Announcement

Collapse
No announcement yet.

Wayland's Current Release Manager Is Stepping Down, Following Samsung's Open-Source Drama

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

  • #61
    Originally posted by MrEcho View Post
    IMO I think the project tried to make everyone happy, and play nice with existing systems, Gnome / KDE. In doing so they pretty much tried to make the impossible.

    The only way I see a new GFX / Desktop GUI for Linux is to do it all in one. Server + Client + WM / Interface
    Really rethink the whole thing. With a lot of the GFX Cards now supported in the Kernel and or Mesa, it should make it easier. Also there is no need to do networking with a desktop, its pointless now days, the GUI is going to be on the same box as the GFX hardware, and if you need remote theres things for that. So that will solve a lot of the protocol issues.

    I think doing your own WM / Interface that no one would support out the gate would be an issue, and you would have to make some api wrappers to kind of get things kick started. But at the end I think a lot of apps would switch from Gnome and KDE to the new system. If you look at the eco system right now, some Gnome and KDE apps require you do load a whole crap load of other sub systems for them to work, so even right now its a mess.

    It would be even more work then what Wayland/Weston has done so far, but going into it as a full solution would make things easier. You could build a Desktop system that is still super flexible, but there would be core systems that you could not load a replacement. I look at OSX and Windows and what they provide for a desktop experience, yes both have general issues, but theres one way to do an desktop application, or theres one way to do a systray, or one way to show active applications. (yes I know windows is in a transition phase). I know the general Linux community hates 'one way' things, ex systemD. But I think its time to start thinking of doing the same thing for the Linux Desktop
    i Completely disagree, network transparency is a necessary feature and given how ubiqitious it has become, people want to benefit from the flexibility of being able to display apps from one device to another. With all of the mobile devices, 5G your statement that networking doesnt matter any more is outrageous.

    The big problem with Wayland is it was a big waste of effort time and resources, given that we already have X. It was a lot of wheel reinventing, that as I and others have predicted, Wayland is far more bloated and wasteful than X was.

    I also completely disagree with your horrendous idea of a Window Manager getting involved with system trays. XDG and Freedesktop have been in fact developing API standards so that there is a standard way to register icons for the system tray and other shared resources that covers the need for this. The desktop environments who render the system tray then support that API. I fully support development of APIs that many different desktop environments can support, that makes sense.

    People claim X is bitrotting. Bitrot is a nonsense term sense code continues to work, it does not decay like an apple will on a table. They say people are not looking into the X code. Thats because they have been wasting all of the time on the clusterfunk and black hole that is Wayland.

    We do need some work on X such as to ensure the Glamor project continues and becomes ubiqitious. Apps on X need to use the vertical sync and double buffer extensions which would provide the same thing which Wayland was allegedly about, defect free rendering, but thats minor compared to the Wayland monstrosity.

    Scrap Wayland, Keep X.

    Comment


    • #62
      We always see all of these bad ideas crop up again and again. Wayland is a bad idea and a result of some of this. People who believe they can outthink and are wiser than the X way will end up making something quite inferior. The paradigms of X were well thought out by wise people and people who question them are doomed to produce very poorly designed window systems.

      The window manager/window system seperation has prevented the emergence of hundreds of completely incompatable window systems. If look and feel was built into the window system then everyone would have made their own incompatible window system. By staying out of look and feel and leaving that to window managers, people can create their own window manager and we do not end up with hundreds of incompatible window systems and we have ensured that applications can work well with all of the window systems so we dont end up with fragmentation of the app ecosystem.

      Window system staying out of look and feel has encouraged everyone to use the same standard code base which increases the amount of resources developing one code base rather than end up with spreading developer resources over hundreds of code bases for a bunch of incompatible window systems.

      Window manager code tends to add more complexity and potential for bugs, so keeping it seperate from window system, protects the window system should a bug occur and allows the window manager to be easily restarted.

      Window manager/window system seperation are vital to adapting to different form factors since different form factors are optimal with different look and feel models, you want a different Model for a smartphone than you would want on the desktop, and people have often learned that the hard way. When people use a desktop they want a taskbar and icons, phone APIs are optimized for that platform and its limitations but do not take advantage of what a desktop can offer.

      The diversity of window managers is what makes Linux great and allows people to tailor there environment to what suits them best. There is no single user interface that fits everyone.

      Network transparency also makes perfect sense since especially now in the 5G world, there is more networking than ever and being able to use a desktop app from a phone device and so on opens up the possibilities. Network transparency can be done easily and without any additional cost, people who think it takes a lot for network transparency are obviously not much involved with the programming because it does not take additional effort to enable network transparency and does not add performance overhead of the display and app are on the same machine.

      Comment


      • #63
        I think this probably has a lot to do with Samsung's terrible earnings report that just came about. I've watched as Wayland/Weston news items came and went, but knew that for me and most others switching from X.org wasn't possible anytime soon. I won't pretend to understand how Gnome/KDE's stacks interface with X, and I'm sure this is abstracted/transparent even to most devs.
        If there's anything that needs major cooperation and oversight, it's the task of replacing X. I know desktop/PC Linux users aren't priority in the big picture, but it seems like there would be major players invested in the future of Wayland. Even the BSD's are tied to it's future.

        Comment

        Working...
        X