If he was still about and in to tech, he'd be the poster guy for this....
.... after all 'I Did It Mir Way'
Canonical is still working away on its own Mir display server, used in several of its IoT product lines. Version 2.14 gains more functionality useful for full desktop environments. Mir is a complex project which has undergone some big changes over its more than a decade of existence, and it has several subprojects now, …
Wayland still lacks some functionality that I use under X11, such as global shortcuts and window ordering. These are implemented by KDE and Gnome, so I would have to target these on a per desktop basis.... yes I know security issues... but Windows implements them and it's not as if an X server still runs as root... (don't forget the desktop implements them....so).
Remote display of a single application over ssh is missing. Very handy to have over a slow link when you can't plug a keyboard, mouse and monitor into that device buried in a water tight box behind a sealed bulkhead.
Still can't get Wayland to work on Arch Linux with an Nvidia T1000. I use Nvidia cards because they are fast, way simpler to debug GPU programs and graphics then anything else.
Still waiting for Wayland to survive a sleep/wake cycle without becoming corrupted. Cinnamon/Xorg survives 9/10 times and I have a spot on the screen to click to restart it. I can't ever get Ubuntu/Wayland to work after waking up the computer. Maybe I can't even log in. I first suspected it was the usual NVIDIA bugs but Wayland isn't using it.
> survive a sleep/wake cycle without becoming corrupted
Just for what it's worth, the go-to option for recovering from a corrupted graphical display on pretty much any Linux box is to switch to a virtual console and back. Ctrl + alt + F1 through F6 should get you to a clean text screen, and then alt+F7 should take you back into graphics mode forcing it to be completely redrawn.
Until they come up with a speedier system for remote displays, it isn't going to be of much use to me.
As far as I can tell, the whole branch of code for Wayland was started because someone wanted to emulate the concepts of DXn for Windows under Linux. i.e. Tie it in to the system more closely to theoretically display a few more frames for video games. But given the paucity of video games for Linux that run native instead of under Wine, I think it is a waste of time. It would have been better to further enhance and build on the X protocols, perhaps pruning some of the oldest and least used features at long last.
> It would have been better to further enhance and build on the X protocols, perhaps pruning some of the oldest and least used features at long last.
I think you make a very good point. I don't think how many of the people who always comment about X11 working over the network realise how close X.org is to death now.
It really is in maintenance mode. About the only people that were really actively working on X.org were the Xenocara team from OpenBSD, & now that there is an experimental version of Wayland for OpenBSD, that could be in danger.
I really wish that a bunch of X11 admirers and enthusiasts would get together and create a wiki style list of features that everybody wants, needs, and uses in X11, and ones that barely anybody and anything uses anymore.
If it were possible to come up with an agreed list of "stuff we really need to keep" and "stuff it wouldn't kill us to lose", it might be feasible to come up with a plan for a (maybe only vaguely backwardly compatible) X12, which preserved the good bits. Otherwise, through sheer inertia, Wayland is going to win.
In the meantime, though, I think that it's really good that somebody else, outside of the core group, is working on an alternative system, one with rich support for touch and so on.
> the very people who are now busy working on Wayland
I don't think so. I think there are (at least) two camps: those who want the new shiny, versus those who want the old stuff to stay working.
I regularly read — and occasionally write — about fancy new functionality that is coming to Wayland or occasionally to gnome or to KDE. I recurring theme for me is that it's stuff that I have absolutely no desire for or interest in for my own computers.
And yet, I look at what macOS manages to do with, for example, two different screens with extremely different DPI ratings – as I am using right now, in fact – and I lament the fact that X11 simply can't do this yet.
My worry is that X11 is going to get hugged to death. People will cling onto X11 because of perceived vital features that in actual fact many of them never use, such as operation over the network, and because of the determination to keep this decades-old code base alive, it will remain too large and too complicated to add vital new features. Until the point that nobody is prepared to pay for it to be maintained anymore, and we all have to switch to this new alternative.
The only way out of this monkey trap that I can see is to isolate two separate positive subsets of features and functionality from X11, versus one negative subset:
* the essential functionality to preserve its unique advantages over Wayland
* the essential new stuff to keep it competitive with Wayland. For me, and I speak for nobody else, that means supporting different DPI is on multiple heads under Xinerama.
* the legacy functionality that nobody actually uses anymore on the 21st century operating system.
I suspect that that third bullet point contains large features, such as font servers and X fonts, some degree of support for different bit depths, support for both big endian and little endian operation, add a bunch of other stuff which I can't even point in the general direction of, but which nobody uses on any machine running Linux kernel 5.x or 6.x.
Step one in "Operation Save X!" would be to try to work out what to save, versus what could be thrown over the side.
But that means making tough decisions, and nobody is willing to step up and make those.
And if they're not made, I think that the writing is on the wall for the X window system in general. :-(
The real problem here is that the people that think continuing to evolve X11 is a good idea do not have the experience of how very problematic that is.
It has been a very successful design over the past 40 years, but lacks any sensible way to evolve rather than simply grow.
X is not fit for the modern world in fundamental ways: for one of many examples, the input and output of every application should not be available to all applications, but attempts to fix that just proved how hard that is in an ecosystem built for a different age.
Wayland has its share of problems, but it has a design that supports the creation, updating and deletion of extensions. That makes it possible to evolve in ways that were simply not possible with X11.
Most importantly, it builds on the experience of working on X11 and seeing what is effective and what is problematic. In other words it is your hypothetical X12.
In any case, the de facto decision is made, both compositor developers (for desktop environments and other shells) and toolkit developers (for applications) have accepted that Wayland is the way forward.