Three signs that Wayland is becoming the favored way to get a GUI on Linux It has taken about 15 years to get there, but there is mounting evidence that the Wayland display server may soon topple X11 as the most common way to get a GUI on Linux. We've reported on growing endorsement for Wayland recently. The team developing Linux for Apple Silicon Macs said they didn't have the manpower to work on X. …
I think its a glaring sign of failure that the open source world is normalizing the blind acceptance of these decadal failures, mostly because nobody bothered to build something better.
That last part is the crux of the problem, as it's core is a lack of coordinated development and resources put to the harder and less sexy parts of the operating system. Hence we have had 4 new file systems since EXT that don't have reliable drive tools to do basic things like repair/resize/compress, and of the ones that do, to many only work on a offline volume. An IP stack that made the routing tables more efficient, but failed incorporate basic security concerns, most developing technologies, and punted the work for "fixing" things to other teams like DNS, DHCP, and BGP. Or left them broken, like the architectural problems IPv6 has with load balancing or fail-overs across multiple WAN links.
We have tapped out trying to fix too many problems, and if and idiot yells loud and often enough, we let them redefine binary math to decimal values(lookin at you Western Digital) or pretend that somebodies proprietary quintuple shielded cable is the new CAT-7,8,9, or 10 standard even though nobody actually uses it. If we can't move past the days of letting wikipedia edit wars define technical specs and prolific but garbage coders flood the unix ecosystem, it's best days will be behind them.
"...the Linux world in general was going to stop complaining and just accept systemd"
That might be an overstatement.
I'm looking at distros as I am considering moving from my current distro due to some sort of weird compatibility problems, and from what I see there are a lot of distros still, blissfully, systemd free. That, and judging from a lot of the comments here and on other sites, although the volume of complaints may have moderated, they are still being expressed.
It may be that the part of the Linux community that is more receptive to systemd are the developers and maintainers. Even so the fact that there are distros actively avoiding incorporating systemd shows that not even all of these people have been assimilated.
That might be is an overstatement.
There you go.
... volume of complaints may have moderated, they are still being expressed.
Moderated because (becoming wise to what systemd is) most of those complaining probably moved on to a systemd-less distribution.
Because, who in their right mind wants that crap in their OS?
... more receptive to systemd are the developers and maintainers.
Yes.
Incredible as it may seem, that may well be so.
There is a lot of moolah behind the push to systemd.
Where is the author of that abortion of nature working today?
Me?
Would not have systemd in my box unless hell freezes over, three times in a row.
.
You shouldn't confuse the number of distros with the number of their users or, indeed, their overall importance in the Linux ecosystem. Every major distro uses systemd. And for good reason. It works well and has useful features, good documentation, and easy-to-write service files. It's not just developers and maintainers who like it but sysadmins like myself who remember the hell of hacky init scripts, especially when it came to dependencies and issues of concurrency. I suspect most desktop users won't care what init system they're running and don't have much of an opinion either way. As it should be!
While there may be a "philosophical" argument against the init, essentially that it's not UNIX-like, I find it telling that those lines of argumentation usually ignore things like the X server or the Linux kernel itself (Linus never meant it to adhere to that ideal, incidentally.)
It is popular to hate on the init system here, no doubt. I fully expect the usual slew of downvotes to this post. But it is worth bearing in mind that it is a highly vocal minority that hasn't moved on from early justifiable criticism or have a, frankly strange, vitriolic dislike of the init's creator. The author's point about people voicing their liking elsewhere plus the widespread adoption underscores how it is very much not an overstatement but just reality. systemd is fine and has been for a long while.
> It is popular to hate on the init system
The objection to, and the best case for, systemd is that it is not just and init system, but an extra layer of the os including the init system.
Wayland is very different. It is a replacement for a limited set of funtionality. No one has a philosophical problem with it, just concerns about whether it is mature enough and worries it may cut out the BSDs.
What an odd reply to my post. In no way have you addressed the original comparison by the article author that was mentioned by the poster I replied to nor the incorrect assumption to that is anything but a loud and unrepresentative group that complains about systemd.
As for Wayland being free of philosophical debate that is simply not true. I've been following discussions (as well as using wayland) for a long time now. For starters, there's been a lot of discussion about the fact that the protocol is decoupled from the actual server implementations, ie the compositors. That point has caused issues that are still relevant with KDE and GNOME; they have issues that are not present in the other or wlroots-based things for that matter. How input is handled and the "every frame is perfect" approach likewise has been the focus of a lot of debate. Hell, even elsewhere in these comments you'll find people complaining about its CSD-first approach.
The person who replied made valid points which were germane to the discussion, of which you are but one participant. The fact you are whining that their reply isn't solely addressed to you further undermines your credibility, as you seem more genuinely offended you aren't being treated as the center of the universe than by any of the points they made. You don't really refute their points either, despite the attempt at redirection.
