice is nice, but fragmentation definitely is not, especially considering Linux on the desktop is already such a small fraction of the market. There is a limited amount of development power. As it stands now, the same work is being done many times over. Not just on display servers, but also on window managers, music players and basically every other part of the Linux desktop other than the kernel itself. If people were to find compromise a little more often instead of fork whenever they run into conflict, I think this would be much better for the Linux ecosystem.

      • niagr

        Depends on how you define fragmentation really. While I do not claim to know exact details of the working, I believe both Mir and Wayland could relieve dependency of the graphics drivers on the display server, so the same code could be used by many display servers. Of couse I might be wrong (please correct me if I am), but if not, this would mean better hardware support while increasing choice. As for the old duplication of effort argument, this is the nature of open source. Forking freely is what makes FOSS what it is, as much as the ability to collaborate. Which project you want to support/use is of course up to you alone. As it is, lower quality forks often end up merging with the more successful competition or go out of development so it tends to resolve itself as the technology matures.

        • Nick G

          Ok, I’ll tell you, you’re wrong. Proprietary graphics drivers will always depend on the display server unless they switch to using the kernel’s DRM, DRI2 and KMS infrastructure, which is impossible because that would require them to be open source (KMS exports GPL symbols, which means only GPL code can link against it).

          Because the proprietary drivers implement their own graphics stack instead of using the kernel’s, they are inherently tied to a single display server. This becomes especially dangerous in the case of Wayland, where the proprietary drivers will likely have to pick and choose which compositors to support unless they come up with a way to get around that limitation.

          • niagr

            Thanks for clearing that up. I had been looking for such an explanation

  • Sasha Shepherd

    I love Ubuntu, and love the idea of the Ubuntu phone, but…if the Ubuntu phone forces me to use Unity, I won’t touch it. I’ll stick to Android. I get a really bad feeling about this whole mir thing, linux is fragmented enough as it is :(

    • Adrian Wechner

      please don’t get me wrong. i dont care what succeeds, Mir, Wayland or both… but really can’t understand your comment: “love the idea of the Ubuntu phone, but…if the Ubuntu phone forces me to use Unity, I won’t touch it” …. the Ubuntu Phone is all about Unity… without Unity you dont have the Ubuntu phone experience. that makes me wonder, I dont see you can love the idea of the Ubuntu (unity-)phone but not like Unity…

      • FFaael

        We are not interested in “the ubuntu phone experience” but the Linux stack on the phones should be nice.

  • Pingback: Links 6/3/2013: HP Pavilion Has GNU/Linux, Apache Exceeds 40 Million Downloads | Techrights()

  • Guest

    “vidia & ATI still providing questionable support for their hardware in Linux”

    Say what now? The binary blob from nvidia is actually light years ahead in support compared to any of the other video drivers for Linux. They’ve just released an update which has support for Titan

    • Grab Boyd

      this is the FUD and lies I have been talking about, spread by the very community we trusted. I have verified myself that the binary drivers from AMD are better performing AND more openGL compliant.

  • Silviu C.

    ” …and Nvidia & ATI still providing questionable support for their hardware in Linux”

    Say what now? The binary blob from nvidia is actually light years
    ahead in support compared to any of the other video drivers for Linux.
    They’ve just released an update which has support for Titan

  • Grab Boyd

    I also have unanswered questions for the status quo- why, after all these years, do we not have desktop xorg drivers that are openGL compliant ? AMD and nvidia have been more than willing to supply them and they have been doing so but the drivers break due to kernel (and xorg ABI) dependencies that change with even minor release version and break these drivers.

    A lot of FUD has also been hurled toward these vendors, with arbitrary ( and clearly false) propaganda that OSS drivers are much better, much faster and much stabler than the binary drivers- in my experience, that is not so. The OSS drivers have only caught up with the binary ones in stability recently; in feature, they are not even close. The only reason for binary drivers to fail is when the OSS people keep moving the goal post for these vendors instead of making it easy for them to develop for.

    In the past, KDE devs also bought into these lies and as a KDE user, I have often come across FUD such as KDE/plasma/kwin will work much better with OSS drivers for radeon – totally untrue, in my experience.

    The rabid, moronic attitude of OSS developers is holding linux adoption to very low numbers and raining on the satisfaction levels of those that brave the odds and run linux anyway. I see them trying to come up with ‘ok’ OSS drivers for the radeon h/w that is 10 years old, and then keep rewriting the infrastructure every 5 years, shoveling everything into the kernel meantime. While we are still wondering why we cannot have decent drivers for 10 year old h/w, new h/w, form factors and devices are appearing every 6 months. Why don’t they understand this and the fact that driver support is best left to vendors, esp when they are WILLING to do that ?

    I am with canonical on this one.

    • gert

      >The rabid, moronic attitude of OSS developers

      As opposed to the rabid, moronic tudes of commenters?

