Announcement

Collapse
No announcement yet.

Canonical Is Hiring More Mir, Unity Developers

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #31
    Originally posted by dh04000 View Post
    How dare they make Mir when Wayland was already around.

    How dare they make PulseAudio when Alsa was already around.

    How dare they make Gnome, when KDE was already around.
    First, just because they can do something, doesn't mean they should.

    Second, just because they can do something doesn't mean everyone has to agree. They have their right to make their choices, but we also have a right to judge them based on those choices. Or do you really think that neither Gnome nor PulseAudio devs got any criticism?

    Third, there is a big difference between something like a sound API or desktop environment, where applications written for different ones can happily co-exist, and a display server or protocal, where they can't. This is much less like making a new PulseAudio or new Gnome and much more like making a new Linux kernel, where libraries and applications written for the Linux kernel will not necessarily run at all on the Ubuntu kernel and vice versus.

    Comment


    • #32
      Originally posted by tuuker View Post
      Future needs one display system to all devices and that is Mir. Until Mir is open source then everything is fine. X.Org is stoneage and should be terminated.

      I am glad they are using pure c++ and Cmake system.
      LOL. Xorg still does whatever u need, so why all the hate?

      Comment


      • #33
        Originally posted by oleid View Post
        server side buffer allocation
        And that was the last feature I didn't know if Wayland offered.

        Comment


        • #34
          Originally posted by TheBlackCat View Post
          First, just because they can do something, doesn't mean they should.
          Your talking to a research scientist. This very concept is frankly offensive to people like me.

          Comment


          • #35
            Originally posted by dh04000 View Post
            Your talking to a research scientist. This very concept is frankly offensive to people like me.
            I am a research scientist, too, and the fact that there are scientists like you who feel that way scares the hell out of me. In fact I was under the apparently naive impression that this was a strawman used by anti-science types to attack scientists. I guess we still need animal and human research ethical review boards after all.

            Comment


            • #36
              ok here is the reason fragmentation is bad in such an important part of the stack

              1.) Neither Qt5 nor GTK+ 3.X nor EFL libraries will support Mir, this translate in no KDE/GNOME/EFL application will work with Mir in Ubuntu for a while unless canonical take the task to support out of tree all Toolkits, in simple english gimp, thunderbird,brasero, calligra, chromium,libreoffice,etc in wayland no problem but in ubuntu say bye bye unless canonical port them themselves
              2.) Mesa will not support Mir officially[same reason as KDE], so there is a big chance in the future canonical will have to fork mesa and maintain it themselves
              3.) Nor intel nor AMD nor nouveau will have Mir in consideration when making changes to their drivers beyond X11 and wayland, so if breaks feel free to fork canonical
              4.) Wine and SDL will not support Mir either, so say bye bye to games unless canonical port them themselves
              5.) RedHat will support wayland exclusively meaning all workstation apps like Catia or Oracle or Maya, etc won't work on ubuntu unless canonical make deals themselves and port them
              6.) and i can keep all day long

              so canonical with Mir have 2 choices, they either break all relation with linux native apps and make their own API for their store[the mac way] or they tackle the massively huge task of porting every app in their repos and PPA to Mir, as a possible 3rd choice could be ubuntu desktop will start Xmir by default and keep until apps stop supporting X11 and Mir native apps[having all the goodies] will only exist for mobile versions but then come the question, why Mir in the desktop then? i mean if the entire desktop will stay using legacy X11 what is the point of keep Mir in ubuntu X86? since Qt5 works on both you can't support wayland on x86 and Mir in mobile and save you millions of LOC having the same result?

              Comment


              • #37
                Originally posted by jrch2k8 View Post
                1.) Neither Qt5 nor GTK+ 3.X nor EFL libraries will support Mir, this translate in no KDE/GNOME/EFL application will work with Mir in Ubuntu for a while unless canonical take the task to support out of tree all Toolkits, in simple english gimp, thunderbird,brasero, calligra, chromium,libreoffice,etc in wayland no problem but in ubuntu say bye bye unless canonical port them themselves
                I believe Canonical said they were going to port Qt themselves. Don't remember the version, though.

                Comment


                • #38
                  Originally posted by mrugiero View Post
                  I believe Canonical said they were going to port Qt themselves. Don't remember the version, though.
                  well they have some skeleton code[kinda partially work] of a lighthouse module for Qt5.0[for 5.1 and 5.2 will need many adjustments tho], as for GTK or SDL2 or EFL or Wine i can't find anything which led me to think ubuntu will have QML applications only through ubuntu store

                  Comment


                  • #39
                    Originally posted by jrch2k8 View Post
                    ok here is the reason fragmentation is bad in such an important part of the stack