So what are you left with? A rhetorical misstep by the prior poster in claiming "No one has a philosophical problem", which then lets to jab at them by creating one. All while ignoring the broader intent to show that the academic concerns remaining with Wayland aren't remotely close to the cancerous trainwreck that is systemd.
If there are problems with Wayland, they could and can be fixed either in Wayland or a relatively easy to swap out replacement, should someone choose to build one. The core issues being that the design and the pace it's been built haven't got buy in across the unix ecosystem, and that there aren't credible alternatives to Wayland or X that really address them despite the age of X and long infancy of Wayland.
"No one has a philosophical problem with it"
From my perspective, the arguments against Wayland are about equally split between technical and philosophical.
"just concerns about whether it is mature enough"
Which it isn't, not by a long-shot, and after 15 years of development (15!), I'm pretty sure it never will.
"and worries it may cut out the BSDs."
Frankly, the BSDs can handle themselves. They'll either make it available to their users, or not. Regardless, Wayland will not "take over" anything in the BSD world because the very idea of "taking over" is anathema to the BSDs. Something the kitchen-sinkware corporate Linux distros would do well to emulate.
We already have it on FreeBSD and its decent. It's not that difficult to get working with a DE either, I use Wayland with KDE as my daily driver on that OS if I'm using a GUI there anymore actually. OpenBSD is working on it from what the article says, it'll probably wind up being the same way over there but I can't speak to it.
"about whether it is mature enough"
Indeed, and I quote this from another forum:
"You failed to read the fine print at the bottom of all the wayland promises over the past 12 years:
"It will improve your performance. Next year. Or the year after that. Or maybe the year after that. If you have the right hardware. And the right desktop. On certain tasks with certain apps. Maybe. Depends on the alignment of the stars and the moon, and if Jupiter is in the 2nd house"."
[Author here]
> The time to watch Wayland is when they merge with Yutani in 2099.
I mean, yes, I chuckled, but still, Weyland/Wayland. Not the same.
I get people angrily "correcting" me because they haven't noticed one-character differences such as this.
Some of it is being forgotten. E.g. in olden times, MacOS and Mac OS were 2 different OSes. Now, it is MacOS and macOS: arguably an even smaller difference.
A MAC is not the same thing as a Mac. There is a unit the kB, but not KB; there is a MB but not a mb; and Mb and MB are very different things.
It matters, in this industry more than anywhere.
[Author here]
You're right.
> You shouldn't confuse the number of distros with the number of their users or, indeed, their overall importance in the Linux ecosystem. Every major distro uses systemd.
100%. There are quite a few distros that don't but they are quite niche. It's over; systemd won. Sorry, but it did.
The interesting question, for me, is: why?
> I fully expect the usual slew of downvotes to this post.
You were right there, too.
:-(
"There are quite a few distros that don't but they are quite niche. It's over; systemd won. Sorry, but it did."
Only two major distros adopted it (RedHat and Debian). RedHat did it because they are trying to be Windows (were trying? gawd/ess only knows IBM's intent at this point). In Debian's case, it was an accident of history, in essence fall-out from a large internal power struggle. In other words, it was a political choice. It certainly wasn't for technical reasons. Thus Devuan.
The rest of the distros to implement it, being mostly clones of those two, blindly followed due to ignorance and/or apathy, with a pinch of sheer laziness on the part of the devs. They certainly didn't spend any time thinking about the ramifications, beyond "I use that software repository, so I must comply".
y
The systemd-cancer didn't "win", rather the community at large is losing. This can still be turned around.
Over the last dozen years or so, I've noticed the SUSEs quietly disappearing from my friends and neighbors desktops. Nor sure what that means in the great scheme of things, perhaps I should conduct a short poll asking why (if I can be arsed, I might). It's probably because it has drifted too far from "mainstream", whatever that means.
I like Arch. If Slackware didn't exist, I'd probably fork it into a variation sans the systemd-cancer (which works quite nicely, BTW, try it!).
Ubuntu is just a corporate rebranding of Debian, NOW WITH ADDED KITCHENSINKWARE!!!! I don't consider it and sub-distros to be serious Linux contenders.
If you honestly think the systemd-cancer has won, how do you explain MX Linux being at the top of Distrowatch's PHRs this last year, and by a wide margin?
Though now begins the dreaded task of finding a valid donation link or passing a check to someone that knows someone that can pass it to him.
Oh, and don't try to "volunteer" to help me here. We've be down that road before. And Yes, I'll start by looking at the patreon, then have to figure out if it's a fake, and if it isn't how much graft their wringing out of it, then end up sending a wad of grants by carrier pigeon. It's a good cause though.
I didn't have the scratch back in the day, but maybe now we can get them a SSL certificate for xmas :-)
Agreed the main reason for objecting seems to be it looks like "Windows".
