Announcement

Collapse
No announcement yet.

It's Time To Admit It: The X.Org Server Is Abandonware

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

  • Originally posted by mdedetrich View Post

    Users dictate what happens, not developers. If users and developers are not aligned then users will abandon the platform or developers will be forced to accommodate.
    Users do not get to dictate anything if they are not paying for it.

    Want something different, learn the skills of the trade and make it happen. Otherwise shut up and use what you are given.

    Comment


    • Originally posted by Sonadow View Post

      Users do not get to dictate anything if they are not paying for it.

      Want something different, learn the skills of the trade and make it happen. Otherwise shut up and use what you are given.
      Wrong, Linux is only as supported as how many users it has. Linux without any users will turn into a joke and become abondonware.

      The reason why Linux is where it is now is because it supports a lot of hardware, more hardware support means more users use Linux and more users means that Linux community gets more feedback on what works and what doesn't (talking about Linux desktop here).

      Linux developers do not have every single hardware configuration, so they need users for this reason.

      Comment


      • Originally posted by mdedetrich View Post

        Wrong, Linux is only as supported as how many users it has. Linux without any users will turn into a joke and become abondonware.

        The reason why Linux is where it is now is because it supports a lot of hardware, more hardware support means more users use Linux and more users means that Linux community gets more feedback on what works and what doesn't (talking about Linux desktop here).

        Linux developers do not have every single hardware configuration, so they need users for this reason.
        Bull. Linux works on a large variety of hardware because large OEMs like Dell, Lenovo, HP etc and companies like Red Hat, Broadcom, Qualcomm, Nvidia, Intel, AMD etc plough in lots of money to have their developers and engineers contribute upstream patches in the kernel for hardware support, or work downstream to provide a driver blob or in ARM's case, a custom kernel with all kinds of patches, proprietary bits and DTBs preloaded so that it even boots.

        It's money, money and always about money. Linux doesn't need a desktop user like you to be successful, and most of its success comes in headless server deployments where there is no such thing as X.

        The average user is nothing to Linux.
        Last edited by Sonadow; 26 October 2020, 05:02 AM.

        Comment


        • Originally posted by Sonadow View Post

          Bull. Linux works on a large variety of hardware because companies like Broadcom, Qualcomm, Nvidia, Intel, AMD etc plough in lots of money to have their developers and engineers contribute upstream patches in the kernel for hardware support, or work downstream to provide a driver blob.
          Your comment is bull. The only 2 companies which you listed that are still contributing to the Linux kernel in a relevant manner are AMD and Intel. Broadcom/Qualcomm supplied a driver ages ago which the Linux community had to work to keep it updated. Also note that Qualcomm's wireless drivers for android are proprietary.

          Nvidia only added support for Tegra to the Linux kernel, which is now dead apart since no Linux phone is really using Tegra (I think Nintendo Switch is the only mainstream device that uses Tegra).

          You completely missed the point, which is that those companies are users which happen to be developers, i.e. they use Linux for their own goals. They also happen to only contribute to Linux in the areas they care about, i.e. companies like Microsoft are one of the biggest contributes but its for windows virtualization due to Azure.

          Also I am talking in context of Linux desktop, so the only 2 relevant companies you listed are Intel and AMD which doesn't cover a lot of hardware

          Originally posted by Sonadow View Post
          It's money, money and always about money. Linux doesn't need a desktop user like you to be successful, and most of its success comes in headless server deployments where there is no such thing as X.
          Also wrong because its reductionist.

          Originally posted by Sonadow View Post
          The average user is nothing to Linux.
          Yeah all of the free testing done by users is nothing Also do note that just because companies contribute code doesn't mean it works in all cases (and this has historically been a fact). There is also something called regression testing.
          Last edited by mdedetrich; 26 October 2020, 05:09 AM.

          Comment


          • Originally posted by Mark Rose View Post

            KDE is hardly an experiment. It's been around longer than GNOME.
            And? Experiments can take even decades!

            I knew many KDE developers, most of them abandoned KDE. They got bored of that "experimental" mindset and relying on a third party toolkit with too much corporate instability.

            I like lots KDE features and consider GNOME lacking too much features in comparison, but it needs to be a lot more stable and focus on certain stuff that some forks (such as kwin-ft) try to fix separately.

