All the faces of Ubuntu

Thursday, March 7th, 2013

Harald,

Of course what Kubuntu and Xubuntu and Ubuntu GNOME Remix et al do matters. If it didn’t, we wouldn’t invest a ton of time and energy in finding ways to share the archives effectively. And I consider it one of the lovely things about Ubuntu that there is room for all of us here. As long as there are people willing to make it happen, there’s room for a new face.

You all make the broad Ubuntu family more diverse and more interesting. For which I’m grateful.

In return, you get the benefit of an enormous and concentrated investment in making a core platform that can be widely consumed (on top of the already enormous efforts of the open source community, Debian, and any number of other groups). That investment brings with it a pace of change, and a willingness to be focused on specific outcomes. Mir, which is a fantastic piece of engineering by a very talented team that has looked hard at the problem and is motivated to do something that will work well, is just one example. Every week, we’re figuring out how to coordinate changes. Why blow a gasket over this one? I’ve absolutely no doubt that Kwin will work just fine on top of Mir. And I’m pretty confident Mir will be on a lot more devices than Wayland. Which would be good for KDE and Kubuntu and Plasma Active.

So, before you storm off, have a cup of tea and think about the gives and gets of our relationship. Seriously.

Mark

38 comments:

  1. luzon says: (permalink)
    March 7th, 2013 at 6:36 pm

    Your blog posting about divisive leadership reminded me of your OpenSUSE “invitation”.

  2. ethana2 says: (permalink)
    March 7th, 2013 at 6:50 pm

    Keep in mind how much help Google was in porting Kwin to Android and ChromeOS…. ;p

  3. e8hffff says: (permalink)
    March 7th, 2013 at 6:54 pm

    Looking forward to the coming great works on Ubuntu.

    Mir and its ability to layer on Android and talk to X11 is amazing and very smart. You are addressing the current means to display and the future.

    Thumbs up to the developer who come up with the idea and the discussion group that must have hashed out the plans.

  4. e8hffff says: (permalink)
    March 7th, 2013 at 7:00 pm

    @ethana2 is doesn’t matter as Google needs to comply with the licensing for Linux, that is the backbone of Android. Canonical’s team and others are playing on the support Android gets to lever Ubuntu into that greater device world. It’s simply a brilliant move that doesn’t need to be called a check-mate. All can win with ‘Mir’.

  5. Mark B. says: (permalink)
    March 7th, 2013 at 7:08 pm

    One concern surrounding Mir is that toolkit providers will not be willing to invest supporting in three competing graphics stacks and that for the Cannonical teams, Unity will be the focus without regard to other DEs/toolkits. If Mir is technologically superior to Wayland/X11 I hope it wins. However, the history of Ubuntu sponsored projects (for example Upstart) being adopted by other distros is spotty and tech history is rife with examples of superior solutions being crushed under the weight of cheaper/easier solutions.

    Another concern regards the ability of Cannonical to push Mir support for toolkits upstream. If Ubuntu maintains the patches in its own repos, Ubuntu will be constantly behind other distros who ship graphics stacks supported by upstream.

  6. Daniel Rose says: (permalink)
    March 7th, 2013 at 7:28 pm

    The problems are obvious:
    1) You are solving a problem YOU have, namely that the Desktop is less and less relevant. This is not a problem your users have.
    2) I don’t think you are strong enough to force Canonical into the tablet/phone business at this point. I also don’t think you are strong enough to unify the GNU/Linux community around Mir.
    3) What you can and will do is increase fragmentation. The countless audio layers are an immense problem for Linux. We don’t need this kind of diversity on the display side.
    4) What this amounts to: The tablet and phone versions of Ubuntu will die a quick death due to a lack of adoption. The desktop version will continue as a largely incompatible distribution that only increases the problems developers already have. And why? Certainly not because of your users. But without them, Ubuntu would be nothing.

    Thank you very much!

  7. fireboot says: (permalink)
    March 7th, 2013 at 7:32 pm

    Hi,

    I just read the wiki page about Mir.

    While I’m far from being a graphic expert, I still don’t see why Wayland cannot be improved to meet all of your requirements, and I therefore don’t see why Canonical chose to develop its own solution if not to have full control over the project.