However when I coded a long running server to run on an early version of WindowsNT I was deeply impressed by the Server API which was well written and reliable with a great API.
Not all of Windows is cr*p so why not copy the good bits.
It's less that the foundations are crumbling, and more that change is happening on the hardware side of things.
If we had stability in that software didn't become more bloated and resource hungry, and new programming and application paradigms didn't keep appearing, then the existing tools would work fine forever.
But change is happening. More powerful systems, with more and different cores, more memory, more network bandwidth, and new storage designs appear, and existing software needs to change to accommodate this. I don't thing anybody thought when Linux was first being written, that we would have the monster 64 bit systems with multi-gigabyte RAM and terrabyte disks.
What is happening here is that the new, young things doing the programming don't want to learn legacy, they want the new, shiny, and feel that re-inventing the wheel rather than just changing the tyres is the way to go.
For the Yanks in the audience who don't do Wiki, the Wagon Wheel (from 1947) is essentially a British copy of the Moon Pie (1917) (in some states the Scooter Pie is more common, but is essentially the same thing). If you prefer ice cream to marshmallow, here in the US you'll want an It's-it ... which is a whole 'nuther kettle of worms.
No, they are not identical, but they all share a common esthetic. Kids love 'em, adults maybe not so much.
All are easy enough for kids to make ... When we host birthday parties, we have 'em make their own.
RC cola optional (personally, I prefer coffee ... strong, black, no sugar. No kid's confection, either, come to think of it ... ).
> Old and showing its age?
I have run Fluxbox for the past 20 years. Old? Definitely. Showing its age? Definitely not! It has stayed true to its original conception - elegantly minimalist, highly functional and extremely configurable - and resisted the temptation to bloat out with shiny. (It is a window manager rather than a desktop - I do not need/want a desktop.) It is responsive, lightning-fast, rock-solid, doesn't throw surprises, and is available on, and portable (along with configuration) to pretty much any *nix. Despite trying every desktop going (I am an inveterate fiddler), when it comes to getting actual work (and play) done, I have found no reason to change. It is not broke and doesn't need fixing.
I'm not sure how it plays with Wayland.
"Would you run a Ferrari on Boudica's chariot wheels?"
No, but I probably wouldn't drive a Ferrari on dirt roads or on ice or snow covered roads here in Vermont. Neither would I use it to plow a field unless the alternative was pulling the plow myself. If I HAD to drive a Ferrari on Roman roads and I had a choice, I probably would select something quite different than the factory wheels and tires and perhaps closer to chariot wheels. And I'd probably try to do something about the road clearance.
I'm curious what you think has changed in the past three decades that requires upgrading from X11. It's not like GUIs have actually done anything other than repaint and rearrange the deckchairs in that timespan.
If you actually have a use case that requires Wayland, by all means use it. That's one of the virtues of Unix. It's relatively easy to replace even major subsystems. But I have to say that from the point of view of running applications and getting results -- which is what I use computers for -- systemd and to a lessor extent wayland -- look to me like solutions in search of a problem.
High DPI
High DPI wouldn't have been a problem if a) application programmers had asked the X server about pixel density like they were supposed to and b) X.org hadn't subsequently wired the server to 96 dpi even if it was running on a 200+ dpi display. The number of times I've had to fight bloody Firefox & Thunderbird to get properly sized controls because they've changed the scaling mechanism yet again for no good reason. Rant, fume!
The X server devs were or are the same people as the Wayland devs and claimed that they changed it to make it bug compatible to Windows.
Windows has not had that bug since the release of Windows 7 but it still is in the X server.
The X server devs have answered all bug reports about the fixed 96 DPI with WONTFIX and said that the fix is Wayland since 2006.
Never mind that Wayland was nowhere near ready in 2006 or 2016 for that matter their intent was and is to force the switch over to it.
It's not really memory or CPU in the case of the display system (I was talking in a more general sense), but hardware wise, the actual display hardware also changes, and is continuing to change rapidly.
The people doing the the most promising Linux port to Apple M1 and M2 hardware have said that they do not have enough resource to both re-write the X.org backend display driver for the new silicon, and also do a Wayland compositor. So they've opted to just do Wayland. And Apple never have done a native X11 driver for their display silicon fro MacOS, so are not going to do one now.
Unfortunately, the different vendors of display hardware all do it their own way, and are developing their own silicon, with different numbers of cores and basic graphic primitives, different levels of abstraction for higher graphics operations and even different ways to send the commands to the display hardware. And they don't keep it the same over time. New cards from Nvidia, AMD, Intel et.al. just don't work with the old drivers, and keeping up with the changes if the vendors themselves don't do it is very difficult for community development projects, especially if the specs. aren't published by the vendors.
This is, unfortunately, the way things will go forward, and an indication that Linux is moving further and further away from being a UNIX-like OS to being something distinct from it.