                    1.) Neither Qt5 nor GTK+ 3.X nor EFL libraries will support Mir, this translate in no KDE/GNOME/EFL application will work with Mir in Ubuntu for a while unless canonical take the task to support out of tree all Toolkits, in simple english gimp, thunderbird,brasero, calligra, chromium,libreoffice,etc in wayland no problem but in ubuntu say bye bye unless canonical port them themselves
                    2.) Mesa will not support Mir officially[same reason as KDE], so there is a big chance in the future canonical will have to fork mesa and maintain it themselves
                    3.) Nor intel nor AMD nor nouveau will have Mir in consideration when making changes to their drivers beyond X11 and wayland, so if breaks feel free to fork canonical
                    4.) Wine and SDL will not support Mir either, so say bye bye to games unless canonical port them themselves
                    5.) RedHat will support wayland exclusively meaning all workstation apps like Catia or Oracle or Maya, etc won't work on ubuntu unless canonical make deals themselves and port them
                    6.) and i can keep all day long

                    so canonical with Mir have 2 choices, they either break all relation with linux native apps and make their own API for their store[the mac way] or they tackle the massively huge task of porting every app in their repos and PPA to Mir, as a possible 3rd choice could be ubuntu desktop will start Xmir by default and keep until apps stop supporting X11 and Mir native apps[having all the goodies] will only exist for mobile versions but then come the question, why Mir in the desktop then? i mean if the entire desktop will stay using legacy X11 what is the point of keep Mir in ubuntu X86? since Qt5 works on both you can't support wayland on x86 and Mir in mobile and save you millions of LOC having the same result?
                    1 ) Fail : Canonical work in a Qt and Gtk backend
                    2 ) They have some patch for mesa
                    3 ) Nouveau, Intel and Radeon work yet on Mir and Wayland
                    4 ) Lol they support it because The dev of SDL work for Valve and Valve support Ubuntu
                    5 ) bullshit again
                    6 ) Yes saying nothing interesting

                    Reason to have more than one project :
                    1 ) No more application depending directly for a display server.
                    2 ) If a project fail we have another option
                    3 ) People can choose what is better depending what are their needs
                    To resume : all the avantage offer by the free software phylosophy... Not from the proprietary philosophy...

                    Comment


                    • #40
                      Originally posted by seb24 View Post
                      1 ) Fail : Canonical work in a Qt and Gtk backend
                      Fail? Point it out. It clearly says 'except they do the porting themselves.
                      2 ) They have some patch for mesa
                      Same, mesa will not support Mir. Mesa is upstream project, not Canonical in-house patches.
                      3 ) Nouveau, Intel and Radeon work yet on Mir and Wayland
                      This all depends on mesa working, and again, local patches to mesa are not upstream support. This means, breakages that Canonical will have to fix.
                      4 ) Lol they support it because The dev of SDL work for Valve and Valve support Ubuntu
                      That's a hard assumption. For a start, there's no warrant this dev will keep being the maintainer of SDL, specially having another full-time job. Note Valve never took a position in the Wayland-Mir conflict. As far as anyone outside Valve knows, they'll stick to X.
                      5 ) bullshit again
                      Can you point any link saying Red Hat is thinking about supporting Mir? Wayland sounds really likely (and if you take the freedom of assuming Valve is onboard of Mir when they didn't even paid their maintainer, and then make SDL maintainer port it to Mir, I can take my own to assume Red Hat will support Wayland, which is a far simpler one), since the one that came with the idea of Wayland is their employee.

                      1 ) No more application depending directly for a display server.
                      That's not a reason, that's a fix. We have to support several display servers, thus we don't target them directly.
                      2 ) If a project fail we have another option
                      If a project fails, we can fork it and fix it. That's free software about, after all.
                      3 ) People can choose what is better depending what are their needs
                      For this, Mir should actually cover needs not covered by Wayland.

                      Comment


                      • #41
                        Originally posted by seb24 View Post
                        1 ) Fail : Canonical work in a Qt and Gtk backend
                        2 ) They have some patch for mesa
                        3 ) Nouveau, Intel and Radeon work yet on Mir and Wayland
                        4 ) Lol they support it because The dev of SDL work for Valve and Valve support Ubuntu
                        5 ) bullshit again
                        6 ) Yes saying nothing interesting

                        Reason to have more than one project :
                        1 ) No more application depending directly for a display server.
                        2 ) If a project fail we have another option
                        3 ) People can choose what is better depending what are their needs
                        To resume : all the avantage offer by the free software phylosophy... Not from the proprietary philosophy...
                        1.) the qt backed barely working is there, show me the bazaar repo for gtk cuz i can't find it
                        2.) no, Mir still can work with noveau,intel and amd in mesa but if tomorrow wayland set EGL 1.1 and ask to remove 1.0 wayland will have support the same day while canonical will have to fork to maintain the previous version until they migrate for example, not official means nobody cares if a change break Mir <-- that is my point
                        3.) SDL2 is in code freeze and only have wayland backend official and Valve have not announced support for Mir nor SDL main developer has either, so in best case scenario SDL will not support Mir at release time because the 2.1 branch have 0 Mir code, so if such code exist at all will land in 2.2 that is far away
                        4.) most enterprise application give higher priority to RHEL, you can call support yourself and certify it. im not saying right are impossible to install in ubuntu but nobody pays that ludicrous amount of cash for a software to install it outside redhat with gold 365 support plan, unless you are those that live in the Oz land that believe anyone win money with ubuntu and redhat evil eyed reviewers to be in the top fortune 500 list of companies
                        5.) wine use its own render api base on GDI and that need to be ported too, is unrelated to SDL