    Any clarification would be very much appreciated

  8. Jonas says: (permalink)
    March 7th, 2013 at 7:40 pm

    Hi Mark, (user since Warty here)

    I think a lot of the criticism against Mir stems from the initial miscommunication in , where it was stated that Mir was created due to certain, incorrect, perceived faults with Wayland; I think it would do great service to the community if you would simply state clearly and publicly that you just wanted to have the freedom that rolling your own solution gives you to do whatever suits your needs, which seems to be the case — rather than justifying Mir’s existence in terms of criticism of Wayland, which is entirely unnecessary.

  9. Fitoschido says: (permalink)
    March 7th, 2013 at 7:56 pm

    I just don’t understand why Canonical is getting so much FUD.

  10. Walter Alejandro Iglesias says: (permalink)
    March 7th, 2013 at 8:00 pm

    My general thinking to all FOSS users

    Coherent with his “GNU defends users freedom” argument, Stallman should be
    against all software that try to keep users ignorant selling the fake “out of
    box” experience. He should be against KDE, Gnome, Xfce just for that.

    But we all know that final users are not programmers, they have no interest in
    reading any code. They just want things work “out of the box”.

    So I wouldn’t argue against Mark Shuttleworth bet. It’s fine and for the sake
    of all that *one* Linux distribution like Ubuntu exists. I am not either
    against the development, maintenance of KDE, Gnome, Xfce and I am not against
    that even a distro like Slackware includes all of them (falling in a clear
    contradiction with its supposed KISS philosophy). I am not against the
    existence of MacOSX or Windows. Because I am not against *everybody* be happy
    doing anything he wants. Because we all are different, each individual is
    unique; that’s not a human choice like Mark think, that’s a nature choice.

    BUT, the true is that the whole concept of *freedom* is indeed relative; the
    same can be said about “free” in the “free beer” sense. All has a price in
    this world and we all affect each other. So, what you think is *your freedom*
    or *your right* could imply hard work or even slavery for a lot of other
    people. In practice everybody is forced to renounce to some of A to get some
    of B and vice versa. In all cases, at all levels.

    The problem fall on the real beneficiaries of FOSS, the *advanced users*, I
    mean programmers, system admins, etc., and all that found on Linux not just an
    urban religion pretending to be an ideology but a real choice. The problem is
    that in practice all that “out of box”, “use and drop”, “just do it” today’s
    religion ends up *taking the place of* the true useful, reliable, simple,
    coherent, proved stable use of the computer. Not just in part; B *absolutely*
    replaces A to the point of making A disappear.

    I feel like the last monkey climbed in the last tree in front of the man with
    the mechanical saw. And it’s not about a snobbish “I want to be different”,
    like Mark think, it’s not about “Linux being hard and exclusive”. It’s about
    Keeping it Simple, Shuttleworth. The Out of Box experience imply a lot of
    layers of software abstraction. All that software must to be developed and
    maintained by the software community. The more dumb his idealized world of
    clones that he calls *people* becomes the more difficult, complicated is the
    job for the software community. Sit Mark to write the code himself and you
    will see him changing of mind quickly.

    That’s why I complain about some popular figures promoting the “out of the box”
    experience, not attacking but *defending* myself. Do not *force* me to follow
    the cows and I will not complain:

    http://comments.gmane.org/gmane.comp.fonts.fontconfig/4462

    I just advising those that think to be clever betting for the pure marketing
    strategy that if all linux distributions follow the Ubuntu way it will not be
    any *reason* to choose FOSS for *nobody*. If you are fine keeping it alive
    like a urban religion, congratulations, I’m not.

    Walter

  11. Antony says: (permalink)
    March 7th, 2013 at 8:49 pm

    I just don’t understand why you don’t want to work with upstream more. Rather than create Mir, you should have worked on improving Wayland. I’m certain the Wayland developers would have been happy to receive patches to help Wayland meet Ubuntu’s needs.
    ATI struggles to support just X.org alone. Do you really expect Nvidia and ATI to support X.org, Wayland AND Mir? It’s simply ridiculous for them to do so based on the overall Linux marketshare.
    You’ve created a mess. And don’t get me started on you not wanting to adopt systemd.
    No wonder Icaza just dumped Linux for a Mac.