"The people doing the the most promising Linux port to Apple M1 and M2 hardware have said that they do not have enough resource to both re-write the X.org backend display driver for the new silicon, and also do a Wayland compositor. So they've opted to just do Wayland."
Possibly smart. When Apple change their mind and do another radical switch in hardware (as is their wont), they'll have to re-write it all from scratch. Again.
I hate treadmill programming. Still, I guess it's a living.
"an indication that Linux is moving further and further away from being a UNIX-like OS to being something distinct from it."
And yet here I am, happily using a non-systemd-cancer distro, and not running Wayland. And I see absolutely no reason why this will change in my lifetime.
Linux is the kernel. It is not X or Wayland or the systemd-cancer or any other init. Linux is just the kernel. Shall I repeat that? Linux is just the kernel. You are allowed, nay ENCOURAGED, to graft the bits and bytes onto it that make an OS that suits you, the way you use a computer. There is no "one size fits all", and never will be, despite all the corporate interests trying to make it so.
"What is happening here is that the new, young things doing the programming don't want to learn legacy, they want the new, shiny, and feel that re-inventing the wheel rather than just changing the tyres is the way to go."
Unless they learn the legacy first they won't know that which they're reinventing is supposed to do. That means there's little chance that their reinvention will do it. Then stuff breaks and they don't care because they don't know that caring was necessary.
> Unless they learn the legacy first they won't know that which they're reinventing is supposed to do.
Like 100% supported network transparency between any two boxes.
As Wikipedia (still) says "a compositor can implement any remote desktop protocol" but will it - further, will *all* of them provide both sides of at least one such protocol?
I’ve been in IT for almost five decades. What strikes me most is the number of times new tech is introduced without considering what has already been developed. Developers love developing, so they re-invent the wheel time and time again rather than tweaking current technology.
The history of X11 is long, and I've used it for nearly 30 years (my first computer which ran X11 (just) was an Amiga A3000 with 10mb of RAM). X11 is OK, but a bit of a funky design, which primarily got adopted because it was a good way for the other UNIX vendors to ensure that Sun didn't fully capture the workstation desktop market with NeWS.
It's design predates many hardware features and concepts and make their use awkward or impossible, but it's most significant problems these days are with security. Many of what were originally design advantages of X11 are now just overhead. I once tried to do some low-level X11 development. It wasn't a pleasant experience.
I switched to Wayland about 2 months ago. Performance is better, stability is worse, a few features don't work. I don't expect to go back.
Submit patches and/or good quality bug reports.
"stability is worse, a few features don't work."
Lovely. Just what I want in a serious GUI for my business.
I know people have been conditioned by Microsoft to think that computers are supposed to fail on a regular basis, but I'm sure you'll pardon me if I choose not to jump on that particular exasperating bandwagon. Life's far to short to deal with computers that don't work... especially when there are computers that DO work within easy reach of all and sundry.
They're not crumbling, they're doing what foundations do: sit there solidly being purposely hard to change.
The problem is that to developer such a system is difficult and boring but creating a whole new thing is exciting AND you don't have to deal with the hard problems for years because there are so many easy/fun problems to solve first (problems that the project you're purportedly replacing already dealt with, which is why it grew all those warts in the first place).
" ... That would explain why the devs are so keen to push the changes. ... "
So could ego and narcissism.
I'm 70 years old. I am not against change. I have seen more of it then many, if not most here have. Change is in fact a good thing. But change simply for changes sake is not. Objectively, not subjectively. Is systemd objectively better? Is it faster? Does it use less resources? Is it simpler? Does it do the job of that which it was meant to replace? Or does it do more, intruding into places that it doesn't need to? If the answers to any or all of these questions, and more, objectively, is no, then how and why is it "better" and why should it be adopted?
Wayland. I have no problem with it at all. Haven't knowingly used it, so I have no objective, or even subjective, reason to like it or dislike it. Change is good. If in fact, that change is good and not change for the sake of change. Or for someones overblown ego or narcissism.
It's not hard to change my mind. Show me that's it is,objectively better. And I will have no problem. But if you are going to be subjective, to offer opinion, or try to blow smoke up my ass, just go the fuck away. I don't have the time or inclination.
I have tried Wayland and will try it again. I have also performed some research. Is seems to me that the biggest difference is that Wayland does not do network display. At this point in time, that will not stop me from using it. But, I have in the past used X network display capabilities. So, I have hesitance to using a display technology lacking the network capability as efficient as X has achieved.
On the other hand, X development is lagging. I am not looking for the new shiny. I am concerned about security. I am concerned about keeping up with display hardware.
Change happens naturally, you have to live with it. Change for change sake as said is useless. There is also change for monetary sake. If the change is only for the monetary benefit of the driver of change, then that is even worse than change for change sake.