            Comment


            • Originally posted by timofonic View Post

              I knew many KDE developers, most of them abandoned KDE. They got bored of that "experimental" mindset and relying on a third party toolkit with too much corporate instability.
              I would be careful about making judgements based on a few developers. Using that same reasoning Gnome is also experimental (actually considering how many things Gnome has broken historically and how many times they have changed their vision of "desktop" I would consider gnome more experimental).

              The fact that a huge number of people are using KDE without problems means that its stable (which is the opposite of experimental).

              Comment


              • Originally posted by mdedetrich View Post
                With performance implications? Yeah no thanks.
                On platforms that Xwayland works
                https://github.com/Plagman/gamescope
                As what was found with gamescope. Increased performance compared to bare metal X11.

                It's getting game frames through Wayland by way of Xwayland, so there's no copy within X itself before it gets the frame.

                Xwayland bypasses the X11 server compositor layers. So there are less buffers to go though before the output gets to screen so better and more stable performace,

                Originally posted by mdedetrich View Post
                The reason why I mentioned Valve is because of steam. Currently the support for games is much better via X11 (also taking into account performance) compared to Wayland.
                This statement is wrong. If you have Intel and AMD graphics your best performance for your games is under Wayland using Xwayland if the applications are X11 only.

                Originally posted by mdedetrich View Post
                Furthermore NVidia GPU's are the dominant GPU's for gaming (this is a fact, check steam hardware survey) and Wayland support here is lacking (to put it diplomatically).
                This is why Valve has been working with Redhat developers over GLX Delay.

                Originally posted by mdedetrich View Post
                This is just reality, Valve has invested into Linux but they also want to make sure that their platform (Steam + Proton) works on Linux.
                Yes it true that Valve is invested into Linux and they want Steam + Proton to work. The problem you have missed to run legacy games valve needs something like gamescope to work well. Like you have a 32 bit X11 graphics server and you have a 8 bit colour game and it now need nice resolution scaling as well

                This is a case that Nvidia is not listerning to one of their biggest pools of users. Valve is saying over and over again they want Wayland. One way Valve could force their hand it take over maintenance of x.org X11 and drop all support for X11 server running bare metal so all output is by Xwayland this would force Nvidia hand..


                Originally posted by mdedetrich View Post
                Also the whole point is to reduce effort from the Linux community, right now KDE/Gnome/Sway are all duplicating the effort which is unnecessary and stupid.
                KDE and Gnome both use libwayland-server from Weston. Sway uses wlroots. There is not as much duplicated effort as it first appears.

                Originally posted by mdedetrich View Post
                The alternative is for Valve to ditch Linux and redhat to reduce its commitment to desktop because simply put Wayland doesn't work with the hardware/feature set that X11 did.
                This totally ignores Valve problem space. Valve has a pool of games they are licensed to sell. Valve wants to be able to sell as many of those games as possible that will work for users without users knowing tweaks have been done. X11 lack of ability to cleanly abstract between hardware and application reduces the number of games Valve can effectively sell to Linux users.

                Originally posted by mdedetrich View Post
                Users dictate what happens, not developers. If users and developers are not aligned then users will abandon the platform or developers will be forced to accommodate.
                Its not users who absolutely dictate what happens. Valve case its about how can Valve make the most money sorry X11 bare metal does not work for Valve, The one big thing you missed that important is users using steam with AMD or Intel on Linux can install and use more games from Steam than the Nvidia users due to the limitations of X11.

                Sooner Nvidia accepts the writing on the wall and support wayland and xwayland the better. Either Nvidia support Wayland and Xwayland or Redhat with Valve will make it work anyhow because that is where the profit is.

                Why would Redhat for servers be interested in getting Xwayland working with Nvidia by any means required. Simple this will allow spitting low consumer cards between multi virtual machines. Redhat interest is not desktop linked and is also about money.

                Comment


                • pal666 you're funny

                  Comment


                  • (Situation: there are too many comments in this thread looking at the whole X vs Wayland situation in a completely wrong way. That's ridiculous! I need to write one universal comment that covers everyone's viewpoints.)