  12. Bart Willemsen says: (permalink)
    March 7th, 2013 at 8:58 pm

    I wish you all the best in the world and I hope that developers will start to support your decisions. Linux needs consistency, and if the community can’t figure it out by themselves than the biggest party must stand up and do it instead.

    I really hope that things will work out!

  13. kersurk says: (permalink)
    March 7th, 2013 at 9:05 pm

    @Fitoschido
    It seems that people don’t approve Canonical’s reasons to invest time and money into a new display server, as it seems that Wayland is just as good, and should be improved instead.

    Community maybe felt, that Wayland is something truely community-made (though it did start as one-man project afaik), but Mir feels more closed, as it was kept as a secret.

    Maybe it would have made sense to be open about Mir from the start, but I can see the Canonical’s point here, which is that they probably didn’t want to announce anything until having a working prototype.

  14. honko says: (permalink)
    March 7th, 2013 at 10:10 pm

    @Bart Willemsen: I don’t quite follow you there. How does adding yet another display server for Linux lead to consistency? It only means that now toolkits, applications and drivers have to support two display servers (on top of X) on GNU/Linux. The fact that Mir is L/GPLv3 licensed, under CLA and was completely developed behind closed doors pretty much assures that it won’t be adopted outside of Ubuntu. All of this isn’t really a big deal as long as propietary driver interfaces are usable by both; altough unnecessary fragmentation because lack of communication is still unfortunate.

  15. keithzg says: (permalink)
    March 7th, 2013 at 10:24 pm

    I’ve absolutely no doubt that Kwin will work just fine on top of Mir.

    As someone who loves the Ubuntu base but loves KDE above all else, I know that I, and many others, will hold you to that Mark.

  16. Meow says: (permalink)
    March 7th, 2013 at 10:33 pm

    Please consider the fact that the difference that have made ubuntu popular exist more than anything in the packaging, management of releases, and brand recognition.

    Many of your internally-sourced projects, such as bazaar, upstart, unity, etc, are not terribly popular with the community. Bazaar, as a good example, is nearly extinguished. Upstart has been superseded in many ways by systemd. I’m not trying to say that your developers are bad, but that there are thousands of skilled developers all over the world working on similar projects. You don’t always have to start from scratch.

    I’m not saying you have to do what everyone else does, but please don’t take your technical ball and go home. You’ve received far more from the community than they have received from you, speaking from a total code and asset percentage. So, if in the future, you would be willing to talk to the other guys before deciding on a system or an interface and throwing your name around, it would be appreciated.

    You have some weight now. Please don’t use it to make everyone else’s lives miserable. As an example, since you’re developing a display server, you could discuss with the Wayland and Android development teams the driver structure before you arbitrarily decide on an interface. That way, everyone’s code can share the same drivers with minimal difficulty. Or you could communicate with the alsa, audioflinger and pulseaudio development teams about sound needs before deciding to implement a new audio system. Also, Steam directly supports ubuntu now, so if you break sh*t and rebuild existing things in an entirely new way, you might force people’s hands even if their current code and roadmap is technically superior than yours.

    We’re not asking for control. Just don’t be such a secretive clan of holier-than-thous and come talk with the rest of us. Maybe we’ll do some of the work for you. There are a hell of a lot of talented developers working in the kernel, and in wayland, and in many other places. But at least we’ll understand why you’re doing what you’re doing and you won’t have to post passive-aggressive PR statements like this.

    Please put that in your cup of tea and think about why people are upset. Fragmentation on Linux is already a problem, and you’re complicating it more, whether it will be a good idea in the end or not. Talk before, so people understand the need for a fork instead of doing your own thing and then having to publicly mince about it later with arguable success.

  17. manny says: (permalink)
    March 7th, 2013 at 10:48 pm

    I believe most (if not all) the decisions you guys are taking are for the best.

    From the 3 month UDS, to the focus on the LTS point releases that make it so much easier for users, OEMs and ISVs thanks to the backporting of software, fixes and features …

    To u-Touch and now Mir.

    The latter of course being the most controversial since many assume they will need to modify their plans.

    But am sure it was a very difficult decision. Is probably one of your biggest investments so far and You more than anyone probably thought that we would be in the wayland wagon by now, but then you had to let go of the idea at this point (maybe the last minute) because the situation always keeps changing…

    But you guys are not afraid of adapting (even if you need to drop your new plans and revert back to what you had. i.e. unity-desktop again in QT).

