You Got A Question? Ask    GNOME Community!

Working on a bigger proposal

This post was made with an older stylesheet

By Matthias

Just to be clear: I do not speak for the GNOME foundation, and what you are citing here are internal discussions of the release team.

Few minutes ago in GNOME’s Mailing Lists by Matthias.

Hi, just to give you a heads-up: I want to propose that we should make GNOME work on Wayland.

We (me, Owen, Jasper and some others) met with krh last week and talked about this idea at some length. I’ve put the resulting information on the wiki here:

So far, we’ve silently assumed that Wayland is the future display system on Linux, and that we will get to using it eventually. This hackfest proposal from last year shows that we’ve been eyeing this direction for a while: (even if it didn’t happen at that time). The recent Mir announcement makes it a bit more urgent that we put our weight behind Wayland and help it reach its full potential. Doing this also fits our mandate from last Guadec to set stronger technical directions.

“Gnome on Wayland” is a bit too big for getting it done in one cycle, which is why I want to propose that we aim for doing it in two cycles – by the fall, have gnome-shell optionally work as a Wayland compositor, and have the gtk backend substantially complete – by next spring, have a complete port of all desktop functionality. Making applications run natively can happen independently at its own speed – we will support X clients transparently.

Let me know what you think

Of course you also need to read the upcoming Follow-Ups.

About Mir

Just a quick notice. Mir isn’t like a fragmentation between a package manager or a toolkit. Mir brings fragmentation in the very low level of Linux Desktop and in the part that Linux hurts most; Graphics stack, including  Mesa, Drivers and Hardware Manufacturers.

GNOME (Shell) will not be able to run under Mir as Shell will turn into a Wayland Compositor.  None (as far as I have read) in FOSS except of Canonical engineers are on the side of Mir.


  We can't watch comments unless G+ provides an API or if you send a notification, e.g +World Of Gnome
     Sometimes is better to place your questions on GNOME Community
  • Rajesh Ksv

    Last line – shouldnt it be ? – “None (as far as I have read) in FOSS except of Canonical engineers are on the side of Mir.” ?

    • alex285

      yeap, thanks!

  • Ace

    I wonder what will happen with apps like Inkscape or Gimp that still use Gtk 2. Will Gtk 2 also be ported to wayland?

    • alex285

      Check here ..but no gtk2 will not be ported in wayland. Not some other TKs

      • Ace

        Thank you.

      • Olav Vitters

        You can run X applications under Wayland. Makes more sense to port applications from Gtk+2.x to Gtk+3.x than to make such changes in a version which is supposed to only have bugfixes (2.x).

        • alex285

          Yes the link I gave was explaining how Wayland is compatible with X Server to run “legacy?” toolkits.

    • Pēteris Krišjānis

      Inkscape Gtk+3 port already happening (coming also with gui redesign, it’s a beauty), and GIMP already plans it’s departure to Gtk+3. Both ports can be done within 1 year time frame.

  • gutigen

    And what will happen with Ubuntu Gnome? :/

    • kazade

      Probably (hopefully?), Canonical will continue to ship Wayland in the repositories and Ubuntu GNOME will have that as a dependency.

      However, GNOME is and will continue to become more crippled on the Ubuntu platform because of the downstream patching that Canonical is doing. I’ve found Fedora to be a nice home for those of us who used to enjoy GNOME on Ubuntu :)

      • gutigen

        But I want Ubuntu as a base and Gnome as DE :P I guess I’ll just move to Unity if shit happens.

        • Luya Tshimbalanga

          Sadly Unity is a shell with toolkit identity crisis, first mutter, Qt, Compiz and now qt/QML.

        • Michael Heyns

          There’s no real advantage to sticking with Ubuntu anymore. Have a look at Arch/Manjaro, it’s still fresh but it has a big future. Also check out the whole ‘systemd’ hype regarding Debian/Ubuntu and regarding Gnome. (Manjaro is full systemd). :)

          • Rowan Lewis

            Manjaro doesn’t appear to provide a Gnome desktop, which is unfortunate.

          • Michael Heyns

            Not by default, but installing it is as easy as 1, 2, 3, reboot! I haven’t looked back since

          • alain

            I don’t know what you mean about Debian and systemd but I would like to remind you Debian works with linux and BSD, whereas systemd only is for linux. Debian does its best to integrate systemd (it enters slowly into the distribution) but people who created systemd and who are pushing it never cared to the troubles it could create for one of the main linux distribution. At the same time, it is pushed by Red Hat that is a Debian-competitor…

            I think that’s also a problem Ubuntu complained a lot. It is true many developpers are paid by Red Hat and so are forced to choose some technologies over another without thinking what it implies for the other distros. It is not only about FOSS!

            (by the way I run Debian)

          • Michael Heyns

            You are very right. Systemd isn’t BSD friendly and Debian is. I understand Debian is trying to integrate but it isn’t a fundamental changeover. I also agree with your statement that enterprises/companies push certain fundamental pieces onto the community. Just like Mir. I don’t want to argue over which init system is more superior nor am I interested in the politics and philophies surrounding the controversies.

            We are talking about Gnome. A huge project that has to base foundations on a certain model in order to survive in a constantly changing and fragmenting environment. One of these foundations is the systemd init system, which is becoming more and more a hard requirement. This has deep rippling concerns for the Debian, Gentoo, BSD and Solaris families.

            Personally, I love Gnome. I don’t mind what is under the hood, as long as it provides me with a smooth, efficient and consistent experience. In this case, systemd is becoming essential for that experience. Backwards compatibility is nice to have but adapting for the future is inevitable.

            (by the way I run whatever works the best for me)

          • Rowan Lewis

            Was systemd developed for nine months in the dark, then released by RedHat into the world? As far as I’m aware it was developed quite openly, and discussed frequently so the situation isn’t really comparable to this.

            The Debian issue is unfortunate without a doubt, however the Debian folkes have also done a fine job porting lots of other things to BDS, it isn’t completely unreasonable that systemd could follow one day.


          • alain

            I am saying the situation regarding systemd and Debian is completely different from what is happening with Mir. Michael said it in a way that seems like Ubuntu and Debian had the same behaviour about systemd. I pointed out Debian is doing an effort to port something not made thinking about its need (BSD and linux, not only linux). The problem with systemd is not it is developped in nine months or in nine years hidden or not but it has an impact on many distros because to integrate perfectly gnome, you must use systemd… You say the situation is unfortunate but it would have been better to first make systemd works with everything before making it compulsory for gnome.

            By the way, Mir development is not that hidden: install Ubuntu and use ppa:mir-team/staging in order to now how it works now and what it can do. There are certainly thousand of bugs…
            As I am not a gnome developper maybe I am wrong but there are things I don’t understand: to me it looks like a political choice, not a technical choice. Indeed this sentence taken from Mir specifications let me believe gnome developpers don’t need a lot of work to make gnome works with Mir: “Mir’s client library should be easy to integrate with existing toolkits. Application authors relying on Qt/QML, GTK3, XUL etc. should not be required to perform additional porting as we will work on providing Mir
            integration for the most prominent toolkit choices. In reality though, certain legacy applications will not be able to transition away from X completely, and we will provide an in-session rootless X server that is integrated with Mir. It acts as an on-demand compatibility layer between legacy X applications and the session-level Unity/Mir instance. ” When I read that I have the feeling gnome should work almost out-of-box.

            There are many things I don’t know but in democracy (let’s think the opensource world works like a democracy), the majority imposes what it wants to the rest or I should say the companies the biggest number of developpers impose what they want for them. If you are Canonical, a small contributor, maybe you would sometimes appreciate your ideas are followed and not rejected because we think you could pay more developpers to work for the community. I don’t know who is right or wrong here and I will never be able to know as I don’t know everything in this story.

        • hashem

          Why do you want Ubuntu as a base? You’re already using something based on Debian; just use Debian proper.

          • PromoteCommonSense

            If switching from Ubuntu, one should also check out Fedora and OpenSuse just to see what they offer, too.

        • PromoteCommonSense

          Wanting to use a base is meaningless. Having a
          list of wants and needs and then finding a distro that best fits them
          should be the goal. If you list what you want your linux system to do,
          you will find many alternatives to Ubuntu or any other distro. Then you
          can list the pros and cons of say the top three to come up with the one
          that best fits your needs.

    • JJ

      Also Jeremy (current UGR leader) is an ardent supporter of canonical. And he commented in another blog that he do not willingly lead the
      project and is waiting for someone else to take leadership. So I don’t think UGR has any future/hope anyways.

  • Marcos V.F.

    Linux is a bigger proposal. I full deslike to see the community divided. For the first time I see a community divided because a new OPEN SOURCE software that was created (Mir)…

    That the strongest survive. Since it is OPEN SOURCE. And honestly? Dont know who is the strongest.

    • Michael

      Being open source does not justify their actions, causing driver fragmentation for any distro that is not Ubuntu is a serious issue, using that argument would be the same as saying Android is open source so the fragmentation to the kernel and in house only development of its technology its bad news for any non Ubuntu distro if Mir gets more support than Wayland.

