How long will Xorg live?

Now if Qt and Firefox/Chrome (who are RedHat loyalists) decided to stop future X11 development, FreeBSD is pretty much screwed. We have no control over kwin, mutter, etc.
I really don't see why. FreeBSD supports Wayland perfectly right now. Even bhyve can run VMs under Wayland natively, even without XWayland (something that Virtualbox cannot do, even under Linux).

I understand that people love to spread FUD about Wayland and X11 but, frankly, this is becoming ridiculous.
 
Besides, fast forward a few decades and far more people will be using Gtk with X11 than i.e Win32 and Cocoa ever have (due to having better alternatives) and those small communities have still managed to implement support.

I don't see that. A lot of various open source projects actually ditched gtk and ported their UIs to Qt. Win32/Cocoa exist in completely different worlds so they're irrelevant to the discussion of the open source desktop. A lot of engineering hours were put in to DWM and Quartz.

I think these are a good example. The sheer volume of very complex patches required to currently build these browsers, absolutely dwarfs what X11 needs for it to be supported inside Gtk. It shows that it will get done. Its not like we all need to support ALSA for our audio, "just coz Linux does" right? Why should graphics be any different?

That's fine and all. But relegating support for gtk2/X11 down to a single person and a small BSD project doesn't really all that sound promising. But hey if there's a will, there's a way right?

X.org works well, but the Linux community has beaten and battered its implementation also. It's just bad.

I really don't see why.

Of course you don't. Because you don't see the trajectory and movement of the issue. Until I saw what the OpenJDK, Firefox/Chrome and the toolkit folks were doing; I was somewhat sure X11 would stick around. Not so much anymore. The ecosystem is moving towards Wayland; this isn't just about X vs Y Window Manager or DE anymore.
 
I don't see that. A lot of various open source projects actually ditched gtk and ported their UIs to Qt.
Sure, but the same discussion stands for Qt as Gtk.
Win32/Cocoa exist in completely different worlds so they're irrelevant to the discussion of the open source desktop. A lot of engineering hours were put in to DWM and Quartz.
It was more an example that those tiny communities (i.e People using Gtk on macOS) were still enough to get Gtk supported. Which means that the open-source ecosystem will have no problem keeping X11/Gtk supported. A lot of engineering hours were not put into Win32 or macOS support for Gtk. If you look through the repo, it is generally only one or two guys.

That's fine and all. But relegating support for gtk2/X11 down to a single person and a small BSD project doesn't really all that sound promising.
If a single person could do it, then a community of a few dozen or so will manage no problem. And thats what BSD likely has. Plus chuck in the Solaris and AIX guys in there and you basically have nothing to worry about.

X.org works well, but the Linux community has beaten and battered its implementation also. It's just bad.
Indeed. I am hoping once Linux has gone away, Xorg can finally get cleaned up again.
 
It was more an example that those tiny communities (i.e People using Gtk on macOS) were still enough to get Gtk supported. Which means that the open-source ecosystem will have no problem keeping X11/Gtk supported. A lot of engineering hours were not put into Win32 or macOS support for Gtk. If you look through the repo, it is generally only one or two guys.

My point is that most users are using SwiftUI/Quartz based alternatives to major gtk2 based applications on the mac. Those individuals can only maintain gtk2/Quartz bindings for so long before RedHat fully migrates to Wayland. Who knows what they'll do then. X11/Wayland and Quartz are two completely different worlds; with their respective toolkits. Once RedHat ships by default Wayland GNOME/KDE desktops; X11/Qt and X11/GTK will be on life support. I highly doubt the BSDs will pickup up that legacy burden once the toolkits follow suit. As long as Debian ships GNOME/systemd by default; they're at the helms of RedHat/IBM.

I think people here are dismissing how much control RedHat has over the entire Linux userspace ecosystem. Thus, for the desktop, over other platforms; ie. FreeBSD.

Plus chuck in the Solaris and AIX guys in there and you basically have nothing to worry about.

Most people are running Solaris/AIX headless. I'm not sure where this is relevant.
 
I mainly tried GNOME on Wayland, but 8+ years my 1000Hz mouse would move inconsistently with high CPU/GPU load.
GNOME 48 Wayland (openSUSE Tumbleweed) may finally fix this :p

I've been using it for some days now and haven't noticed anything with the mouse (1000Hz, 73Hz display) with general stuff, and was playing some Oblivion last night in Wine vsync no issue.

Even Dota 2 mostly works with SDL_VIDEO_DRIVER=wayland! (no direct chat/Console typing; can copy/paste text; probably some SDL3/Wayland thing but FPS is fine and no xwayland).

I really don't see why. FreeBSD supports Wayland perfectly right now.
I'm not sure how it is on FreeBSD, but generally speaking I'm not confident in general desktop usage on Wayland unless it's on GNOME (I'm finally seeing good results after 8+ years :p), and FreeBSD's GNOME looks to be 42.

Plasma 6 wanted to push ahead with Wayland without having session restore, so I'm not sure how that's usable with an expectation of stability (OOM killer made it easily unusable Fedora).