                    First of all, X11 is old and deprecated, period. It's by no means useless or broken yet, and it won't be going away any time soon, but there's a reason it is currently in an unmaintained (and, going by its developer team's comments, unmaintainable) state. And no, you don't get to call an unmaintained and unmaintainable software "stable" on merit of it not being completely broken yet, so please stop doing that. The only reason we're still using X11 is because its replacement, aka X12, aka Wayland, isn't ready yet - and it might never be "ready" the way we would want it to be; which leads me to my next point.

                    The second elephant in the room is that unfortunately, with Wayland we do not have a cross-{desktop,distro,OS} display server anymore like we used to have with X11, i.e. one that can be used and shared by all interested parties. Instead, AFAIK, we have this weird situation where the former X11 team got mostly integrated into GNOME/Red Hat, and thus the new version of X (because that's what Wayland really is) has thrown all DRY sensibilities out of the window and has essentially become GNOME's own display server, with every other X user (KDE, the BSDs, et al) left dead in the water and trying to play catch-up with their own Wayland implementations. Yeah, there's the generic Wayland protocol. Yeah, there's Weston. But in reality, nowadays the "official" X display server is mostly a GNOME project and the "official" X developers are mostly Red Hat's/IBM's employees, so for the most part their work only benefits GNOME - or, as some of them like to insinuate, and as the GNOME fanboys like to openly declare, the "default Linux DE".

                    This lack of cross-{desktop,distro,OS} interoperability would already be problematic on its own but, to add insult to injury, by integrating with GNOME the X team has also inherited all the known "qualities" of the GNOME development paradigm, i.e. they only care for their own use cases, they dismiss open standards if they don't intend to make direct use of them (e.g. server side decorations) and/or implement custom, GNOME-specific replacements for them, they make arbitrary decisions behind closed doors, and they tell people who "whine" about all that to stop "whining" and take a hike.

                    Fortunately, there's also some good news: lately, this GNOME mentality seems to be shifting for the better (they've actually been addressing a number of papercuts that until recently they used to dismiss as wontfix) so we can only hope that the Wayland situation will improve as well. Still, for the time being, it's what it is.

                    Anyway, the real issue here is that most of you people are already aware of this stuff I just wrote, but instead of forming an objective opinion based on hard facts, you instead choose to ignore some of those facts and pick a side in this endless and fruitless debate. Why is that? Well, as far as I can tell:

                    - Those of you who are understandably fed up with Wayland as a protocol/design; or with its progress as a software and its adoption by third parties; or with GNOME's/Red Hat's/144Hz's elitist attitude; or with clueless people and/or GNOME shills declaring Wayland as ready when it's clearly not even close; you choose to blissfully ignore the sad state of X11 and proclaim it as "stable" and "state of the art" software, when even my dear ole grandma could tell it's really anything but.

                    - Those of you who are fed up with Wayland naysayers who proclaim it as useless; or who constantly bash GNOME's/Red Hat's and other corporations' work; or who behave like OSS zealots and casually dismiss companies like e.g. Nvidia or initiatives like e.g. systemd with seemingly no regard as to the need/wish/desire for Linux to at last grow up and become a valid OS choice for the masses; or who stubbornly keep trying to describe as "stable" a crap software like X11 that even its own developers dismiss as crap; you choose to ignore the simple fact that at this moment in time Wayland is indeed a F*CKING MESS, and that GNOME's hegemony of the project (instead of it being a neutral, shared library like X11) is anything but good and will only hurt Linux in the near future.

                    So, there you have it. And because this is already a very long comment and I'm getting tired of writing, I'll not bother waxing philosophical here and I'll end it with a simple TL;DR that will have to do as an epilogue.

                    TL;DR: X11 is not at all "stable" but rather a f*cking mess. Wayland is not at all "ready" but rather a f*cking mess. But Wayland is slightly less of a f*cking mess than X11 because it's more modern and it's being actively developed, while X11 is all but declared dead by its developers and is slowly but steadily being deprecated across the board in favor of Wayland. So everybody, on both sides, please stop f*cking around and admit the hard truth, no matter if you like that truth or not. Also: alas, woe is me! etc etc. But really, things are not as bleak as they may seem to some people, and there's more than a good chance that, God/Nvidia/GNOME willing, within a couple of years or so everything will have just clicked neatly into place. Just have a little patience and in the meantime please don't feed the trolls.