    • JJ

      If I create a virus and release it under open source license, will you support it (help me spread it)? Should the community support it because it is open source?

      • Sepehr Aryani

        PFFF, you don’t seem to understand FOSS, anything open source is appreciated even your imaginary virus.

  • foobar

    To be honest, I actually try not to be a Ubuntu/Canonical basher. But they make it a way too hard for me. They use software I wrote. And that’s perfectly fine. But they always make me feel like I should alter the licence a little without warning. Why? Because they make our live hard by being unpredictable. They make us waste our time. And that makes me angry.

    I specially love how the Space guy starts to spread lies on g+ about their competitors. This is simply not fair play. Back in the days, Gnome and KDE tried to live well together for the greater good. But they just don’t care about anyone else.

    The way they are doing Open Source is more a marketing strategy anyway.

    • gutigen

      Because they want to compete with Apple, M$ and Google. They can’t always do what Open Source community expects of them if they want to succeed. Unified vision under strong leadership is always better then democratic and fair collaboration, at least in big projects.

      Without Mark and Canonical Linux would be at 0.05% of desktops and things like Steam for Linux would not be possible.

      Most of Linux community is made off developers, admins… IT guys/girls. They make software for those like them, not for some random normal user which have no idea what partition or terminal or display server are.

      Canonical with it’s user centric aproach and very focused design and marketing is sole reason of Linux increasing popularity in last few years.

      • alex285

        @gutigen:disqus I don’t agree to that. Looking some graphs/statistics on internet you can see that the rate of Linux is increasing with stable rates before and after Canonical. Nothing change and Canonical didn’t seem to play any major part. Canonical just took Linux Users from other distros. That was at least what happened till mid of 2011 which I was gathering the data. Which can be wrong ofc!

        • JJ

          Well said. The ubuntu fanboys make it sound like canonical is some kind of saviour to the linux, but in reality they do more harm than good to the ecosystem.

          And Steam for linux happend because MS moved with their own game store and steam was afraid; nothing to do with canonical.

          • Ian Monroe

            This is off-topic, but I’m sure Steam for Linux is about the Steambox.

          • alex285

            hahaha good point, what display manager steambox will have? ;)

          • gutigen

            FOSS fanboys make it sound like Linux could get any non-IT (as hobby or profession) users without Canonical and Ubuntu :)

          • Craig

            Sorry, who’s the fanboy here? It certainly seems to be you, you pathetic cretin.

        • gutigen

          I would like to see those stats and how they were gathered.

          • alex285

            I don’t have the charts somewhere. You can Google for many of them and in every chart you will find about the Linux Desktop growth you will see a stable increase. Don’t care about the absolute values but only for the percentages.

      • foobar

        My point was that it doesn’t matter what they do, it is about how they do it. They are working with several guys from the community and simply block information from them. They make their schedule and do work to add value to ubuntu. And then they get screwed. Not because they start to use their own homebrewed software but because they promise to implement things and don’t tell them in the background when they abandon the idea. Some people even implement stuff for ubuntu after they got asked to do it right before they get screwed.

        They wouldn’t even need to induct external peoples about their plans. They just have to tell them — in private — that they are not going to do what they promised. They always give us the finger and wonder why people are pissed off.

        This is the issue. They simply act like assholes.

      • Michael Mistretta

        I use Ubuntu at home. I like it, it works, you could customize it the way you liked and it was great. When 11.04 came out and they switched to Unity I did not like it and continued using Gnome2. When 11,10 was released I had to pick between Unity and Gnome Shell, I chose Gnome Shell. When Amazon Ads came to the Dash in Unity, I knew I made the right decision.

        Now for quite some time I heard about Wayland and read Mark Shuttleworth’s praise of it and how he planned to adopt it for Ubuntu, now it turns out that Canonical contributed zilch to it and have decided to create Mir. While I do believe Ubuntu is very nice for it’s “just works” qualities I also feel at this point Canonical takes a LOT from the Linux community and gives back very little. It makes me wonder about the future of Ubuntu as a distro, part of me thinks I will have no choice but to switch eventually.

        Ubuntu is turning into something that just isn’t for me, I guess according to Mark I’m some elitist geek nerd loser who has no place in his operating system anyways. As long as I can run the desktop environment of my choice on Ubuntu I will continue to do so but I have to wonder how long before that isn’t an option.

        For the record, Gnome Shell discourages me as well, it seems like for every dubious feature we get we lose five useful ones. This is bad for Linux, I hope that things get back on the right track again. It seems like the Gnome developers are listening better than they once did. The “we know better” mantra needs to stop, give us back the features we need, Linux is supposed to be about creating an operating system that works for you, not our way or the highway. If I wanted that philosophy I’d use a Mac.

        /end rant

      • Craig

        Ok, you can take your tongue out of Shuttleworth butt crack now.

    • Dang Ren Bo

      > Back in the days, Gnome and KDE tried to live well together for the greater good. But they just don’t care about anyone else.
      “Back in the day” was obviously after 2002ish, if you believe that. The hatred between the two camps from 98-2001 or so was epic.

      • sramkrishna

        I wouldn’t say it was so vitrolic. No, what you saw was the fans of both desktops being vitrolic. We recognize smart and great coders from both sides. But man, slashdot for instance was a cess pool. The press also would love to bring up the whole KDE vs GNOME thing at every chance.

        Yet nobody seems to have an article on VI vs EMACS? WHY? Conspiracy?

  • Rocker

    Practically speaking, what are the advantages of Wayland over Mir, and vice versa? I know Wayland will provide flicker-free and tear-free graphics, but is that all? My Mint install doesn’t have flicker or tear that I can notice. Also, does anyone know what advantages Mir claims to provide in regards to what the end user will experience, and I’m not talking about what’s under the hood.

    • Michael

      Thats the big issue, there are no advantages to using Mir, Wayland can already do what Canonical wishes to achieve with Mir. All Mir is going to do is duplicate Waylands work, fragment Linux drivers, and give full control to Canonical vs a community controlled project. If canonical is successful it will slowly turn Linux into the same situation with Android, where its technically Linux under the hood but everything is developed in house and controlled, released, and licensed by a single company rather than a community project that can benefit every distro (which Wayland is).

    • PromoteCommonSense

      Wayland’s advantages are that it works with almost all distros and desktop environments. It is a lot closer to being deployable and it is fully tested. Mir’s advantage is that it will be optimized for Unity Next on mobile devices.

  • jon_downfromthetrees

    I’d hope that Gnome, and any other effort, would adopt the display server that’s best for its users, rather than obstinately stay with one or another because of some perceived violations of the unwritten rules of developer culture. To be blunt, users do not care about developer culture.

    The need to replace X has been obvious for years, yet the developer community has not appeared to be in any hurry to replace it. Wayland has been under development for several years and is still a long, long way from being ready. Why should users cut that effort a break if Mir delivers something better than X sooner?

    More to the point, 99 percent of users have no awareness that X, Wayland or Mir exist. They do not know what a thing called a “display server” is. They do not need to know. All they care about is that they expect Gnome or or KDE or Ubuntu or Fedora or whatever to work on their device. Why it works is irrelevant to them. The notion that users would adopt Gnome if it delivers second-rate performance due to misplaced loyalty to one effort’s developers is laughable.

    • Michael

      users who dont care about the developer culture probably are not running Linux in the first place they are either using Windows, Mac, or an iPad. So your point doesn’t really make sense to me. If Canonical decided to support wayland everyone could be working on a unified architecture instead of them causing even more fragmentation, dont blame the developers for supporting an open standard instead of supporting a single distro/companies desire for control which has far less shpped code than what wayland currently has. Canonical is hurting the users with their actions not the developers.

      • jon_downfromthetrees

        My point makes a lot of sense to me. Developer freedom is not always beneficial to users.

        I’ve run Linux since the early nineties. I see no reason why I should care about developers beyond their ability to produce original and innovative applications for me to use. The kind of developer culture whose members we’ve seen get angry at Canonical has nothing to do with the rules of FOSS. It’s just a culture that’s developed that wants to imposes its choices and values on Canonical. Canonical is a business and it needs to make a profit. Being a business and making a profit offend many Linux developers. Fine. None of Canonical’s business. But, that’s got nothing to do with being faithful to open source.

        I have no idea if Mir or Wayland will be any good. Neither are anywhere near completion. I want the freedom to choose the better product. I do not want that freedom arbitrarily taken away by, by Canonical or by the developers currently mad at Canonical.

        • Arron Washington

          “Neither are anywhere near completion.”

          No offense, but you should do a little Googling before making a statement like that. It very much weakens the positioning of your argument.

          • jon_downfromthetrees

            I know some early tests version of Wayland are out there. So, since I use Fedora, le’ts say Fedora 20 adopts Wayland. If, 6 months later, Canonical releases a finished Mir that is recognized as demonstrably better than Wayland, and if Fedora refuses to adopt it, then Fedora’s decision gets in the way of my freedom to choose.