    Of course none of the ubuntu devs just woke up one day with the intention of “pissing” on anyone’s plans, in fact it could be a better alternative for many. But yes, like some others have mentioned, it would be a better idea to communicate any indecisiveness or a change of heart a little earlier.

    Anyway, am glad to hear Mir will be available for other desktops. Because just like u-Touch is optimized for mobile devices and unity for desktops, the other DEs are optimized for other usage scenarios or type of users. It all makes for a healthy and diverse Ubuntu ecosystem.

    More than ever we’ll need to stay united to make this great vision of convergence and open source a reality! Good luck and success to all your team and contributors! :)

  18. Aaron Seigo says: (permalink)
    March 7th, 2013 at 11:15 pm

    “I’ve absolutely no doubt that Kwin will work just fine on top of Mir.”

    That is highly unlikely.

    http://blog.martin-graesslin.com/blog/2013/03/war-is-peace/

    Unless there is a port of libwayland-* APIs to Mir’s protocol (assuming KWin stays within those boundaries) allowing KWin to run untouched on Mir, I don’t see this happening. The rest of the Plasma workspace processes face the same issues. This is a very bold claim to make, and I’d love to see the detailed technical explanation to back it up.

    “Which would be good for KDE and Kubuntu and Plasma Active.”

    The only way this could be accurate is if getting Mir on more devices made it easier to get Wayland on those devices.

  19. Dan Williams says: (permalink)
    March 7th, 2013 at 11:24 pm

    Given how the maintainer of Kwin has basically refused to support Mir (based off his prediction that only Ubuntu will ever make serious use of it), why do you have no doubt that Kwin will “work just fine on top of” Mir? Do you expect that the Kubuntu community and/or Blue Systems will create and maintain their own backend for Kwin?

  20. Simon Strandman says: (permalink)
    March 7th, 2013 at 11:27 pm

    “And I’m pretty confident Mir will be on a lot more devices than Wayland”

    I’m not so sure considering the tizen project, backed by Intel and Samsung among other, will be using Wayland. Benefiting from that would have been great for canonical. :/

  21. Aaron Seigo says: (permalink)
    March 7th, 2013 at 11:32 pm

    @manny: “The latter of course being the most controversial since many assume they will need to modify their plans.”

    Outside of Ubuntu, I don’t know of anyone modifying their plans at this point. Every single developer I’ve spoken to thus far is completely fine with the plans they had a few weeks ago. Some have no plans beyond x.org at this point, and those who do are continuing to move towards Wayland. This has been a couple of years in the making now. We finally have Wayland at 1.0 as for October last year and it is a competent system and we have QtWayland that works very nicely. Mir really isn’t of interest.

    So that means our software won’t run on Mir. Since our software is not shipped with Ubuntu anyways, that’s a non-issue. As long as x.org or wayland packages are available (via Debian is nothing else), Ubuntu derivatives will be just fine as well, so that too is covered.

    As long as Canonical shoulders the burden of maintaining Qt, Gtk+, XUL, vcl, etc. ports to Mir, then application developers (who don’t write to the display system, anyways) will also be fine.

    So, no, there is no worry about changing plans other than people who have been planning on using Ubuntu in their own products. That is a different story.

    “Is probably one of your biggest investments so far and You more than anyone probably thought that we would be in the wayland wagon by now, but then you had to let go of the idea at this point (maybe the last minute) because the situation always keeps changing”

    We all knew Wayland was headed for 1.0 for some months before the release .. which means that for a very significant % of time that Mir was in devel (perhaps nearly all of it), there was no real question here. As the Mir developers were apparently unaware of Wayland’s design (claiming, e.g., that the input method was the same as x.org demonstrates clearly they had not looked into Wayland much at all), I don’t see how this decision could have been influence at all by expectations about Wayland.

    Moreover, if Canonical had invested in Wayland over that same period of time, they would have been able to show Ubuntu Phone on Wayland at MWC rather easily.

    “Anyway, am glad to hear Mir will be available for other desktops.”