Every time I try Wayland (as recently as this quarter, both Fedora and Debian KDE) things don’t work. For now, I’ve retreated into tiling-WM-land (StumpWM is very nice) and as it stands now, it’s gonna take a forcing action like, say, Firefox going Wayland-only before I consider it again.
In my experience this is because people keep fiddling with stuff.
My ancient AMD graphics card worked fine for years. Then someone changed something (probably something completely pointless for that card, like 3D ray tracing) and now I am in 1024x800 mode.
Another weekend of debugging!
You had to update the device driver, didn't you?
Both AMD and Nvidia retire the drivers for old cards. Chances are your card is no longer in the AMD universal driver, and you're back in un-accelerated VESA mode. Reinstall the old driver and pin it so that it doesn't get upgraded, or switch to the open drivers.
"Reinstall the old driver and pin it so that it doesn't get upgraded, or switch to the open drivers."
Try both. Pin the one that works best for your needs. It's your computer, it doesn't belong to your video card manufacturer.
Note that one or more of the current LTS kernels might be your best option for older hardware. For example, LTS kernel 4.4 (released in very early 2016) will be maintained until at least 2026, and probably until 2036 ... and possibly beyond, if there is a need. If your hardware runs nicely on that, it might be an option for you.
Indeed. The ultimate OS is the one that blissfully gets out of the way and allows me to do my job with no histrionics. For my needs, the systemd-cancer free Slackware (sans Wayland) does exactly that. Try it, you might like it.
The only time I actually think about which OS I am running is when I'm responding in threads like this one.
I also use Slackware. Wayland is present but an option that I periodically evaluate. I have not yet been sold. I have not yet been completely turned off.
On systemd, it may not be present, but it has left it's mark. In order to provide a reasonable selection of tools, it has implemented workarounds to appease various packages. I have successfully installed packages which supposedly require systemd, but only require "infrastructure" of systemd. Lest us not forget that Slackware uses elogind as a workaround for hard-coding for systemd dependencies.
"Maintaining sysvinit would be less effort than dealing with systemd related problems by many, many orders of magnitude." -volkerdi 08-16-2012, 04:42 PM post #164 to thread Slackware and systemd on linusquestions.org
I keep hearing about fractional scaling and suchlike, but what I really want to know is this:
On a technical level, is there anything stopping a window manager or compositor from freely scaling windows independently of each other? I've wanted this for a long time - and I think it was kind-of-possible using Compiz (something else that's on life-support, if even that.).
Basically, what I want this is:
When, say, a calculator app (i.e. something non-resizable) is using an unreasonably large amount of screen space, I want to hit a key combination (or hold down a key and roll the mouse wheel) and have just that one app shrunk by 10%. Shrunk, not resized - as far as the app's concerned nothing's changed - instead the window manager / compositor is just presenting a shrunken view of the window, and transforming input events accordingly.
Likewise, found an app with unreadably tiny text? Hit a key combination to resize the application to, say, 75% of the screen size, then scale the smaller window up so it fills the screen. (On high-DPI displays - especially with some suitable filtering - I'd expect the result to look reasonable, even if not as razor-sharp as it'd be without scaling.)
Is this possible today, or is anyone working towards making this possible?
Windows can do it per application per monitor.
But it's a pity that MS's own applications, such as Edge, get screwed up by it. I'm afflicted with Win10 at work, hover over a link, and Edge gets really confused and hides the pop up address nowhere near the mouse cursor, but on a different screen (the laptop screen that I keep half closed but can't fully close for different reasons) where I can't see it. I'm fairly sure it's related to the screen scaling which is 50% for the laptop screen (a poxy 13", unusable, but our corporate standard).
Don't confuse "does it" with "it works" or is even used by the same vendor's applications.
It's possible in KDE to launch gtk3/qt5 apps with a specific scaling factor, so if you know beforehand you want .75 or 1.50 for this application, you can edit the .desktop file for it. You can't change it 'on-the-fly' once it has launched though.
Info found on this thread: https://forum.manjaro.org/t/scaling-individual-applications-in-kde-plasma/109075
Reminds me of the Dell/EMC RSA SecurID app. The, admittedly, old version I'm running is only 8cmX3cm. The update to this old version increased the size of the application window to take 1/4 of the display......for an 8-digit number....WTF...AYFKM.... Why would an application who's raison d'etre is displaying 8 digits need to take up 1/4 of a 1920x1024 display? Ignorance? Stupidity? Blindness? I've got it! They wrote it for full-screen on IOS/Android and the backported the code to MacOS/Windows because they're total fuckwits.
Yeah, technically it's very possible. When the application makes a drawing call, the window manager just rescales it based on what window it is and and what monitor. Like AC said, Windows does it no problem, even for windows that are across multiple monitors with different resolutions and scaling factors.
The big hassle is putting in all the UI and hooks for all this, and handling the edge cases (literally in this case, har) where the window goes across multiple monitors. Basically, you just need some developer to care enough about it to put in the work.