            That’s the essence of my point. I see developers asserting *now* that they aren’t going to use Mir. To me, that means that even if Mir turns out to be better, my options will be restricted as a result of the choices those developers made today.

          • Olav Vitters

            Wayland has a stable version out. Mir has a big TODO list. Gtk+3.x has a wayland backend. A few GNOME applications run on it.

            Freedom to choose: Ehr, just use Ubuntu. Fedora’s goal is not to provide you with every choice out there. Try Gentoo or something. Same for Canonical actually, they’re not using Wayland.

          • Olav Vitters

            Seems my comment was lost when Disquss went down.

            Wayland has a stable release. Gtk+ has a backend for Wayland. Various GNOME applications already run under Wayland. We already did a test port of Mutter to Wayland.

            Early test versions: It is a 1.0 release, and one that is actually stable and usable.

        • Michael

          What about driver developers at Nvidia or AMD? Do you care about not being able to have drivers for your graphics card because they had to choose between supporting a single distro or everyone else? Its not about developer freedom its about convenience and standards. Why should a single companies freedom (canonical’s) trump developer freedom?

          People go where developers put their apps, thats why Windows still dominates in many areas because there are tons of programs and games that are windows only, if office, games, and photoshop were on Linux as well as windows I can guarantee more people would be migrating over, so if you are defending Mir for the point of adoption developers matter most. So developer freedom = more applications on more platforms = more choice for the USER which is better for the USER.

          Sorry if I sound rude in my response, this is just a topic I am obviously passionate about and i am trying to explain why its better to support an open standard rather than a closed one :)

          • jon_downfromthetrees

            How is Mir going to stop non-Canonical developers from working on what they want to work on?

            You’re absolutely correct about users following the applications. But, you are absolutely wrong to say I’m “defending Mir”. I’m defending my right as a user to choose what I think are the best products, unconstrained by the wishes or desires of developers. If Wayland turns out to be better, I want to be free to choose it, If Mir is better, I want to be free to choose it. I *do not* want to be placed in the position of finding that the products I choose to use have been compromised because their developers refused to use the better display server.

            As for Nvidia… Well, I expect Nvidia would continue to make the same calculation they must be making today. I.e., that the cost of producing Linux drivers is outweighed by the profit made selling cards to Linux users. If something changes that calculation, I’d expect them to stop releasing Linux drivers. If that meant I could not use Linux (not that it would) I would not use Linux.

          • Michael

            if Mir is successful (unlikely) it will force developers to either have to program for Mir and Wayland or choose between them. Lets say Nvidia developers for example only support Mir and Mir only works in Ubuntu, that means people who buy nvidia hardware will be forced to use Ubuntu to get their stuff to work right, not really giving people much choice.

            I apologize for accusing you of defending Mir, I see now what you are saying that you want to be able to choose Mir. As of what we know right now and the future plans for Mir it wont be anything much different than what you could get from Wayland, also no matter what the display server you would notice almost no difference in performance at all between Mir and Wayland. Which is why the community (myself included) is against this because they are forking very low level tech for the sake of control.

            I’m not sure if you understand exactly what Mir or Wayland do….it wont matter if you are running Mir, or Wayland or X11 on a recent desktop you wont notice any difference, even on a phone or tablet you probably will have no idea which display server you are running as a user. its not something the user notices at all since they with both perform almost exactly the same, only developers and driver creators care/notice which display server is being used because they have to code their stuff specifically for it. It could take over a year of hard work to port a large project to just one display server and you are asking developers support both for no reason? Sounds kind of selfish, I think you should try and write some low level code before making a statement like that.

            I am still unclear on your argument because the user wont ever notice any difference in the server being used only developers will so why should developers be forced to give users choice in something that doesn’t matter to them at all since they are not interacting with it? Users are interacting with what developers write on top of that. Its kind of like giving the users choice on which brand of tires the UPS or FEDEX truck driver uses to bring packages to your doesn’t matter as long as it works. You should be asking Canonical to support what developers are supporting so that you as a users are not left out and have a choice in what you want to run on top of it.

          • jon_downfromthetrees

            >>”if Mir is successful (unlikely) it will force developers to either have to program for Mir and Wayland or choose between them. ”

            That assumes both will have equal merit and capability. If they don’t, why would developers, including Nvidia’s, choose to support the least capable driver?

            I think Canonical’s decision reflects a desire to do what they think is best for their products, even if it means something else for everyone else. Since Canonical is a *corporation that needs to make money*, that’s exactly the choice I would expect them to make. That doesn’t mean I oppose Wayland or despise the Linux developer community.Or that I’m a Canoncial booster. I just think it’s naive to expect people to act against their perception of their best interest. That is exactly what both Canonical and the Wayland community are doing. Our interests are often at odds, legitimately.

            I’m not prepared to assume Wayland won’t be better than Mir, or vice versa, Both are assumed to be better than X. Otherwise, why bother? Still, if Canonical went with Mir solely to control what happens to *their* products, that’s a legitimate decision.

            A tension has always existed between the broader Linux community and the for-profit companies that use it. However, companies like Red Hat and Suse, (and Canonical, too) have dealt in software. When they fork a bit of code so their servers run better, everyone gets the code. Canonical is going off into hardware. This exposes and emphasizes that tension, They want their software to be used to sell proprietary hardware products. The economic imperatives they face, frankly, aren’t amenable to the cultural expectations that have grown up around FOSS. Everyone will get the Canonical code that makes someone’s phone run better, but may or may not do them much good.

            You have no reason to apologize. You don’t need to agree with me.

          • Michael

            But the point I was trying to make is Mir will have no significant improvements over Wayland, so the same thing could be said, what people have an issue with is deliberate fragmentation of the driver stack, no one cares what Canonical does if they want to work in their own projects, or if they are a business. Canonical is getting a backlash because their actions could potentially harm everyone else’s hard work and damage any part of the community that is not Ubuntu.

            Merely having open source code does not justify their actions….if that was the case then would you support an open source GPLv3 licensed virus? What if that virus gave you a 5% increase in how fast your phone runs but steals your information? Is that ok? Should developers support this virus because it makes your phone run faster and is open source? Using your logic above it fits your criteria of what is justified to support.

            I’m not saying that Mir is a virus this is just an analogy but just because something is open source and may or may not have some benefits does not mean they should be allowed to damage something else, or that the damaging effects of its existence should not be pointed out and dealt with before it becomes an issue. Red Hat, Suse, and many other Linux companies make money just fine without damaging the community, there is no reason why they need to do that to make money.

            I was apologizing for accusing you of defending Mir in my previous post, not apologizing for us not agreeing, I actually like to get the perspective of others with a different opinion, helps keep me from getting short sighted on important topics. But I think you need to also look at the damaging effects Mir could have and not just the possibility that it could be good, even if it is good its not going to be a night and day difference over Wayland, and its highly likely that they will perform at near the same exact speed.

            Canonicals decision to make Mir is strictly business, not to deliver a better product, they are doing this so that vendors are locked into making Ubuntu/Mir specific drivers and that users will then be locked into running Ubuntu, thats all it is. I dont ask that you agree with me but I do hope that you consider some of the points I brought up and let me know what you think :)

          • jon_downfromthetrees

            I agree Canonical wants/needs to lock people into their products. (So do Red Hat, Suse, etc. Every business needs to retain customers and is not going to make it easy to move to a competitor.) I wouldn’t expect them to do anything else. Neither would I expect them to bet their business on the work of developers who they don’t pay and don’t control. The technical attributes of either server, or their current status, are, I expect, much less important to Canonical than the fact they will control Mir. They really have no choice in the matter. (This is loosely analogous to Apple’s decision to write its own server for OS X, while still shipping X on the install DVD’s.)

            I think the jury is out on which server will be “better”. As a user, I don’t see why I should want to replace the monopoly of X with the monopoly of an untried Wayland. I recognize the existence of both will annoy developers. Who knows? Maybe someone will create something new and better out of the bits and pieces of both.

            No, I woudn’t “support” a GPL’d virus. I don’t “support” any virus. Or, Wayland, or X, or Mir, for that matter. “Support” is the wrong word. Frankly, I don’t see my use of Linux as requiring me to “support” anyone. I just use what I like.

            I still think this is an issue for a part of the community that is predisposed to be hostile toward Canonical because it has not adopted the accepted culture of that community. I’m not a member of that community. I’m a user. I don’t think that what is seen as good for developers is always the same as what users think is good for users, or vice versa. This is one of those instances.

            Meanwhile, Canonical has, all along, been moving toward an objective that is fundamentally out of synch with open source culture. This episode just highlights the tension that exists between that community and any business that leverages Linux for its own profit. That tension exists for Red Hat and Suse, as well. They have nurtured a different image, though, than Canonical.