    At least for the KDE workspaces (desktop, netbook, tablet, media center), you should not expect that. (See my link to Martin Grasslin’s blog in my comment above).

    If Canonical wishes to invest in Mir, that’s obviously fine and I wish them luck. I’d just like to see it done without misinformation being spread about Wayland, KWin, Plasma, etc. in the process of trying to justify those decisions.

  22. will says: (permalink)
    March 8th, 2013 at 12:48 am

    While I appreciate the fact that you are addressing the concerns, what worries me is that this is going to fragment the community even more than it already is. What I forsee happening is we’re going to have 2 camps in the Linux community: Ubuntu and everyone else. So what’s going to happen is you’re going to have apps that run on Mir but don’t work on Wayland and vice versa. This is beyond Upstart vs SystemD. This is going to fracture user space and application compatibility. Keep in mind that Ubuntu still relies a lot on upstream development and app developers suporting the Ubuntu platform. If you really don’t think Wayland can support what you need why not contribute changes back upstream to it? That is how the open source community works. Ubuntu has a reputation already tarnished in the community due to poor handling of changes to the platform and perceived lack of contributions back to upstream. Further damaging this reputation is a very bad idea.

  23. celso says: (permalink)
    March 8th, 2013 at 1:39 am

    @ Mark Shuttleworth :

    When you decide to make the unity interface, i was a little ceptic but i decided to give it a try. And here i am since ubuntu 6.10 to 12.04. But seriously, making MIR will benefit more ubuntu than wayland? I mean, the cost of human work, the schedules, drivers (proprietary and open), the companies that will support it, will bring more benefits for ubuntu? Ĩ don’t care if ubuntu uses mir as long it benefits ubuntu AND the community BUT isn’t this a “waste” of human work since we have wayland? And wayland have more people working on it!
    ( this doesn’t mean that i dont trust on the ubuntu devs! I only have to say thanks to them but it seems that ubuntu is wasting resources in this case)

    Best regards!

  24. Israel says: (permalink)
    March 8th, 2013 at 2:42 am

    Something people should remember is that this is FREE software as in FREEDOM. If you want to run Wayland on your system, you can. If you want to compile your very own Kernel you can. Ubuntu does not limit this freedom in the slightest. In fact Raring ringtail has even MORE Desktop Environments than before (Weston and Wayland are in the repos as well)

  25. Ben says: (permalink)
    March 8th, 2013 at 2:43 am

    What happened here? I thought Wayland was the future for Unity:

    http://www.markshuttleworth.com/archives/551

  26. F i L says: (permalink)
    March 8th, 2013 at 6:41 am

    I just want to add that, in light of the new Phoronix article about the possibility of Nvidia creating a Display Server agnostic EGL driver, this situation wouldn’t be nearly as bad. The important thing is that commercial games, and applications like Blender which work much better on, or require the proprietary drivers are not restricted to Mir-based systems (when Wayland is so far ahead), but instead could simply target EGL for creating OpenGL graphic contexts and then only need to support a thin difference between Mir-based systems and Wayland-based ones. I still think the divide is pointless and annoying, but if that is indeed the situation, then things are not near as bad as my previous two posts make out.

  27. Reply to “All the faces of Ubuntu” | Martin's Blog says: (permalink)
    March 8th, 2013 at 7:19 am

    [...] you “have absolutely no doubt that Kwin will work just fine on top of Mir”. This is great and I totally appreciate that you think Mir is a great system. But I’m [...]

  28. Karmo says: (permalink)
    March 8th, 2013 at 8:26 am

    I’d like to see explicit examples why Mir is better than Wayland. So far I have read only articles full of unsupported statements. Including the one above.

  29. Jasna Benčić says: (permalink)
    March 8th, 2013 at 9:26 am

    The more I read your blog calmer I get. I’ve got to stop reading “articles from around the web portals related to FLOSS.” Especially when something like this occurs – the vawe of “where Canonical is heading to.” Those kinds of aritcles can be really influencing…

    I’ll be back with contributions but not yet.

    Stay well,

    J.

  30. Kai-Uwe Behrmann says: (permalink)
    March 8th, 2013 at 9:27 am

    Mir looks like a project to get Ubuntu to Android fast. That is a reasonable and nice thing to have. But will it outwight the reduced community support, as this controverse here suggests?