Windows isn't a good example as it's implementation is stupid.. there are about 3 of them that yield different results, and they have to be be implemented by the apps.. a proper dpi scaler would be done at the OS layer not forcing apps to implement it.
The result is that some apps do it right, some half do it, and some not at all. If you're developing and you pull in a library it may or may not be hidpi compatible and even if it is it might use a different method so not be in sync with your app, leading to bug reports and annoyances for users.
A better example is mobile where the UI system was written to scale from the ground up and you largely don't have to think about it.
Lot's of support folk want a (network) remote host to display a GRAPHICAL window on a local host -- like:
$ ssh -X -C -l remote_username -p remote_port_number 192.168.NN.MM
So do I.......and I've looked, and it seems Wayland does not help with this sort of remote GRAPHICAL access.
Please advise!!
It doesn't have it, and it's not going to have it as part of the Wayland protocol.
What there is is an X server called XWayland that runs on top using the Wayland protocol to talk to the compositor (think display driver). But it's not perfect, even though it is mainlined in X.org's X server. And I believe that it has problems with remote sessions, and many things like window re-parenting and keymap modifications, and possibly cut-and-paste will not work exactly the same to anything outside of the XWayland display. I've not played around with it much, so I don't really know.
Having got used to using X for the last 30 years, this seems to be a backwards step to me.
"No wireless. Less space than a nomad. Lame."
But seriously, sometimes stripping out support for things is the way you go forward, particularly when those features are used by a vanishingly small percentage of users AND there are still suitable alternatives. The alternative is basically Windows, where weird and often contradictory ways of doing things end up butting into each other, confusing users and developers alike. Apple eventually removed the ability to run 680x0 code from OS X. Likewise, it's now removed the ability to run PowerPC code. One day it will remove the ability to run Intel code. This is how things move on: keeping in everything that's ever been in there just because there's someone on the planet that still uses it is a way to software stagnation.
@ianbetteridge
Quote: "...how things move on...."
Hah!.....you will be telling me in the next ten minutes that we don't need the wheel.....after all the human race has been using it for...what....five thousand years......Yup!....Ian Betteridge has a MUCH BETTER IDEA!!!
To the point....like the wheel, X works JUST FINE!!!
Wayland overall is a huge step backward. That's why it's still an also-ran after 15 years. Nobody really wants it outside the Corporate world.
Remember, X is still available. And will remain available until roughly the heat death of the universe. That's one of the beauties of FOSS.
Yes, X has problems. It needs an upgrade or replacement. But Wayland is not that replacement no matter what the fanbois tell you.
I use nomachine when I can install it on the server (windows & Linux).
I use remmina for windows remote desktop otherwise. Windows RDP turns off graphics acceleration and the only way to turn it back on is reboot or local logon.
vnc I've not used in a long time.
Not convinced - for us, Wayland still seems to have fundamental issues, like not actually working. The number of times we've had apocalyptic desktop weirdness happening, second or third screens going AWOL, or unexplained 'treacle wading', it seems to have Wayland always. Flip to X11 and everything's happy and normal again. This is on recent Ubuntu versions. I'm guessing in another 10 years it'll be fully ready for primetime.
Interesting, yesterday I booted an Ubuntu 23.10 nightly build, on a shiny new Intel i5 laptop, and it still kicks into X11 by default, even now.
My beard is more white now which is a disappointing development. However I am one of the very few old ones here who enjoys Systemd (bring out torches and pitchforks) and now also Wayland. The KDE compositor had the last bug that was bothering me fixed just recently.
Wayland progress may have been glacial in the past but development have accelerated a lot now, not just the Wayland protocols but also the different compositors using it as well as GFx drivers.
The new Thunderbird is also great, btw.
The X11 system is old and it has its shortcomings but its fundamental strength it its not a remote desktop. It can be used to run a remote GUI, or at least give the impression that you're running a remote GUI. Its a really subtle distinction that I fear its one gets lost in the wash with alternative designs. I haven't done anything with Wayland but it appears to be "yet another windowing system" which like the infamous Microsoft product appears to be heavily integrated with a system's kernel. It will probably do screen rendering perfectly in ways completely unforeseen by the original X11 developers but somehow it will completely miss the point.
Still, I suppose being "old and in the way" (like my mother I went gray in my 30s, BTW) I daresay my ideas about computing are irrelevant. Most of my work never had a human user anyway -- it colors my perceptions of what a computer is -- so it explains why I tend to think of GUIs as a bolt-on rather than something fundamental to a system.
I've found that Wayland lacks a number of use cases that work on windows and x11 that I require for applications I write.
I usually run several programs independently controlling thier stacking order while also sending events to them as needed.
The primary reason for this is to ensure the programs run at max speed with no interference with each other.
The Wayland way is to not permit at the protocol level due to security... At the desktop level it's implemented but differently by each desktop...