                        Comment


                        • #42
                          Originally posted by BO$$ View Post
                          Is it really that hard to believe that Wayland isn't the greatest piece of software in the world and that others consider that they can do better?
                          As long as no one can provide any valid reasons why Mir is needed, yes.

                          Canonical has provided not even one valid reason why they couldn't use Wayland, or why it's necessary to create Mir. Vague PR doublespeak about "unforeseen problems" and "not fitting our business plan" or whatever do not count.

                          The justifications of Ubuntu fanboys/girls are even more pathetic. They all go along the lines of:
                          - "it's good to have alternatives" (no, there's no inherent value in alternatives if there's no valid reason for them to exist and good reasons for them not to)
                          - "there are already X number of package managers/sound servers/desktops/file systems/init systems, so why not display servers" (because the graphics stack is not comparable to any of those things, and compatibility is crucial in this area)
                          - "Mir is good because Wayland started developing faster after it was announced" (a - not true, and b - if the only benefit of Mir is the supposed speeding up of Wayland, then logically it'd be even better to forget Mir and help with the development of Wayland!)
                          - "everyone is just jealous of Canonical because they're the bestest and neatest Linux" (stupid, juvenile, fallacious argument that tries to distract from the real issue)
                          - "Mir needed to be done because Wayland was taking too long" (yeah, real smart... Wayland is taking too long, so let's not contribute to Wayland development, instead, let's start a NEW project...)

                          etc. etc.

                          I'm looking forward to the day we can put all this Mir nonsense behind us and all concentrate on Wayland.

                          Comment


                          • #43
                            Originally posted by dee. View Post
                            As long as no one can provide any valid reasons why Mir is needed, yes.

                            Canonical has provided not even one valid reason why they couldn't use Wayland, or why it's necessary to create Mir. Vague PR doublespeak about "unforeseen problems" and "not fitting our business plan" or whatever do not count.
                            To be fair there is one real difference between Mir and Wayland: Mir is using test-driven development.

                            Of course they haven't pointed out any flaws in Wayland that test-driven development would have corrected, nor is there any reason they couldn't have made their test-driven version API compatible at the very least, nor have they provided any reason to think that test-driven development can compensate for rushing a central part of the software stack (poorly-thought-out tests are not much better than poorly-though-out code), nor is there any reason to think that test-driven development can compensate for not having anyone on the team who has any familiarity with this sort of project (you can't write good tests if you don't know what to test), nor has it even been established conclusively that test-driven development is really superior.

                            But it nevertheless is a real difference. As you probably guess, though, I don't think this is a valid reason to start over from scratch.

                            Comment


                            • #44
                              Originally posted by TheBlackCat View Post
                              To be fair there is one real difference between Mir and Wayland: Mir is using test-driven development.

                              Of course they haven't pointed out any flaws in Wayland that test-driven development would have corrected, nor is there any reason they couldn't have made their test-driven version API compatible at the very least, nor have they provided any reason to think that test-driven development can compensate for rushing a central part of the software stack (poorly-thought-out tests are not much better than poorly-though-out code), nor is there any reason to think that test-driven development can compensate for not having anyone on the team who has any familiarity with this sort of project (you can't write good tests if you don't know what to test), nor has it even been established conclusively that test-driven development is really superior.

                              But it nevertheless is a real difference. As you probably guess, though, I don't think this is a valid reason to start over from scratch.
                              Also, test driven development is something different only at the implementation level. Not only this means there is no reason to fork from the protocol (they could still do their own implementation, which wouldn't mess anything), but this technique could be applied from now on if desired. Write the tests, and continue the development as test driven, and problem solved.

                              Comment


                              • #45
                                Originally posted by TheBlackCat View Post
                                I am a research scientist, too, and the fact that there are scientists like you who feel that way scares the hell out of me. In fact I was under the apparently naive impression that this was a strawman used by anti-science types to attack scientists. I guess we still need animal and human research ethical review boards after all.
                                I meant along the lines of its ok to reinvent the wheel and come up of new tools that do the samething. Sometimes the new tool is better in the different situtions than the first tool. Even better, we learn that the first tool was biased, thus we learn the truth. Same concept with more than one way to approach or solve a problem. Reducancy is important sometimes.

                                Comment

                                Working...
                                X