Announcement

Collapse
No announcement yet.

Is Compiz On Its Deathbed?

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

  • #11
    IMHO the problem with compiz is that the code is bloated as hell and has become a nightmare to maintain.No wonder about it, the project has suffered from being the hostage of countless dev fights, a fork and a merging, then later another fork and a merging again!
    Think about this: The dev community has gone through so many fights that caused the forking of the project (beryl and compiz-fusion) but later they merged back. Merging projects doesn't mean the code between the forks will be all compatible with each other. And what's more important, merging doesn't mean the devs will stop confronting themselves so easily to work on the project.
    Seeing the project like that, is even a miracle that compiz is still alive today.

    We should remember that compiz is one of the oldest compositing window managers in the history of gnu/linux. And that has a price to pay. The linux (And more specifically, the mesagl) community has changed over the years and the compiz project has to adapt every time it sees it neccesary.

    The project needs to stop this madness of patching something made to be compatible with the now dead 2.4 kernel. More likeable, the project needs a full restart with both blobs and gallium3d in mind. There are desktops like lxde, xfce and mate which can still get a benefit on using compiz.

    Comment


    • #12
      Originally posted by Teho View Post
      Large projects like Ubuntu don't just jump from window manager to window manager...
      You'd think that except it doesn't seem like they make decisions based on technical merit. There original move from Mutter was supposedly based upon the an engineer's discovery that mutter had fundamental perfomsnce issues thus the move to the more "mature" compiz. Of course this decision has come back to haunt them. My guess is that the real reason they wanted to stay away from mutter was that they wanted to control the entire user facing codebase (Unity). If they had stayed with mutter they would currently have ogles2 support for free, and a very good codebase for a WM.
      As it stands now, I think moving to kwin is the logical step. It supports plenty of backends and is modular enough to provide the effects they need. However, I have doubts as to the long term ability of Ubuntu to create a viable DE. I just don't think they have the technical will (IMHO, this is b/c of Shuttleworth).

      Comment


      • #13
        Originally posted by liam View Post
        If they had stayed with mutter they would currently have ogles2 support for free, and a very good codebase for a WM.
        Mutter was forked for Cinnamon so I doubt things would have gone that smoothly; the control is definetly an important factor and communication between Gnome and Ubuntu doesn't seem to work at all...

        I do see moving to KWin as a best option from technical standpoint as well as deprecating Unity 3D in favour of Qt. If Canonical goal is to conquer all the possible formfactors they really should just go with Qt/KF as a base. But again I doubt that that's not going to happen.

        Comment


        • #14
          Compiz is not the original (linux) composite manager, Kwin was with KDE 3.2 in tendem with the first Xorg release (when XFree was king). That code have been dropped for KDE 3.8(Krash2) and rewrite for 3.9(KDE4 beta). It was based on a KDEified version of xcompmrg. Of course, this was before XGL/AIGLX/NVIDIA-TTP so the number of effect was limited to compositing, transparency, shadow and little more.

          Comment


          • #15
            Again? Another flame war inciting nonsense "article" to drive clicks and ad revenue?

            Comment


            • #16
              Yeah, I don't really see a problem with compiz. It's feature rich, more so than any other windowing manager or compositor.

              It doesn't really make sense to start over unless the code base is just unsustainable.

              Yes it has bugs, but only if you use obscure compiz plugins. It's been working great here along with VMware gallium drivers.

              What is a shame is how the various compositing managers conflict. I'd like to have a compiz backend along with cinnamon as the front end. Or mix XFCE's simple window manager along with compiz.

              Comment


              • #17
                Originally posted by LLStarks View Post
                Compiz has fallen hard from its Beryl and Fusion days.

                It used to be the face of the 3D Linux desktop.

                But we can't have it anymore since Gnome 3 and Unity make it impossible to use it.
                Harsh words for the software that put windows to shame for many years in the looks department.
                For some reason, the leading window managers have chosen to implement their own compositing engines. Inevitably, this means a much narrower audience for Compiz. From there, there's not a big stretch to predict Compiz developers loosing interest and the project stagnating. I'll always have fond memories of Compiz and my friends faces whem I showed them Linux can look 10x better than Windows.