My other use case is remote display of a program... via ssh... rather then using nomachine for the whole desktop... Primarily due to limited bandwidth as I occasionally use radio comms while solving problems as I can't plug a laptop in as it's wet, the platform is moving and I can't get onboard.
Here we go again, same old scenario, 2nd, 3rd, 5th, 8th verse. Somebody gets a bug up their butt that their next great killer app is ready for prime-time and starts marketing chatter like the My Pillow goof claiming their new app is the best thing since sliced bread and they dupe a couple major distros into foisting the tech onto the user-base who then deals with black-screens, crashes, frozen displays, hardware incompatibilities, app incompatibilities and feature loss for the next 10 years. The backdrop in the Wayland camp is they just want to get the damn thing out the door and claim it's done and at release-quality. (just like KDE 4.0.4a was release quality in openSUSE 11.0... or Gnome 3 was release quality.... take your pick)
Now it wayland's turn to be the my pillow goof. Problem is there are very few desktops that are actually supported and there is no easy replacement of X with Wayland -- period. Wayland breaks every desktop and requires extensive mods to make work. Gnome is basically the only major desktop that fully supports Wayland -- and that isn't a good thing for Wayland either. Basically a my pillow quality Gtk GUI with libadwaita hacked on top and all XDG menu categories blown away so every other desktop is left with a royal menu-mess. Gnome's attitude -- "sucks to be you, you should run Gnome..." Wayland isn't much different, the fact it isn't even close to a 1:1 replacement for X and breaks a lot of stuff -- just means "sucks to be you devs, now just re-write your desktop for Wayland... in your spare time..."
If you run NVIdia graphics, you are out of luck. If you rely on xinput or xmodmap for configuring your input devices -- "sucks to be you", there is no similar functionality or utilities for Wayland. You just can't configure it to your liking the way you can X (starting to sound a lot like Gnome 3 on...)
Don't get me wrong, I'm not against Wayland at all. I'd love to see it mature, provide the features we have come to expect in X, support the most common high-end graphics cards and provided a much easier port from X than it currently does. But, until it can, don't blow smoke up our skirts about how ready it is for prime-time without being candid about its current short-comings and limitations.
"But this sort of thing feels to us like it's trying to push users towards Wayland, rather than actually attracting anyone."
This is exactly the problem. The folks forcibly pushing Wayland are the same ones forcing hideous, mandatory, touch-centric UI elements upon everybody through GTK, removing any element of user choice through their devout belief that their 'new' way is the only way.
If Wayland as amazing as they say it is then it will stand on its own merit and people will naturally migrate to it as they "wake up" to whatever is apparently so hideously broken in X11. Threats and FUD to try and scare users into switching isn't going to wash.
And for the folks behind Budgie to declare MATE as dead... That's just bollocks. The team behind it is constantly releasing improvements and someone will have to prise it from my cold, dead hands before I stop using it. I don't care whether it's running on X or Wayland so long as Wayland doesn't remove any of the features I use or introduce any instability to my desktop.
I put my money where my mouth it with this one: it's my daily driver and I run my business on it, so I pay a monthly subscription to the project
MATE is an excellent desktop which not only balances functionality and performance but it also gives users huge flexibility with its user interface, which can easily be switched between classic layouts like the original GNOME 2 on which it's based, through to looking and behaving like Windows, Mac and several other desktops.
Sadly, just like XFCE and any desktop either based on GTK or running GTK applications (yes KDE users I'm afraid this affects you too), MATE is now tainted with awful (and probably irreversible) Hamburger Menus and buttons nested inside title bars along with some other UI horrors, which were all forced in by the GTK developers. A major driving reason was because.... at the time, Wayland required "CSD" (Client-Side Decorations) - i.e. the application has to draw its own title bar, scroll bars etc. The sad part is that I'm led to understand that since then Weyland has gained support for Server Side Decorations which basically means that the whole story which led us to our desktop UIs being irreversibly effed up, was actually a lie.
Dear reader, you might be able to tell that I'm a tad miffed about this whole situation. I'd possibly go as far as to say "angry". I'm not angry about "modern" UIs per-se, I'm angry that major players in the desktop Linux ecosystem are carefully stripping away the user choice which has always been one of the biggest selling points of desktop Linux - this could all have been implemented in a way which embraced the spirit of user choice, but instead it has been hard-coded on a wide scale which means folks will be getting these changes whether they like it or not... That's not the spirit of FOSS / Linux, that's the way the likes of Microsoft decides to dump 'TIFKAM' on the world.
If you prefer Unity or KDE, GNOME, Cinnamon, Budgie, XFCE or some other desktop, it's YOUR choice. I'm on my high horse about this because I see this choice being eroded and with the changes being pushed out under the banner of "Wayland is better", don't be surprised if your choice of desktop begins to lose its unique character and features :-(
I hope I'm wrong!