          • FFaael

            You don’t understand what opensurce is. FOSS give you right to do things with the code. It do not give the user right to obtain more than the developer (or the company how pay the dev) is ready to give.

          • jon_downfromthetrees

            I’m a user. I don’t want to do things with the code. I could, but I think the argument that FOSS is valuable to users because *users* have the right to alter the code is silly. By definition, once you start editing code you are no longer a user.

            I don’t have the right to make anyone do what they don’t want to do. But, we all certainly have the right to say we hope that Developer A doesn’t limit our future choices by ruling out something made by Developer B. How is it to the advantage of users when developers limit the choices available to users for reasons that are important only to developers?

            Having viable systems that run on three different display servers certainly gives more options to users that having viable systems that run on one or two display servers.

            The imposition of the standardization that would see everyone move from X to Wayland is good for developers but it is not necessarily good for users.

          • FFaael

            obviously. But it’s not good for the user if you don’t can use the applications you like because they support wrong display server. Maybe in the future people you need to shutdown MIR and start wayland or the contrary to use some apps.

            Has you read the blog of the creator of upstart (ubuntus init system) way he don’t use Ubuntu on his homeserver anymore. Apparently no Ubuntu dev cared to maintain the patches for some package that Ubuntu needed (anymore after he left the company) so his server failed sporadically on boot. Both Debian and Red hat has their own pathes for their own init systems but Ubuntu ned patches for upstart instead of systemd or the init in debian.

        • PromoteCommonSense

          Mir doesn’t complicate developers lives. It just means that they now have to use limited programming resources to program to yet another display server instead of fixing bugs and adding features. Either way, the programmers are programming. It is the users who suffer in the end, because a lot of effort is being expended on a compatibility issue that didn’t have to exist.

          Now a very easy solution would be for Canonical to develop a Wayland compositor for their Mir display server. That way developers just need to develop for Wayland as they had been planning on doing and the balance is maintained. However, it is highly unlikely that Canonical will do that because if they do, then their efforts on Mir are wasted. Why? Well, if I develop for Wayland, I can run on all distros that support Wayland, including Ubuntu (if Mir supported Wayland). If Mir didn’t support Wayland, then I must develop for Mir specifically. Canonical’s assumption is that they are the biggest distro, so developers will choose to develop for them. Maybe, maybe not.

          It does mean, as a developer, I have a choice to make and that choice boils down to do I want to develop for one specific distro or all the rest of them or do both? Of course, if I do both, then the programmers maintaining the code for the multiple display servers aren’t available to code for new features and bug fixes.

          So as a user, what would you prefer developers work on? Do you want them fixing bugs and adding features or adding compatibility with incompatible display servers?

    • Adonis K.

      Maybe 99% of the Ubuntu users

    • foobar

      It is not only about “the best working solution”, it is also about the confident one. We simply don’t have the manpower and the will to reverse our decisions every year. We prefer to do stuff once and do it good / in a way that allows us to refactor it good.

      Mir is the unsecure choice. Canonical proved itself to be unreliable about their own technology choice several time. We don’t have the ability to change anything without signing a contract with a player who doesn’t play fair all the time. We have to ask them to accept patches and make sure that it has values for *them* as well.

      It is like building a house without a geological profile. It sometimes work quite well but you’ll never know.

    • JJ

      Android display server repackaged from canonical? No thanks. May be this time next year they will be chasing firefox drivers.

    • ScionicSpectre

      I think there are clear advantages to using Wayland from a pragmatic point of view, outside of the cultural repercussions. Wayland’s design is well thought out by technical experts in the field of display technology. The best minds from the Xorg community and Intel are working together and they have been diligent in doing it correctly. This is part of the reason it hasn’t happened overnight- they’re really paying attention to doing it right, without imposing unnecessary structures on the system.

      Mir, on the other hand, has not proven that it will be ready in less than five years- that’s the hope. It may well be that Mir just uses XMir for a long time until it’s fully capable of supporting everything Ubuntu needs. If Mir does prove to be the more functionally promising technology, that will warrant some merit.

      But if this happens only due to the support that Ubuntu has, I think it is a great loss for all of us that we would allow Canonical that much power. If NVIDIA makes binary drivers for Mir but not Wayland, for example, it would be easy to pretend that Mir is better merely because it has support, and then we’re fighting uphill to get what users and developers really want. I think it’s impossible to separate the social aspect of Mir from the situation- it just so happens that Wayland is the better decision, right now, from a technical and ideological standpoint. It may well be that Mir and Wayland find some sort of compatibility with each other in the future if it’s necessary. Of course, if GTK and Qt are both supported across all of these technologies, the core interface and drivers will be the only exclusive features.

      • jon_downfromthetrees

        I’d guess the appearance of 3 display servers in Linux might well convince Nvidia to stop writing Linux drivers altogether. It’s not like it’s a profit center for them.

        What if the circumstances were reversed? What if Mir was nearing completion and, all of a sudden, part of the community announces they’ve been working on a different, new, display server for 6 months?

        Like I’ve said, I’m a user and I look at things from the perspective of a user. I don’t think that perspective is, or can be, always the same as a developer’s. To me, Mir looks like competition for Wayland, That’s a good thing for users.

        • ScionicSpectre

          I think your personal opinions are on the mark. The main issue for me is which is better, and from the design perspective, as well as demonstrations of actual usefulness and performance, Wayland is way ahead. The only feature Mir offers over Wayland is that Ubuntu is planning to adopt it. Also, Mir will likely focus on Unity first, so porting other desktops probably won’t be feasible for a while.

          • jon_downfromthetrees

            I’m not a Unity fan. I use Gnome Shell on Fedora and like it. Canonical wants to leverage as much common software as possible across desktop, tablets, phones, etc. That’s rational. I’m really doubtful, though, about their chances for long-term success in tablets and phones.

      • alex285

        > But if this happens only due to the support that Ubuntu has

        That is quite true. I read somewhere about the concerns that Ubuntu has for releasing the 2 years LTS + the Rolling release. Someone was referring that System76 weren’t happy with that. That is the size of support that Ubuntu has from manufacturers. Okay I am trolling ;)

        I still believe the only support that Ubuntu has right now is from community projects and community projects only (ie Libre).Also I think they risk to go against everyone, even if they had/have right..

        • jon_downfromthetrees

          While I think no one outside the enterprise arena has much of a chance making money in open source, as long as Ubuntu consists of open source and as long as Shuttleworth wants to pay his developers, I expect Canonical to produce Ubuntu. I do not think he will be successful in turning Ubuntu into a profit center. So, I expect, eventually, to see Ubuntu follow the same path as Fedora and OpenSuse: Become the open source quasi-testbed wing of a for-profit business focused on the enterprise. Or just vanish.

          • alex285

            I agree 100%, also I agree with a previous comment of yours that you were saying that you don’t believe that Canonical will make it in mobile market. The reasons are more than obvious. All big software vendors are now targeting to make hardware. And even if Canonical make it, that would be the 8th miracle of the world. It’s like one in the millions.

            I just can’t understand how Mark S. is so optimistic. He always saying they will make amazing things, in Ubuntu Touch videos “says” that Ubuntu (which does’t even work) is better than Android and Win8 etc etc. There is a mental illness is called Vanity ..I don’t know what else to say.

            However as ex-Ubuntu user, I can’t but thank Canonical for their amazing releases

    • Olav Vitters

      This is not proposed to annoy Canonical. We planned to go with Wayland
      for a long time. We’ve been testing how to use Wayland. We had already a
      working implementation. Wayland is dependable, Mir: only announced
      almost a year after the plan was made. It is not something you want to
      depend on.

      I am not affiliated with any open source or free software company.

    • Brandon Watkins

      Mir won’t deliver anything sooner, mir is significantly less complete than wayland, and wayland developers are mostly former X developers that have far more experience. Don’t take canonical’s lofty claims as something that’s actually going to happen.

      • jon_downfromthetrees

        Pretty sure I’m not cheerleading for Mir or Wayland, If I’m cheerleading for anyone, it’s users, who I want to be able to use whatever turns out to be the better display server, in the distribution of their choice. I think users like me might be forgiven for believing some developers have already ruled out using Mir, no matter what. That’s as hostile to users as Canonical ruling *in* Mir, no matter what.

        • PromoteCommonSense

          As a developer, changing display servers is not trivial, so many developers have already ruled out using Mir, not on technical reasons, but because a lot of time and resources have been invested in Wayland and Unless contracted to a specific setup, I try, as do most others, to target the widest userbase as possible. That is not Mir for 2 reasons. First, it doesn’t exist yet. Second, it is specific to only a single distro.

          So, as a developer, I will continue to work with as the display driver for Ubuntu systems. Currently that’s all they support and when Mir is available, it has a compatability layer. For everything else, I am moving towards Wayland, because that is where everything else is going.

          That’s not being hostile, it’s being practical and is in the best interest of users.