    Given that Canonical had quite some man power behind MIR, I imagine what could be achieved with that behind Wayland in the input event handling, Android porting and driver parts. Perhaps a according time line could be even shorter than with MIR?

  31. Jack McCaw (@JacketMcCaw) says: (permalink)
    March 8th, 2013 at 10:40 am

    What do I want? Mark, I want a secure, solid operating system, that allows me to run EVERYTHING I want, for low cost.. no, make that NO financial outlay at all!No effort on my part, no fiddling or ‘compiling’ or whatever the hell that you have to do. Real time kernel, blistering fast response times, in one packaged distro. And I demand it no.. oh – it’s already here. Hi there UbuntuStudio… Um.. never mind, carry on.

  32. Pascal says: (permalink)
    March 8th, 2013 at 11:27 am

    As somebody who does not even use Unity on a daily basis, but is a great fan of the KDE Plasma desktop, I simply admire the vision and direction Canonical is taking. And what efforts everybody in the community and in the Canonical teams are making.

    Leadership can never be popular with all people, because even if they share many interests, they differ in their opinions on other matters. And that is totally fine – free software is all about choice, and Unity along with its related projects is a great choice and a great chance to bring Linux to a wide range of devices. I think Ubuntu has already proven that controversial decisions can lead to great success at the end.

    So as a KDE fanboy, I wish you all the best, and do not get irritated by people who think only they knew what free software ideals mean.

  33. Марк Шаттлворт высказал сомнения в правильности перехода Ubuntu на rolling-модель обновления | AllUNIX.ru — Всероссийский портал о UNIX-система says: (permalink)
    March 8th, 2013 at 11:28 am

    [...] критику разработчиков KDE, Шаттлворт назвал недавно анонсированный дисплейный сервер Mir, [...]

  34. Fabio Rosa says: (permalink)
    March 8th, 2013 at 1:05 pm

    As long as wayland and mir can share the graphics drivers, I don´t see a problem with Mir.
    I think that Canonical could have handled this whole situation better. From what I see, it was a complete PR disaster with the FOSS developer community.

  35. Arian Fornaris says: (permalink)
    March 8th, 2013 at 2:41 pm

    I’m good if Canonical do MIR or Unity or any other thing if they will share as free software. They are in his fully rights of do whatever they consider better for his business. Something I love of the linux world is the vast diversity. There is more options out there, if you don’t like Ubuntu, install then other distro. However, I think all these comments should be taken in consideration by Canonical, specially technical comments.

  36. Celso says: (permalink)
    March 8th, 2013 at 2:42 pm

    @Israel:
    I am not saying that isn’t free. In fact i never doubt that since its one of the ubuntu promises: “Ubuntu is free and it always will be”. And i always trusted on Mark shuttleworth on his decisions so, that isn’t a problem for me. My question to him is if it will worth it? considering the costs of the people work, everything has to be made from the beginning (related to the window manager). I know, in the tek world, we need to rush on the innovations, because not doing it, we may loose against the others. I hope Mark gets all the support that he will need it.

    Best regards,

    Celso

  37. mark says: (permalink)
    March 10th, 2013 at 10:38 am

    @Ben

    We thought so too, until we observed a lot of the same dynamics in Wayland that we see in X. Mir is a small component, we are confident it will be fast and secure and reliable, and this whole kerfuffle is a non-event.

  38. mark says: (permalink)
    March 10th, 2013 at 10:41 am

    @Will

    Don’t be naive about the kerfuffle raised by Wayland developers. For a start, consider the competitive interests of the people who shouted ‘Fire’. They work for Tizen and Red Hat, two direct competitors. They are also people who have in the past been deeply antagonistic to Ubuntu and Canonical, *refusing* to land patches in their code needed to support Ubuntu, then arguing that we’re bad citizens for carrying those patches in the distro, while doing exactly the same things themselves.

    Apps will largely be ignorant of the distinction between display servers. OpenGL is the future, and Canonical is working closely with the Khronos Group, which supervises OpenGL. By contrast, Wayland is strongly lead by Intel, and while Intel employees (who are raising a stink about Mir) claim to have code which supports other approaches to graphics than the Intel one, they say they are not allowed to release that code.

    Methinks she doth protest too much.