Anything else (non-specific/general) doesn't have the eye-candy I'm used to from GNOME :p But there's also Wacom tablets I don't expect to plug-and-play work from non-GNOME/KDE with libinput stuff and finger/stylus switches.
 
Of course you don't. Because you don't see the trajectory and movement of the issue. Until I saw what the OpenJDK, Firefox/Chrome and the toolkit folks were doing; I was somewhat sure X11 would stick around. Not so much anymore. The ecosystem is moving towards Wayland; this isn't just about X vs Y Window Manager or DE anymore.
So what?
 
Not sure about the original person, but loosely:

  • Non-Windows OS exists
  • X11 arrives
  • Everyone codes on non-Windows group-effort alongside the only display server
  • (years of ecosystem built)
  • Wayland shows up out of nowhere
  • Loud virtue signaling/etc makes X11 sound like crap (random eg: Phoronix comments)
  • Everyone is either questionably motivated to go with Wayland, or begrudgingly has to follow since no serious OS also wants to maintain X11 beyond what exists today
The virtue signaling thing largely comes from Linux. Mainstream Linux is GNOME. GNOME's very interested in getting rid of X11 support DE/GTK (likely maintenance burden reasons). GNOME is funded by big parties (notably Red Hat). Fedora's a mainstream Linux distro that has more interest in getting rid of X11 support. Mainstreams follow, and X11 is now deprecated no-mans land.

What I don't get is consumers/end users praising it with technical issues for years, but maybe it's bots pushing the narrative and nobody realistically uses desktop Linux/non-Windows :p I don't get the rush for a worse solution, and don't quite get why X11 couldn't be still continued (why re-invent the wheel and then ship a square?)

If I noticed that mouse thing on GNOME 48, I'd still be on Windows now; but quite frankly I question every part of Wayland's development still if that was allowed to exist as long as it did. Do people working on the input stack not know low-latency or use above 125Hz mice?
 
Those individuals can only maintain gtk2/Quartz bindings for so long before RedHat fully migrates to Wayland.
Quartz support is in Gtk 2 and 3 and has never been the focus by Gnome, RedHat, etc. Basically if the Quartz community can do it, so can the much larger X11 community when the time comes (if the time ever comes).

I highly doubt the BSDs will pickup up that legacy burden once the toolkits follow suit.
They definitely will if needed. This has been proven via their regular replacements to Linux-only ALSA support and patches to Chromium as mentioned prior in the thread.

Most people are running Solaris/AIX headless. I'm not sure where this is relevant.
For AIX, yes but for Solaris, strangely no. It has decent NVIDIA upstream support and has traditionally been involved in CAD, etc. Alan Coopersmith was also on the X.Org board for a time.

Its relevant because they will also chip in to support Xorg (if needed). They have large development teams, much larger than most Wayland compositors do in fact.
 
I don't get the rush for a worse solution, and don't quite get why X11 couldn't be still continued (why re-invent the wheel and then ship a square?)
I see it differently. X11 development was mainly backed by Red Hat in the past few years. They decided that the security issues that X11 has are not fixable -> they don't want to invest any more money in it. It's simple as that (I'm not a software develper but I read that in X11 you can write a keylogger in python with as little as less than o dozen lines of code. Is it acceptable in 2025?)

Anyone want to continue X11 development? Just go on, it's open source, it has a very permissive license, I don't understand why there are so many people praising X11 but no one steps in.
 
I think X11 has the same problem that roads have, and Wayland is this flying robo taxi idea.

People spent a lot of time and work to build roads, and with a little work they can be in good condition for centuries. Roman roads are still in use today! Come along some bright young thing, talk about flying taxis, declare roads obsolete. Maintainance may slip a bit, roads fail a bit. But in the end we have no flying taxis and still have usable roads.
 
I'm not a software develper but I read that in X11 you can write a keylogger in python with as little as less than o dozen lines of code. Is it acceptable in 2025?
The keylogger argument is an old one. Effectively it has been proven not to work because the program doesn't have access to the MIT_MAGIC_COOKIE so cannot connect to an existing X11 server.

Likewise, what if I *wanted* to record my key presses, as part of a testing process? The Wayland community needs to understand that not everyone lives in web browsers or only uses a computer to play Steam DRM platform games. Flexibility is what is important for general purpose computers.

Unfortunately the most noisy communities *are* those that don't need flexibility from their PCs, they just need access to reddit and games.

Anyone want to continue X11 development? Just go on, it's open source, it has a very permissive license, I don't understand why there are so many people praising X11 but no one steps in.
They do. The Xorg project is better maintained than 90% of Wayland compositors. The Xorg backend alone has had hundreds of commits this year already from at least 8 different committers (Including Alan Coopersmith from Oracle); that is more developers than most Wayland compositor projects.

What the Wayland community means by "Xorg is unmaintained" is that their tacky whizzy effects and bloat are not being added. Xorg really is much cleaner than people think. They don't want compositing in there (there are external ones if people really need them); so lacking compositor features and other gimicks is perfectly fine.
 