          • jon_downfromthetrees

            All reasonable and understandable. My point, though, is this: If Mir turns out to be the better driver for users (and we don’t know because, as you say, it doesn’t exist) then resistance to Mir does not serve the best interests of users. The “only available on Ubuntu” business won’t matter to users who would move to Ubuntu w/Mir if it offered obvious benefits.

          • PromoteCommonSense

            You are thinking like a single user. I’m involved with a 10,000 seat deployment and this really is a serious issue. Serious enough that the client is re-evaluating whether or not to use Ubuntu. Mir will probably be fine, but it’s not even ready and the way the change came about creates a lot of uncertainty for users, particularly enterprise users (which is where Canonical needs to be if they want to be successful). Enterprise users can deal with change, that’s not the issue. What they need, however, is predictability.

            In our situation, we were about two-thirds of the way done with converting from X to Wayland. We still have the X code so we can bring it back to life if the client wants to go that route. The other viable alternative is to use a different base than Ubuntu. At this time, it is not even an option to develop for Mir, because Mir won’t be ready in time for their schedule. They are currently evaluating their choices.
            Like many others, the client doesn’t understand what all of this is about or why it matters. They shouldn’t have to, either. But, it makes all of us involved with FOSS and Linux look like idiots and that does nobody any good (except Apple and Microsoft). The sad part is that all of this could have been adverted with better communication. In the past, that is something Canonical was very good at.

          • jon_downfromthetrees

            Well, yes. I’ve been in similar positions in the past, albeit with fewer than 10,000 seats. The people paying the bills do not like this kind of change. With a few exceptions, open source churns much faster than Microsoft and Apple. I learned early on that no one wanted to hear about Linux. So, rather than risk my reputation, I shut up about it.

          • PromoteCommonSense

            While an NDA keeps me from saying too much, this is the largest Linux deployment I have been involved with. Most are between 100 and 1,000 workstations. Linux on the desktop isn’t dead, at least not in certain specific use case scenarios. I’m not sure about its viability on consumer desktops, but there is definitely a lot of demand for entities that are in a position to want to maintain their own source code once a project is completed.

            Where I have found a lot of use, albeit in smaller installations is when it is no longer possible to economically maintain small mainframes or minicomputers, those shops are much more willing to go with a linux solution versus a Windows or OS X solution. For 100% Windows shops that have outgrown their current system, a non-Windows solution is a much harder sell (which is why we also develop and deploy for that, too).

            What I find really fascinating, given all the fuss that gnome-shell has received in the linux community, is that it is really well received in the real world in the enterprise. From the IT side, it is easy to support. From the user side, it is modern and functional and from management, real world case studies do support that productivity goes up. Can it be improved, for sure, but then so can most things.
            What I have also found is that since the late 90s more and more kids have been exposed to Linux in colleges and universities and they have now been in the work force long enough where they are in decision making positions. That doesn’t mean they will automatically choose Linux, far from it. But, they don’t have that knee jerk reaction that was common just a few years ago when anything but Microsoft was suggested. In the 1970s the slogan was “Nobody ever got fired for buying IBM.” By the 1998, it was “Nobody ever got fired for buying Microsoft.” Today, there is no slogan, managers just want the right tool for the job at hand and very often, that tool is a Linux solution.

          • jon_downfromthetrees

            I’ve come across reports — I can’t say how valid — that new users take to gnome-shell more than current Gnome 2 users. That doesn’t surprise me, given our general reluctance to change unless we are given a reason. Gnome developers had reasons to move on from Gnome 2, but not so much Gnome 2 users.

            Good luck with that deployment.

    • Mike Manilone

      I’d say the opposite. Mir performs as a very low level actor in the graphics system (as the post said). To make GNOME work on it, there is too much work for it. Yes, driver, toolkit, window manager… However, I don’t know what will we get from supporting it. (@_@) Um, just because it’s a product from C?

      • jon_downfromthetrees

        Gnome probably gets nothing from using Mir, and I gather it won’t. That’s not my point, though. I think it’s OK for Canonical to go off on its own with Mir to support its own line of products. Clearly, that also means a great deal of code in open source will not be rewritten to accommodate itself to Mir and Canonical. Nothing wrong with that. Canonical does not seem to be particularly stupid, so I will give them credit for knowing that. It also means Ubuntu and Ubuntu-derived code will become increasingly distinct from the rest of open source. I suspect Shuttleworth, himself, crossed that particular Rubicon a long time ago. I take him at his word about open source, but he has always positioned Ubuntu as something distinct from the rest of Linux, even when it has’t been.

        So… why get upset about Mir when you plan to ignore it, anyway? Sure, it’s possible that Nvidia/AMD might choose to support only Mir. But, the possibility/probability that proprietary hardware vendors would follow their wallets has always been there. I still find it difficult to understand why they bother to write Linux drivers at all.

    • PromoteCommonSense

      The real problem is what if KDE needs display server A and Gnome needs display server B and XFCE needs display server C and e17 needs display server D and Unity needs display server E. While X has its problems at least it was consistent regardless of the desktop and applications. You are correct, that users do not care about such things, but they do expect things to work and if logging out of Gnome and going into XFCE means you need to drop to a command line to kill the display server how will that be perceived. Or right now, I can run gnome or Unity and still use K3B to burn DVDs. How will that work if KDE and KDE apps use a different display server than the default desktop the user has chosen?

      Sure, for now, if everybody still uses X, Mir has an X session compatibility layer. But, Mir, being a Canonical specific project, that layer is not guaranteed indefinitely and if the other environments go to support Wayland, they won’t work with Mir.

      For that last point, Canonical says, well, if somebody wants to write a compositor for Wayland, that would be fine, but they aren’t going to do it. So, in other words, if you want your stuff to work on Ubuntu, with a modern display server (not X), you must use Mir and if you want it to run everywhere else you must use Wayland. Canonical’s position makes sense, if they write the compositor for Wayland for their own display server, there won’t be a need for anybody to build in support for Mir. On the other hand, unless debian is going to go with Mir, which it appears they are not, at some point Canonical is not going to be able to be debian based anymore.

      • jon_downfromthetrees

        That last “If your goal is to use the best software, use it” is directed at users. Of course, as Canonical moves Ubuntu more and more away from the Linux mainstream, that means deciding you want to use Ubuntu will probably lock you out of using an increasing amount of mainstream Linux software. I don’t think that’s at all important to Canonical, and maybe of surprisingly little importance to many (most?) Ubuntu users.

        Rather than worrying about the impact of Canonical’s choices, why not accelerate Ubuntu’s isolation by ignoring it? Write for Wayland, not for Mir. Put aside the angst about Ubuntu’s current popularity. Shuttleworth obviously has believed for a long time that he can’t base his products on software from developers who aren’t on his payroll, and, hence, can’t control. His decision to go it alone with Mir is surely not going to be the last.

        • PromoteCommonSense

          Ignoring Canonical’s choices is what people effectively did with Unity. Canonical went their way, Gnome went their way and people pretty much kept on developing with minor modifications for changes for SSO and notification APIs. However, the switching display servers has a lot bigger impact than switching desktop environments. I could still develop for KDE, Gnome, XFCE, etc. without making any changes to accommodate Unity and my app would function just fine, but without Unity specific features. I can’t do that if the display driver changes as my app has to work with the display driver, regardless of what desktop environment is in use.
          The problem with ignoring Ubuntu’s choices is that is an end-user decision. If I am a developer or integrator, I have to be in a position to support my end users and they may want to run on Ubuntu, so I, as a developer or integrator, can’t just ignore what Ubuntu does. There choices can impact my bottom line. Now, let’s say I am System76 and I want to also offer Redhat pre-installed. Now my hardware has to support Mir for Ubuntu and Wayland for Redhat and I have to be able to troubleshoot and offer support for both. That drives up my cost (I am not affiliated with System76, I just use them as an example), which either I must eat or pass on to my customers or choose to only support Ubuntu or Redhat, thus turning away customers who want the other. No matter my choice, it impacts my bottom line and my customer’s bottom line.

          Canonical operates in the real world, not in a vacuum and their decisions they make have real world implications. It is not enough to simply “ignore” them. Like it or not, it is more akin to a big pond and their decisions are like a pebble thrown into it. Eventually the ripples spread across the entire pond. This is true whether a small pebble, like Unity or a big rock like Mir. Their choices impact everybody in the pond.

          This isn’t unique to Canoncial, Redhat, Suse, Gnome, KDE and all the rest share this same pond. It’s just that Canonical seems to be the one who most often throws the most disruptive rocks.

  • JJ

    I love it when someone stands up to challenge canonical.
    And when that someone is an emperor like gnome I love it more than ever. Way to go.

    After porting, “Most parts of GNOME will still work under X (some parts may be hard to keep working under both X and Wayland)”. I see a bright future :)

    • PromoteCommonSense

      I think by that time, most distros will have switched over to Wayland.