"And for the folks behind Budgie to declare MATE as dead... That's just bollocks."
Indeed. In my travels in and around Silicon Valley (including the Unis in the area), I have never seen anybody using a Budgie desktop in the wild. Not once. MATE is almost common, although Cinnamon seems to have taken over that part of the desktop in recent years.
Don't listen to Marketing, folks! It's their JOB to cajole, nudge, bully and obfuscate you into listening to them, regardless of the suitability of their product for your needs.
There's a lot wrong with X, despite the fact it works.
There's a lot right about Wayland, despite the fact it doesn't work.
That's a little unfair, now is the time to get into Wayland and attempt to influence its direction before it becomes impossible to change. However, there's many things wrong with it :
It's Linux centric, and most of the discussion on the Wayland mailing list assumes a Linux architecture. Unix is more than just Linux and this is frequently ignored.
There are multiple compositors. Windows and other platforms have one compositor, meaning functionality can be enhanced vastly more easily.
If the compositor is based on one of the base standards such as the wlroots library or Weston (probably wlroots, because despite Weston being the 'reference compositor' it doesn't work on much other than Linux).
Let me repeat that again : the *reference* compositor is effectively Linux only. Fuckwits. There's a FreeBSD port in progress and version 8 (i.e. old) supposedly runs on NetBSD.
If functionality is added to a reference compositor library, such as a new protocol addition in wlroots, it does not automatically flow down to compositors based on it. This means that the FreeBSD suggested compositor, Hikari, still doesn't have the protocol extension for sensibly managing multiple monitors, three years after it was written. If you're on FreeBSD, use labwc instead.
On X, the xorg.conf file was standard. On Wayland each compositor has a different config file format. This frequently omits functionality, such as the ability to set up multiple monitors exactly how you want. The functionality is in Weston, for instance, but that's effectively Linux only so you're stuffed on other platforms and need to run programs after the compositor has started to do what you want.
Network transparency is compositor specific, not part of the reference compositor. Colour management hasn't been tied down yet. Lots of other things are in flux.
Wayland has been around for over a decade now and is missing some very basic functionality.
Due to the compositor design, power is being concentrated in the hands of the large desktop environments rather than small window managers. This is also an issue with X - installing an application that expects the ability to create icons, on a window manager that doesn't support that, leads to error messages and other issues. However with Wayland the issues are more noticeable.
If the Wayland ecosystem has any sense they'll coalesce around one compositor that functionality can be bolted on to, and that is tested on multiple platforms for every release. Oh look, a flying pig.
It works - with some effort. If you have the right drivers (generally not a problem these days), the right compositor (probably not an issue if you choose one of the popular ones), and don't need remote access (there are some ways to do this, I don't know the specifics).
We have to accept that although X works well this is in spite of, not because of, its design. There is so much cruft in X it's unreal, and its basic design has a lot of flaws. Years of bodges and accepted conventions means the end user experience is generally good.
Wayland generally has a less broken design, has traction from the major players, and a web browser and productivity suite will in general work Just Fine. However there are a lot of edge cases and modern functionality that are still not covered. At some point it's a good idea to try moving to it, so its future direction can be influenced, and the rough edges minimised.
"At some point it's a good idea to try moving to it,"
I have kept an eye on it since the initial release.
"so its future direction can be influenced"
From what I have seen, only the anointed few are allowed to make acceptable suggestions.
"and the rough edges minimised."
Yeah, good luck with that. I'll stick to X, which works for all my use cases, and has done for nearly 40 years.
Yes, X has problems. It needs an upgrade or replacement. But Wayland is not that replacement no matter what the fanbois tell you.
...deprecated...
I just compiled and installed Aqualung music player,
because it works....
I tried a number of active alternatives which were crap.
It has been forked and even has new dev work
I use bits of software that haven't been advancing for many years,
but they work...
and there's nothing else does the job I want, the way I like doing it.
don't fuck with it if it works
Because there is nothing wrong with using a *nix as a desktop system.
Some would say there are many reasons to be using a *nix as a desktop system.
Properly used, a GUI allows one to compartmentalize many desktops, each of which is targeting a specific task, on a single monitor. (Yes, I can do this with a dumb terminal, but that gets unwieldy when you're switching between more than ten or so virtual terminals.)
So far I can't use it. My laptop didn't come with a switch in the enter key, and I can't figure out how to remap caps lock to enter on Wayland like I can with xmodmap. I've tried gconf and setxkbmap, and neither worked. On-screen keyboards, or at least Onboard, isn't an option either. GNOME's does work, but it's annoying in other ways, and Plasma's just doesn't function (as of Fedora 37).
I will say this isn't the end of the road for X anyway; as far as I know all of the BSDs still use it as their default (or usually only) display protocol. I don't know if it'll be the end of e.g. Plasma on BSD, though.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