They do. The Xorg project is better maintained than 90% of Wayland compositors. The Xorg backend alone has had hundreds of commits this year alone from at least 8 different committers (Including Alan Coopersmith from Oracle); that is more developers than most Wayland compositor projects.
Good. So where is the problem then? Why do people care about what the "Wayland community" has to say? Red Hat does not want to maintain Xorg anymore, someone else is doing, it's all good. What am I missing?
 
Good. So where is the problem then? Why do people care about what the "Wayland community" has to say? Red Hat does not want to maintain Xorg anymore, someone else is doing, it's all good. What am I missing?
I don't think anything is missing. No-one can predict the future but it is always interesting to discuss potential directions with other techies and getting all sides of the story rather than just Red Hat (and reddits) narrative.
 
  • Non-Windows OS exists
  • X11 arrives
  • Everyone codes on non-Windows group-effort alongside the only display server
True, but that is not just Linux and FreeBSD. There is also (Open-) VMS, which for many years used its own graphics/windowing systems (VWS). There is the NeXT with its postscript-based graphics system. And then there is the Mac, which first used a modified flavor of X, and then switched to a non-X graphics layer.

  • (years of ecosystem built)
While VWS and Display-Postscript were alive, there were web browsers for them. And today, most of the big web browsers are available for Windows and Mac, without using X. Other graphics-intensive applications are available for things other than X also. For example: I use KiCaD for schematic capture and PC board layout, and it works great on my Mac.

But: Having said that, there are lots of minor graphical applications that ONLY run on Windows, and are not available on Mac, nor on portable OSes (iOS, Android). Examples include hardware-specific interfaces such as configurators for PulseWorx, Omega controllers, OBD-II/J1939 engine management for specific diesels (Deere or International), and so on. I keep a windows laptop around for such tasks.

So the idea that X is the center of the graphics ecosystem is valid on Linux only, and Linux has a marketshare of low single-digit percent in the desktop OS market. Even the second largest desktop/graphical OS (Mac) has only incomplete support.
 
True, but that is not just Linux and FreeBSD. There is also (Open-) VMS, which for many years used its own graphics/windowing systems (VWS). There is the NeXT with its postscript-based graphics system.
Can go with Sun's approach of *everything*. Xnews at a time in its life supported:
  • NeWS
  • SunView
  • PostScript (Ah, came with Xsun)
  • X11
I bet that codebase was... mature by the end.

And then there is the Mac, which first used a modified flavor of X, and then switched to a non-X graphics layer.
That's interesting. Did you have a name for Apple's X flavor? I sense an archeological deep dive is waiting for me here.... :)
 
They definitely will if needed. This has been proven via their regular replacements to Linux-only ALSA support and patches to Chromium as mentioned prior in the thread.

I just don't see it. Only time will tell. Upstream has drank the GNOME/Systemd koolaid a bit too much. I still think Wayland is a chance for BSD to write a clean, grassroots implementation of a display server. X.org is a mess.
 
It really is though. The while GTK/GNOME/systemd/Wayland sandwich is practically control by our IBM/RedHat over lords. They have their tentacles on the entire userspace ecosystem outside the Linux kernel. That was their entire plan. So we can just forget about GTK. Now if Qt and Firefox/Chrome (who are RedHat loyalists) decided to stop future X11 development, FreeBSD is pretty much screwed. We have no control over kwin, mutter, etc.

Unless someone decides to fork gtk2, who's going to maintain it here? The committers aren’t going to bother touching GPLd code outside of drivers. The toolkits and apps are whats at stake here, less so DEs. its either fork the toolkits or write our own Wayland backend for them and browsers.

IMO, I think FreeBSD is ripe for disruption in writing its own API Wayland compositor. The project could penetrate other markets like IoT, embedded, mobile, etc. I’d like a base system I could easily port to ARM/RISC-V phones and tablets. The license, API/ABI stability, and features would fit the concept of Wayland.

Thanks for coming to my TED Talk.
That would be interesting whether the FreeBSD OS developer team would write its own X or wayland alternative.
Wasn't there X12 developed somewhere, or was it just a meme ?
 
IMO, I think FreeBSD is ripe for disruption in writing its own API Wayland compositor.
Recently the fund had a lot of money to improve laptop support. How much money is needed to write API-Wayland? I think that without money nothing like that will happen.

On the other hand, as a regular desktop user, I see no reason to worry. Today I already have a full set of toys: the native Wayland itself, sway, a set of drm-*-kmod, a bunch of basic applications that work with Wayland. Yes, something is wrong somewhere, but 90% of the functionality suits me. If I need something, as ralphbsz wrote, I can always get a PC with Windows from under the dusty sofa and run an exotic program. In any case, we are indirectly dependent on Linux drivers, on the eco-systems of corporations. But the plus now is that the bundle "Wayland + WM + drm-*-kmod + basic software" already exists. And this can ALREADY be used. Our system does not have systemd or any other direct control. I see it this way, it's just that with age, many do not want to relearn. It is not easy for me as a user to switch to another paradigm because of my habits, but I am doing it. If I don't die, then in 2-3 months I will get used to shortkeys, workspaces and master keyboard navigation at a good level. This is a special case. And it WORKS for now.
 
Back
Top