  • ScionicSpectre

    I think it’s important that we offer support for Canonical insofar as we are responsible, just as we would support Fedora or openSuSE. But with Mir, I don’t think anyone should make concessions when we already have well laid plans for Wayland. If it were any other distro, people would laugh it off- we shouldn’t let Ubuntu’s popularity destroy what we have as a broader community.

    Besides, there is no obvious way of supporting Mir on other distributions at the moment, and no proof that it would be useful to do so. I think Canonical is going to use form-factor convergence as their leverage to gain support from the community. We need more contributors to make sure companies won’t dictate our technology, since Canonical’s strategy is outpacing KDE and GNOME in several areas.

  • Matthias Clasen

    Just to be clear: I do not speak for the GNOME foundation, and what you are citing here are internal discussions of the release team.

  • Guest

    What do you need to do to make Gnome shell run on Mir?

  • jeroen_neth

    What work needs to be done to make Gnome Shell run on Mir?

    • alex285

      From Christopher, Mir Developer

      “Most likely, you’ll never need to care – X isn’t going away and we’ll need to support X applications for the forseeable future. Given that, I expect that you’ll be able to run GNOME Shell in an X session on Mir for a good long while.
      Only if GNOME decides to reimplement Shell as a Wayland compositor would this be at all problematic; at that point you might not be able to run Shell on Ubuntu, at least until someone submits a Mir backend for their compositor.”

      • jeroen_neth


      • FFaael

        Ehh… I don’t use gnome but I pretty sure that is what they do according to the article we comment…

  • Philip Witte

    *sigh*… I’ve wasted too much time talking about over the last few days. Such a silly move by Canonical, and one they keep trying to justify with “technical” reasons instead of just admitting that they wanted full control over the project. Oh well, hopefully Nvidia & AMD will make EGL Display Server agnostic drivers which support both Wayland and Mir in the near future, and then it wont be such a huge mess.

    On a side note, It’s good to hear that the Gnome developers are actually sane :-)

    • Vincent O’Neil

      And so what if they want full control of the project? What’s that to you and me?

      • Rowan Lewis

        Nothing at all, and that was his point.

        The problem is that they keep hiding behind bullshit excuses, or even outright lies, and that for nine months everyone was under the impression that they would use Wayland when Canonical knew full well that they would not be.

        • Vincent O’Neil

          “bullshit excuses”? “outright lies”? Got anything concrete to back that up?

          • Rowan Lewis

            It’s not their anyone elses concern, and again that is not the point.

            The problem is that again they failed to communicate with the rest of the community what their plans where, and then when they did “communicate” with the rest of us they chose to do in in the form of a press release that claimed “Wayland isn’t good enough.” Unfortunately it was quickly shown that their reasoning was faulty, but the damage was already done because for those nine months they never though to ask, or talk about their plans.

            Since then Mark Shuttleworth has made some outrageous claims on Google Plus like “Unity came before Gnome Shell”, which is one of the outright lies that I’m aware of.

            But of course you chose to come here and “dispute” this while obviously not being fully aware of the facts, and then turn around and call everyone else “self-righteous.”

          • Vincent O’Neil

            If it’s not anyone’s concern, so what if they miscommunicated? What “damage” was done? How is Mir damaging Wayland? Why must Canonical divulge or discuss their plans with any other party outside of the company?

          • Rowan Lewis

            Just a few things; driver fragmentation and massive duplication of effort, wasted effort developing Wayland based software targeted at Ubuntu, making Wayland out to be broken when it’s not.

            All of those and more are discussed on this very page, and in the comment sections of the news sites reporting on this, and on the Google Plus pages of the people involved.

          • Vincent O’Neil

            Canonical never said Wayland was broken nor implied such. People, being highly defensive and quick to jump to conclusions, imagined that that was what they meant. As for fragmentation, Linux has already got a lot of that. Some people like to call it “diversity” in the Open Source world. True enough, drivers may need to be specifically compatible with Ubuntu but that’s not necessarily a bad thing. Legacy X-based apps will be able to run on Ubuntu with Mir according to MirSpec.

          • Rowan Lewis

            They may not have intended to imply as such, but their list of issues with Wayland (that turned out to be their own misunderstandings) lead people to believe that it was broken, and that is the angle that was reported.

            But all of this is discussed on this page already, if you truly wish to understand the issue, I suggest you continue reading instead of sticking your fingers in your ears and crying “nah-uh!”

          • Vincent O’Neil

            I already know what incredible leaps of reasoning anti-Canonical naysayers are using. Such as these actually want Canonical to fail. The main concern here should be the users and having read MirSpec I don’t believe there is anything wrong with what Canonical is doing.
            I’ve already read most of these posts as well as those by members of the Wayland team and of course they aren’t happy that Ubuntu will use Mir but that’s not their call to make. The authors of MirSpec already sorted out the issue but still others insist on being hypersensitive and acting deeply offended for the sake of turning more people against Canonical and Ubuntu. To what end they engage in such pettiness is beyond me.
            It is no secret that the majority of this Linux community hates Mir but I think that if Mir succeeds (and no one knows if it will) it will be a positive direction for fulfilling Canonical’s vision for Ubuntu. I think that Canonical needs complete control over a project like this and cannot afford to leave it in the hands of the whim and fancy of external developers, especially when something like this lies so close to their goal of platform convergence.
            What if Wayland team rejects a change or a patch that Canonical wants? Or they refuse to trim off certain things that Ubuntu doesn’t need? Canonical will then be forced to fork or heavily modify Weston and will be subject to the same hatred that they’re receiving now. There is simply no way to win when playing this game: you succumb to criticism and use (and modifiy) Wayland: people will complain. You create your own display server to avoid that: people still complain.

            At this point, Canonical has stopped being sensitive to such complaints and is doing what they believe is best for Ubuntu.

          • Rowan Lewis

            So the hypothetical “Wayland may not submit to all of our whims” trumps the very real issues that everyone except you appears to understand.

            Got it. No wait, fuck that.

          • Vincent O’Neil

            Huh? Did you read what I wrote? It’s because the display server is so critical to what they are going to do (Unity Next) that they need this level of control. Nothing I’m saying here is new: it’s all in the MirSpec document. Part of the reason they are choosing to create Mir is so that they don’t have to come into conflict with the Wayland devs who don’t particularly care about Unity Next or Ubuntu for that matter. Canonical is trying to do what is best for Ubuntu, not necessarily what is best for the rest of the community. That may be a bad thing or a good thing depending on your outlook.

          • Rowan Lewis

            Holy shit, did you just try to make that point again?

            Nobody could give a rats arse what Canonical do internally, however when it affects the rest of the Linux community, they do owe us a little respect, which they totally failed to provide. Of course they don’t have to be respectful, they don’t have to fit into the wider community, but if that’s their choice then you’re going to continue to see us moan about it.

          • Rowan Lewis

            Further, because of the way Canonical have behaved, they’ve begun building walls between themselves and the rest of the community. This is probably going to make it more difficult to do what they need to do for their users than if they had just done the right thing in the first place.

          • Vincent O’Neil

            Yes, I do foresee that this and further developments may contribute towards isolating Canonical and their ecosystem from the rest of the Linux community. I don’t, however, view this as something necessarily bad. Remember, Canonical is operating under loss right now, I wouldn’t be surprised that being liked or accepted by the Linux community is somewhat low on their list of priorities right now.

            This is just pure conjecture but I think that gaining a competitive advantage against other Linux distributions on the various computing form-factors may be their current aim as well as increasing their market share on the desktop.

            As they progress along the path towards distinguishing their product from the rest, it is inevitable that changes like this will alienate various stakeholders in the Linux ecosystem. Can’t get by without treading on some toes :-)

          • Rowan Lewis

            You’re right, but to be honest I’m more concerned about Mark Shuttleworths sanity. He’s been coming across as very out of touch on the Google Plus discussions about this, doing a bit conspiracy mongering.

            Fingers crossed that they get it done though, I’m not a fan of Android or iOS, and the N9 was end of lined, so I’m not sure who else could pull it off at this stage.

          • PromoteCommonSense

            Canonical is not trying to do what is best for Ubuntu. They are trying to do what is best for Canonical. If the desktop is truly dead, then so is Ubuntu. That is what it appears that Canonical believes, anyway — desktop linux is dead, tablets and mobile are the new world and that is what they are going for. However if the desktop is dead, then so is Ubuntu. Even Unity is dead. Canoncial bet on netbooks, which is what Unity was designed for. However, now we have Unity Next, which targets tablets and mobile. Switching to Mir is only because they are abandonding the desktop. Oh, they will make Mir work with the desktop, but by their own admission, that is no longer their focus.

            So, quit repeating that Canonical is doing what is best for Ubuntu, they aren’t. They are doing what is best for Canonical.