                Comment


                • #18
                  Originally posted by Teho View Post
                  Mutter was forked for Cinnamon so I doubt things would have gone that smoothly; the control is definetly an important factor and communication between Gnome and Ubuntu doesn't seem to work at all...

                  I do see moving to KWin as a best option from technical standpoint as well as deprecating Unity 3D in favour of Qt. If Canonical goal is to conquer all the possible formfactors they really should just go with Qt/KF as a base. But again I doubt that that's not going to happen.

                  Cinnamon forked mutter b/c of small differences in opinion. They weren't serious technical reasons (http://forum.pinguyos.com/Thread-Cin...nome-3?page=13) according to that. So I don't think we can extrapolate much from that. The Mint guys aren't about to take on a full on FORK (as in major changes, rather than a moving set of patches with frequent upstream sync) of a WM.
                  I agree that Ubuntu would be best served by moving completely over to a QT base, but not b/c it CAN be used at multiple form factors, but simple b/c the codebase would be easier to manage. EFL seems to be the best choice for embedded devices (even ignoring the Tiago change, EFL is pretty dominant in that whole area), but its stack isn't as comprehensive as QT's, and I don't see Ubuntu as a company that can build infrastructure level services.
                  From 30 000 ft it seems as if Ubuntu is losing mindshare which was its biggest commodity. If they want to remain a strong desktop force they have to get that back (or at least have the perception that they have it back).

                  Comment


                  • #19
                    Originally posted by liam View Post
                    Cinnamon forked mutter b/c of small differences in opinion. They weren't serious technical reasons (http://forum.pinguyos.com/Thread-Cin...nome-3?page=13) according to that. So I don't think we can extrapolate much from that. The Mint guys aren't about to take on a full on FORK (as in major changes, rather than a moving set of patches with frequent upstream sync) of a WM.
                    I agree that Ubuntu would be best served by moving completely over to a QT base, but not b/c it CAN be used at multiple form factors, but simple b/c the codebase would be easier to manage. EFL seems to be the best choice for embedded devices (even ignoring the Tiago change, EFL is pretty dominant in that whole area), but its stack isn't as comprehensive as QT's, and I don't see Ubuntu as a company that can build infrastructure level services.
                    From 30 000 ft it seems as if Ubuntu is losing mindshare which was its biggest commodity. If they want to remain a strong desktop force they have to get that back (or at least have the perception that they have it back).
                    Ubuntu Netbook was based on EFL before it moved to Qt. It is why they given up and rewrote the thing.

                    Comment


                    • #20
                      Originally posted by liam View Post
                      You'd think that except it doesn't seem like they make decisions based on technical merit. There original move from Mutter was supposedly based upon the an engineer's discovery that mutter had fundamental perfomsnce issues thus the move to the more "mature" compiz. Of course this decision has come back to haunt them. My guess is that the real reason they wanted to stay away from mutter was that they wanted to control the entire user facing codebase (Unity). If they had stayed with mutter they would currently have ogles2 support for free, and a very good codebase for a WM.
                      As it stands now, I think moving to kwin is the logical step. It supports plenty of backends and is modular enough to provide the effects they need. However, I have doubts as to the long term ability of Ubuntu to create a viable DE. I just don't think they have the technical will (IMHO, this is b/c of Shuttleworth).
                      Agreed. Mutter was pretty slow in its early development, but ubuntu's decision to up and drop it was extremely shortsighted. Mutter has improved massively and gnome-shell runs super smooth (smoother than unity) on the machines I've ran it on. Mutter is also leaner and more modern. Unity really could have benefited if they stuck with it imo.

                      Compiz strength is in the tons of plugins it has, but unity doesn't use most of them anyway, and playing around with ccsm can break unity.

                      Comment

                      Working...
                      X