      Any time an asshat like you decides to call devs idiots I feel free to talk about their whore of a mother who pulled a train of NBA players.

      You might think its an over the top reaction but calling this shit stain out IS the point. Whether its 1,2 or 3 insults doesnt make it different.

      Any value this anal wart might have had in this post is lost because he’s an ass.

      Your not Linus, so learn to fucking talk to others or STFU until you do.

      • Grab Boyd

        really ? seems your ass not only talks; it can type too. Now only if you could HEAR yourself…As for me, I cannot hear anything sensible coming out of there. And wasn’t it enough that all you had to spout was insults at me ? Why drag other members of my family ?

        • Arthur Dent

          You were called out for insulting devs and did nothing to prove what you are: a person in love with himself as only purveyor of truth.

          Calling a group of developers rabid and moronic shows the level of discourse you are capable of.

          I dont believe the way to debate is to attack someone since it ends up being nothing but a waste of time.
          I do think however that the majority of anonymoud commenters would never dare call an open source developer a moron to his face.
          That is the courage of anonimity.

          Each one of those devs has the courage to put his name next to his work. Because of that they will be bigger men-women than you can ever hope to be.

          • No

            Not one of vrm’s points was addressed. It is very well and all to point out that ad hominems do not support an argument, but focusing on them is non-sequitur and only an attempt to derail the discussion.

            What is more, I can find plenty of times where the linux community would willingly insult Microsoft and/or its developers. By placing special emphasis on open source developers you’re turning yourself into some kind of nationalist; which shouldn’t be viewed as a good thing.

            It’s not even Microsoft really, on software that clones the functionality of existing proprietary software, even the devs will resort to petty insults over the proprietary version while expecting full respect for their own. I particularly love octave’s “do_braindead_shortcircuit_evaluation” shows REAL respect for Matlab’s developers huh?

            If I want to call a developer an idiot, If I want to insult said developer’s decisions, I reserve the right to do so. But being a fascist and fervently protecting them SOLELY because they produce open source software, while spitting on those who do not, is NOT acceptable. Developers are human and DO make stupid decisions, some don’t even know what they’re doing and just test if it compiles [Well, funny opensql bug I perused over, removing a line of code without knowing what it is doing is not intelligent].

            But whatever… I’m just bored and talking to no one.

  • minnesota linux

    What’s a Mir? No where is this discussed in the article or why we should be concerned with Mir and KDE.

    > specifically because it is the most closely related to Ubuntu.
    Who cares? Ignore Ubuntu, use Debian Testing + KDE, done.

  • MichaelADeBose

    Technical merits of Mir or Wayland aside, the real question is, in the time frames Ubuntu(Canonical) has identified, could the changes required for Wayland to be a viable solution across all platforms identified by Canonical, be implemented? The followup to that is then, could it be implemented in the way that we go about new directions in the opensource community?

    I’m not even sure Mir is up to the challenge, but I do know that the kind of decision making that is paramount in commercial ventures isn’t likely to happen as quickly as its needed to, in the opensource community. Whether Wayland was being developed in as wide a scope as Canonical is looking at I don’t know. What appears clear to me is that no one was looking nearly as far down the road as Canonical except KDE Active, whose time frames appear to be ‘it will happen when it happens’. Opensource is a constellation of different takes on various ideas and we’ve even done that with the licenses. Ironically we seem to think there’s only one way to discuss change and there’s only one time frame, dare I say glacial; which allows a few rounds on the mailing list and a couple circuits + talks at various *ni(u)x events before we can feel good about it. That is a strength of opensource, but not when going commercial.

    Apple and MS have been moving towards a ‘unified experience’ and Ubuntu + Android is a decent implementation towards that end but now Canonical decides to go all in. We’ll have to work on their communication, but seeing how the iPhone got Consumers and Enterprise customers to rethink Apple kit, Canonical’s moves may if for once benefit opensource as a whole. This needed to be done and it needed to be done NOW, which is not quite our way of doing things and I think that might be the main issue.

  • Pingback: Two Major KDE Developers Weigh In On Mir, Wayland | Linux A.I()

  • Pingback: Linux Mint "Not in the Business of Picking Winners", Continues With Xorg()

  • Pingback: What’s Good For Canonical Is Best For Ubuntu « FOSS Force()

  • Pingback: Plasma Workspaces 2 Coming To Wayland, KDM Not Invited()

  • James LaBarre

    WHAT ABOUT NETWORK TRANSPARENCY IN WAYLAND???? The continuing refusal of the Wayland developers to decide to make network transparency enabled **by default**, or even the willingness to discuss the topic *at all*, makes me suspect that we will be screwed in that respect, and our machines will be forced into walled gardens, with the only way of interacting with them being some completely useless “full-desktop” solution like VNC or RDP. If I wanted to be stuck with that limitation, I could just run MSWindows.

  • Pingback: Python 3 and Wayland()