          • Vincent O’Neil

            Where on earth are you getting this info? Are you making it up as you go along or are you spending too much time in Phoronix forums? You mentioned FUD earlier but do you know what you’re doing right now? Spreading misinformation that will directly feed into FUD. I don’t think anyone who has read anything documented by Canonical with regards to the future of Ubuntu will take you seriously so I hope that you don’t genuinely believe anything you wrote above. Let me show you the light:

            First read the entire Mir Specification:
            Then read the entire Unity Next Specification:
            Then read this post:
            Concerning fragmented drivers:
            Concerning effects on other Desktop Environments:

            If after reading all of that you still believe exactly what you wrote up there, I’m afraid there is nothing anyone can say or do to help you.

          • PromoteCommonSense

            Announcing a direction to give you time to work in secret on an alternative direction and then come out and denounce the original (Wayland) as inadequate creates FUD. One only has to go to various sources to see that Canonical has created a lot of uncertainty and doubt about the future of linux, gnome, kde, xfce, basically everything not Canonical. If it isn’t FUD what would you call it?

            From Canonical’s own account, they did not work on Wayland at all in those 9 months, but instead worked on their own Mir. That seems to indicate that the announcement regarding Wayland was a stall tactic. Businesses do it all the time. Canonical isn’t evil for doing it, but they aren’t noble, either.

            Canonical has made it quite clear in their official and unofficial documents that they see the future in mobile platforms, not the desktop. There current base, however is the desktop, so they need to say things to keep the base engaged, while they develop for the future. Again, every business does this.

            Likewise, when other businesses do this and companies and customers that rely on the directions that are given find out they were wrong, people are pissed off. Cannonical pissed off a lot of people by telling them they were going to do one thing while secretly planning to do something else. Again, that happens all the time in business. What is surprising, though, is that they and their defenders think that people shouldn’t be upset by that. That does not happen elsewhere in business.

            Canonical should look at this positively. They are being treated like a real company instead of just some hobby entity. Of course, that means, that like a real company, they no longer have a user base, but now have a customer base and customers are much more fickle than users.

            As for the links you posted, you need to canvas a lot more than sites outside the Canonical sphere of influence. In the recent US elections, their republican party seemed to rely too much on their own spin results and were quite surprised. Canonical needs to listen not only to their supporters, but to their detractors if they want to be successful. I wish them all the luck, but in the end, it isn’t luck that will determine if they are successful, but actual business accumen.

          • Vincent O’Neil

            “Announcing a direction to give you time to work in secret on an alternative direction”
            Again with the speculation, you have no proof. If you do, I’d love to see it.

            “and denounce the original (Wayland) as inadequate”
            You’re now intentionally using words with negative connotations to twist a situation to conveniently suit your point. According to Mirriam-Webster to denounce is “to pronounce especially publicly to be blameworthy or evil”. Show me where Canonical does this.

            If you’d read any of the links I posted you will know what they’ve said regarding desktop environments and drivers but it seems that you are no longer interested in finding out facts and are instead content to theorize and believe the worst. Believe what you want.

          • PromoteCommonSense

            If you want proof, just look at the commit logs. Usually if one is actively working on a project an objective measure like commits and checkouts is a good measure. Lack of this activity from the project would indicate that one is not working actively on that project.

            I’m pretty sure if Canonical felt that Wayland were adequate for the task they intended, nobody would be having this discussion. They did, in fact, state that Wayland was not adequate for their purposes, thus it is inadequate.

            And yes, I did read all of those links. But in any serious discussion, using your own position’s generated documentation to support your position instead of independent support is kind of like using the word you are trying to define in the definition itself. There is merit to what in those links, but they are not by any means definitive. That is why I posted the comment regarding the US presidential election and what happens when you only listen to your own side’s spin.

            There are those who believe Ubuntu is the best thing that every happened in world and we should all be honored to eat the scraps that fall from their table. I don’t happened to be one of them. I do think Ubuntu has done many good things, but they have also done some pretty lousy things, too. If people are unwilling to see that, then that is their problem not mine.

            I would suggest that we just agree to disagree. It seems that your point is that Canonical did nothing wrong and mine is that the way they did made their Wayland/Mir decisions and announcements caused difficulty for people. 100 years from now, nobody will care.

          • PromoteCommonSense

            The problem has nothing to do with Mir. Mir may be better than Wayland, or it may not. Mir is not the problem. The problem is that when Canonical announces a direction, people make decisions based on the announced direction. So, when Canonical switches direction again or as with Wayland, never intended to go that route in the firt place, it causes real problems for people because Canoncial cannot be trusted. There is an allegory about a certain boy who cries “wolf” once too often and if Canonical isn’t careful, they will find themself in a similar situation with developers and integrators. If that happens, it won’t matter how good their product is.

            Your Wayland/Mir scenario isn’t realistic. We aren’t talking about applications, but core parts of the system. What if Linus rejects a patch that Canonical wants, will they just create their own kernel? Going to Mir impacts all desktop environments. Currently none of them will work with Mir except Unity. Going Mir seems a little drastic if the situation was really about a patch or two.

            Hopefully, you are wrong about Canonical ignoring complaints. If they aren’t careful, Ubuntu might be the next Betamax.

          • Vincent O’Neil

            The Wayland/Mir scenario is realistic, in fact it has been cited several times by the Canonical reps so I don’t see how they would be saying that if it wasn’t a genuine concern. The whole “create their own kernel” thing is getting a bit old. There is no logical reason to speculate that Canonical will ever need to create their own kernel. They already patch the kernel that’s shipped with Ubuntu and the kernel is an extremely versatile piece of software that has been proven to be customisable to fit a vast variety of needs. Wait until you see the changes next year before you start to panic and hit the Eject button. If things don’t work out, you can always move your apps to Red Hat Enterprise Linux, Fedora and OpenSUSE.

          • PromoteCommonSense

            Well, Canonical did say they were pursuing wayland about 9 months ago, so my projects and others started to make changes to support Canonical’s announced change. Now, it turns out that while they said that, they were not planning that at all and instead were planning an entirely different infrastructure than what they publicly disclosed. That means I have wasted time and resources on paying programmers and staff, so I would call that damaging.

            Do they have the right to misdirect by saying one thing while secretly planning something else? Yes, they do, just like Microsoft and Apple do. But please, don’t pretend that it doesn’t impact anybody. FUD impacts everybody.

          • Vincent O’Neil

            It happens from time to time in technology when things change. Perhaps the Mir thing was handled a bit badly on their part. The way I see it, Mir was probably just a pet project they were toying around with until the decision was suddenly made to actually utilise it.
            All your time and resources have not been wasted. All X-based apps will still be able to run on Mir so please read the Mir Spec and stop contributing to the FUD, whether intentionally or not.

          • PromoteCommonSense

            Since you do not know what my staff has been working on, you have no idea as to whether or not our time and resources have been wasted, so please, don’t patronize us.

            The issue at hand is not Wayland versus Mir. Mir, from all that I have read, seems like it will fit the bill for what Canonical is wanting to do, and in some ways much better than Wayland ever could. What is at issue is that less than 1 year ago, they announced that Wayland was their direction, then did next to zero work on it, at least as measured by commits, while pouring all their resources into Mir. This is a different situation than what happened with gnome-shell and Unity where Canonical was trying to get things changed and couldn’t, so they went their own way.

            But when one of the largest distributions announces a strategic direction, and yes, a core component like X/Wayland/Mir, is a strategic direction, with no intention of following that direction from the outset, it causes real problems. This has nothing to do with the Mir spec.

            And yes, it is FUD – Fear, probably not in this case, but definitely the way they handled it has caused a tremendous amount of Uncertainty and Doubt.

            Apple can get away with crap like this because their fanboys are, well fanboys and will take all sorts of abuse. As gnome and kde developers found out Linux users don’t have the same type of loyalty. I commend Ubuntu in trying to reach out into new markets and even new user bases, but if they keep alienating their existing users and developers, they will be in real trouble.

            In business theory, there is a thing called “fat cat syndrome” where a business basically gets too big and arrogant for it’s own good. It happens in all different industries, but in the tech fields, It happened to IBM, it happened to Microsoft, it happened to Yahoo, Wordperfect, HP, Lotus and any number of companies that were at the top of their field or industry. And, if they aren’t careful, it could easily happen to Canonical.

          • Vincent O’Neil

            No, I don’t know what your business is working on, it’s none of my concern. If I were in your position I would educate myself properly, reach out to the Canonical devs and try to salvage what can be salvaged. What I would not do is make a rash decision without knowing the whole picture and only based on hearsay. I assume you did the former, of course. In addition, 12.04 is LTS so there’s no reason why you shouldn’t release your application on that and it will give you more time until you need to move to the next platform.
            Mark announced Wayland would be the direction in 2010:

            Businesses sometimes make premature announcements and have to backtrack when their aims change. That’s something you have to expect in the rapidly changing world of Linux.

            How can you know they had “no intention of following that direction from the outset”? Did you speak with Mark Shuttleworth or another person at Canonical with knowledge and authority to make that statement? If not, then you are offering your own speculation as fact.

            As for your Fat Cat Syndrome analogy, I think you need to look at the size of Ubuntu as compared to Red Hat and Novell before you assume that they are becoming too big. If anyone, Red Hat and Novell are probably more susceptible to Fat Cat Syndrome.