                    (Situation: there are now (too many comments + 1) in this thread. Oh well, at least my comment is Rightâ„¢, even though it's also so long that nobody will bother to read it :P)

                    Comment


                    • Originally posted by oiaohm View Post

                      On platforms that Xwayland works
                      https://github.com/Plagman/gamescope
                      As what was found with gamescope. Increased performance compared to bare metal X11.

                      It's getting game frames through Wayland by way of Xwayland, so there's no copy within X itself before it gets the frame.

                      Xwayland bypasses the X11 server compositor layers. So there are less buffers to go though before the output gets to screen so better and more stable performace
                      Its not just about buffers, but also about exclusive full screen (which still doesn't properly work on all wayland's implementations). There are other issues as well which I don't fully remember since I stopped using Wayland due to such issues.

                      Originally posted by oiaohm View Post
                      This statement is wrong. If you have Intel and AMD graphics your best performance for your games is under Wayland using Xwayland if the applications are X11 only.
                      I am not talking about just Intel+AMD. The whole point is that Intel+AMD is not the only hardware out there, there is also NVidia and if we are talking about gamers its actually the minority that use AMD.

                      Originally posted by oiaohm View Post
                      This is a case that Nvidia is not listerning to one of their biggest pools of users. Valve is saying over and over again they want Wayland. One way Valve could force their hand it take over maintenance of x.org X11 and drop all support for X11 server running bare metal so all output is by Xwayland this would force Nvidia hand..
                      Which is irrelevant to Linux gamers that have NVidia hardware. They care about the system working with their hardware not about internal Linux politics, so as long as this is the case the X11 won't be abandoned in reality (if issues need to be fixed). It will just be forked and maintained.


                      Originally posted by oiaohm View Post
                      KDE and Gnome both use libwayland-server from Weston. Sway uses wlroots. There is not as much duplicated effort as it first appears.
                      Interfacing with the hardware is what is duplicated since Wayland is just a protocol and no one uses the reference Wenston implementation.

                      Originally posted by oiaohm View Post
                      This totally ignores Valve problem space. Valve has a pool of games they are licensed to sell. Valve wants to be able to sell as many of those games as possible that will work for users without users knowing tweaks have been done. X11 lack of ability to cleanly abstract between hardware and application reduces the number of games Valve can effectively sell to Linux users.
                      Versus not being able to run Linux DE properly at all because you have an NVidia graphics card? I think that problem is much worse then the one you are describing

                      Originally posted by oiaohm View Post
                      Its not users who absolutely dictate what happens. Valve case its about how can Valve make the most money sorry X11 bare metal does not work for Valve, The one big thing you missed that important is users using steam with AMD or Intel on Linux can install and use more games from Steam than the Nvidia users due to the limitations of X11.

                      Sooner Nvidia accepts the writing on the wall and support wayland and xwayland the better. Either Nvidia support Wayland and Xwayland or Redhat with Valve will make it work anyhow because that is where the profit is.

                      Why would Redhat for servers be interested in getting Xwayland working with Nvidia by any means required. Simple this will allow spitting low consumer cards between multi virtual machines. Redhat interest is not desktop linked and is also about money.
                      Yes and thats the point, if Valve wants to make the most money they need the largest possible amount of people using Steam which means they need to support NVidia, thats the reality.

                      So this means that either the Linux community needs to mature a bit and actually work with NVidia in trying to get a solution working, e.g. EGLStreams (instead of just them figuratively and literally fuk off all of the time) or Valve needs to deal with the fact that a lot of its Linux users will be using X11 (and even currently they are using X11) and hence in the worst case they might be forced to support X11 if the status quo doesn't change and X11 needs fixes/updates (or they just ignore Linux, hence my previous point).

                      Ontop of this, certain distro's will not move to Wayland unless its up to the distro's standards and the most popular distro (Ubuntu) is sticking to X11 for NVidia hardware for obv reasons.
                      Last edited by mdedetrich; 26 October 2020, 06:45 AM.

                      Comment

                      Working...
                      X