          • shanefagan

            Actually Unity did come before Shell in the form of UNR it just evolved into what Unity is right now with 2 full rewrites.

          • Rowan Lewis

            That’s not true at all, it’s a completely different codebase. If Unity came from UNR then it also came from many mobile devices. Marks claim was that it existed before Gnome Shell, which it did not.

          • alex285

            Not really. Shell was before Unity. You can see the whole story at McCann’s comment.

            Unity announced in April of 2010, I think.

          • shanefagan

            It doesn’t matter when the name came out the idea was already out and the principles that were in UNR went right into what became unity. Its about the design and the iteration on those designs which definitely were before Shell was talked about. The difference is that after Shell started to go down roads that even remotely sounded like they wouldn’t be what Mark wanted they started working directly on plan B which was Unity on the desktop which was what was branded Unity. Obviously it couldn’t have been called UNR on the desktop and there was no specific name other than launcher and netbook interface flying around so it was branded Unity then.

            Rowan just because it is a different code base doesn’t mean its any different either because they went through radical design changes as well but still a lot of the underlying ideas were the exact same. Things like giving a lot of vertical room and the dash were all designed well before Shell was talked about.

          • alex285

            Yes you are right, my bad. I thought UNR was newer.

          • PromoteCommonSense

            According to Canonical, Unity did not come from UNR, but came from differences between what they thought the interface in gnome-shell should be and what the gnome-shell developers thought it should be. Again, according to Canonical, they did learn a lot from UNR, but UNR was specific to netbooks and Unity was not.

            It really isn’t a question of which came first as they aren’t even related. Today’s gnome-shell and today’s Unity are both inspired from the early Gnome 3 GUI work. Much of that early design work didn’t make it into the final gnome-shell or Unity. Likewise, both gnome-shell and Unity built upon those ideas but went ins similar but different directions. But neither was first.

  • Xiaojun Ma

    People bet on GNOME, a DE mainly developed by a non-player company in desktop / mobile market.
    This post is yet another example to show why GNOME unseating Mac OS X/Windows is a hopeless dream; Wayland migration to accelerate because some alternative FOSS appears rather than proprietary software has been working better for years.

    People also bet on Wayland developed by several, if not one, staff in Intel, which is a display server that you cannot technically attack but not usable after end of the world, after the release of Windows 8, etc.

    I hope people understand how economy works.

    If some people want to follow the RMS way, I respect you. But I’d remind you that the current official GNU desktop seems to be GNUstep, I don’t know why.

    • FFaael

      opensuse and fedora is both communitydistribution and they both is community driven.
      At least opensuse is a very stable distro and not some sort of bet distro for sled.

      • Xiaojun Ma

        Well, I do admit that openSUSE is a decent distribution. Then what? You ever see it pre-installed? Who to provide commercial support? Any ISV cares about it? I hope you could think about the real reasons why Linux desktop as a whole is still a 1% niche. Somehow better than peer Linux distributions doesn’t matter much for most people.

        • FFaael

          Can not say I care much what is preinstalled. I don’t think any of the linux companies we have today has any real chance to threaten Apple and Microsoft on consumers desktops.

          The so hyped Canonical, as it is today make a pretty good but not better linux dist than most of the rest of the linux alternatives you can choose between. Besides that what have they more, Ubuntu TV? Did they ever produce anything better than the community alternatives like xmbc or mythtv? No I’m not sure Canonical really can surpass the boarder foss community.

          • Xiaojun Ma

            Yes, Canonical probably die right before the release of Ubuntu 14.04. I just feel that the community or semi-community (e.g. Red Hat) effort is largely a failure; I haven’t seen notable Linux adoption for more 10 years. (IIRC, Linux around year 2000 already have seemingly good looking desktop.) So what, let Canonical try wild things and see whether they die or survive.

          • Helder Pereira

            While it would be great for Linux to gain market share on the desktop,
            it’s no disaster if it doesn’t. Linux has been serving me well so far,
            even with a 1 to 3% market share. Granted I still need proprietary
            software Linux doesn’t have for some stuff, but it’s been the OS I spend
            most time with. Semi-community, as you put it, is doing great things
            with Fedora and OpenSüSE.
            I applaud Cannonical’s ambitions for the
            desktop, but the ends do not justify the means, they do not have carte
            blanche to rain on the FOSS/Linux community parade because they’re
            bigger. Nowadays, Canonical seems to have issues even admitting Ubuntu
            is a Linux (GNU/Linux, if you prefer) distro. What’s next, replacing the

          • Xiaojun Ma

            Why GitHub don’t have a native client for Linux? Not to mention other software.
            I’ve been accustomed to CLI to a point that I have Terminal kept in Dock when I’m using OSX.
            But I still want GUI in many case, but Unix/Linux in general often get poor support for GUI part, if they are supported at all.
            Also not to mention hardware issues when buying and using some hardware, way too painful.
            The above are reasons why I want to see Linux gain market share.

            For Canonical case, spare me now and I have a mid-term several hours later.

        • Bartowski

          Linux desktops do not need to have an important market share to be useful. Without GNOME, KDE and other desktop tools the Linux community would have been much less numerous. And without the Linux community there wouldn’t be ~50% of severs and ~75% of smartphones running under Linux for example. I mean having a 20% market share is not necessarily an end because the development of Linux desktop has indirect positive impacts. Moreover Linux and its ecosystem are not built around one single company and cannot be marketed in the same way Windows and OS X are, even if Canonical thinks so.

  • Vincent O’Neil

    This is so pathetic. Trying to slander a genuine attempt by one of the leading Linux companies to innovate in a new direction. I suppose Canonical, a company, needs permission from the Linux community before they make a decision? So unrealistic and ridiculous it’s no wonder Linux is being held back. I wonder where people get off being so self-righteous and presumptive without knowing and understanding the entire issue. It makes me wonder if anyone was giving Apple this much crap when they implemented Quartz.

    Yet everyone is quick to jump to conclusions before the thing is even released. I know that success tends to breed jealousy and hatred but the amount of vehement venom being spit upon an unreleased project in its infancy by the Linux community is disgusting.

    GNOME is free to do as they please and it is understandable that they’ll support Wayland. Most distributions will probably choose Wayland as well if the reactionary feedback is to be taken seriously. If most of the distros choose Wayland and only Ubuntu uses Mir then the “fragmentation” done by Ubuntu means next to nothing.

    • Rowan Lewis

      “I wonder where people get off being so self-righteous and presumptive without knowing and understanding the entire issue.”

      Pot, meet kettle. I don’t think you have a clue what’s going on.

      • Vincent O’Neil

        Then educate me oh wise and omniscient One.

    • hashem

      The point of “slandering” Canonical is not because they’re trying to innovate, but rather because they’re being antagonistic to the larger Free software community, which has been moving toward Wayland with the assumption that all the major players were on board.

      This kind of solitary “innovation” serves to fragment the Free software community, waste work, and undermines one of the major advantages we have over proprietary software — the ability to share code.

      • Vincent O’Neil

        I don’t see the problem here. Don’t like Mir? Don’t support it. Lack of apps and community support will kill it and, by extension, Ubuntu. Then you can go support whatever distribution you think will have a decent chance of doing what Canonical was trying to accomplish for Linux.

        • Rowan Lewis

          You not seeing the problem, is well, your own problem. Because it’s been well written about by now. Are there any particular points you’re not sure of?

          Edit: Not that I think you’re wrong that in the long run it may only hurt Ubuntu, it’s just that in the meantime we’ve got to deal with these issues.

  • martial

    One day, only Unity will work on Ubuntu, unless Ubuntu and other continue support of X. Once they have migrated to Wayland, Gnome Shell and other DE won’t work anymore in Mir.

    One Day, Ubuntu will be like Android : a linux base but with a lot of in-house applications making it incompatible with other distributions

  • Pingback: Revista Espírito Livre » Blog Archive » GNOME aposta no Wayland como servidor gráfico()

  • Sepehr Aryani

    I see a day when some Gnome community members start patching Gnome Shell to make it work under Mir, that day I will laugh out LOAD!
    It seems that any changes to Wayland depends on some sorts of Canonical’s announcements, like the attention it got back in 2010 when Canonical announced moving to Wayland and once again when they announced they are moving to Mir.

    With this pure community driven vision I don’t see Wayland getting anywhere. Don’t get me wrong, I’m not against Wayland, but I’m sick and tired of all WAPOR_WARES in FOSS communities. Yes, Commercializing Linux is not all that bad, Linux is already a successfull server platform, because it was well commercialized by companies like Red Hat and others.

    • Rowan Lewis

      Maybe you could try reading, because I’m pretty damn sure you’ve got all of the things you’re talking about wrong.

      • Sepehr Aryani


  • Pingback: openSUSE: Ubuntu Killer? | The Linux Action Show! | Jupiter Broadcasting()