Ubuntu Is Switching to Wayland (omgubuntu.co.uk)

An anonymous reader shares a report: Ubuntu is to ship Wayland in place of X.Org Server by default. Mir, Canonical's home-spun alternative to Wayland, had been billed as the future of Ubuntu's convergence play. But both Unity 8 the convergence dream was recently put out to pasture, meaning this decision was widely expected. It's highly likely that the traditional X.Org Server will, as on Fedora, be included on the disc and accessible from whichever login screen Ubuntu devs opt to use in ubuntu 17.10 onwards. This session will be useful for users whose system experience issues running on Wayland, or who need features and driver support that is only present in the legacy X.Org server session.

Comments

But is Wayland better?

[Anonymous Coward]

As long as Linux can remember where I positioned my monitors after I put the laptop back into the docking station, and as long as I can wayland-over-ssh, and as long as there are performance gains, then I don't care.

I'm sure this post will be littered with "I hate change" type posts where people lament the loss of X for no other reason than passion and nostalgia, and I'll have to dredge through loads of nonsense before someone actually puts together a point-form list of pros and cons comparing Wayland to X

Re:But is Wayland better?

[squiggleslash]

Network transparency. X11 has it. Wayland doesn't. Wayland's devs tend to handwave the problem, either claiming it will somehow be implemented once they work on the other laundry list of things they want first, or claiming it's a niche requirement nobody wants or uses.

On top of that they're doing the #1 thing you're not supposed to do in development: completely rewriting a working system.

X11's main flaw is that it's supposed to be inefficient. It might be, but I've never noticed any significant difference between user interface performance on Ubuntu vs Windows or Mac. I think much of it is "This sub-nanosecond operation that is only called once or twice every frame takes THREE TIMES AS LONG under X11 as it should!" type purism.

I'm not happy about this.

Re:But is Wayland better?

[Anonymous Coward]

X11 is a turd. The future is using javascript and rendering to HTML. Rust, go, python, and all the other important languages can be compiled into javascript which runs everywhere - the browser, the desktop, the server, the phone, chromebooks, and even embedded devices that only have 4 GB of RAM.

Re:But is Wayland better?

[caseih]

How this will be solved in the long run remains to be seen. In the short run, toolkits that support Wayland still support X11. Mainly I'm talking about GTK and Qt. Thus KDE, Gnome, GTK, and Qt apps will all run either on Wayland or X11 without recompiling. So for many people, remoting needs can be accomplished by simply using X11 on Wayland and tunneling X11 over SSH. Simply ssh into your remote machine and run the apps. Locally on wayland things are silky smooth, remotely they still work, though a bit choppier (X11 over ssh isn't fast enough for anything but LAN anyway... I use X2Go for WAN remote X11 stuff.

Of course in the long run if Wayland is successfull the X11 backend bits will languish in the toolkits and this will not be a sustainable future. I think essentially RDP will be adopted as the standard remoting protocol for wayland desktops. This will be used to forward individual apps or whole desktops. RDP is already a lot faster than X11 over ssh, due to the way X11 works and the fact that all modern toolkits essentially just push bitmaps these days anyway.

Before criticizing Wayland and extolling X11's virtues, consider watching this talk by Daniel Stone who was formerly intimately involved with X.org and seems to know hist stuff. He makes a good case for Wayland. https://www.youtube.com/watch?...

Re:But is Wayland better?

[Anonymous Coward]

I am not managing a computer remotely, I am using a computer remotely. Yeah, I ssh in. Always start with command line work - just like I do with a local login.

But then I need some word processing. So I start the word processor on the remote machine, and view the window locally. Common case, do it several times per week. No, I don't want to transfer the file here, run a word processor locally, and then transfer it back. That is hell, especially when there is a large set of dependant files. Figures and whatnot.

Surely this sort of thing can be done with wayland too. Wayland renders to memory and then blits to some display. All we need is to transfer that memory - or at least the changed altered - over the network. Then render on the display at hand.

It is a necessary function, but something that shouldn't be all that hard. I don't expect to run 3D games this way - but word processing used to be fine over an ADSL line to the office 6km away - surely wayland can be made to run over the fiber I have today.

Re:But is Wayland better?

[SumDog]

I too use X11 forwarding over SSH (ssh -Y) all the time. Sometimes I'm at work and want to quickly tag some music that's on my machine at home. I can remote SSH with X11 forwarding and start up easy tag.

It's a pretty common use case among some of us Linux users.

Re:But is Wayland better?

[serviscope_minor]

There is no fundamental benefit with respect to performance as Wayland and modern X clients basically work in the same way when operating locally.

I think what he's referring to is this:

For a "modern" X system (i.e. using a compositing WM---though I don't use one), the event (say, mouse) goes to X, to the compositor program, back to X and then to the focussed program, compared to the old version of X where it goes from X to the program directly. I think Wayland, not having the compositor as a separate program, skips one of those steps.

I've seen that touted as an advantage of Wayland, but FFS, context switches are FAST, and the time taken to process events is down in the microseconds. It shaves a few microseconds of latency off events happening at 10s per second, and for which the minimum perceptible latency is about 0.05 seconds.

The advantage is therefore so minute as to to be irrelevant and touting it as a big advantage is pure FUD. I agree with the GP.

Re:But is Wayland better?

[Rob Y.]

Well, okay. But it sounds like you could figure out how to do the same thing via a remote desktop session. So it boils down to whether Wayland improves things enough for the millions of everyday users to make it worth making users like you figure out another way to do that thing you occasionally do. I don't know enough about Wayland to know that it really will improve things - though I've read reports that it makes the desktop 'feel' smoother. If it makes it easier to get drivers for the latest video cards with fewer bugs, though, I'm all for it.

X remoting was always a good-sounding idea that was implemented in a way that made it not much more efficient in practice than VNC-type remoting. I use remote desktop on my Linux box at home when I need to access my office Windows system from home. I run local stuff on one virtual desktop and RDP on another. The RDP desktop is pretty awful - but useable enough, I guess (since I use it). It'd be better if 'grab all keys' actually grabbed all the keys. But somehow I don't think remoting via X Windows would be any less awful...

Re:But is Wayland better?

[OrangeTide]

X11's network transparency is not terribly useful. It doesn't allow detach/reattach of sessions, so you're forced to exit every program and start them back up again at the new location. While this was very memory efficient back in the 1980's, it's not a very good user experience when dealing with unreliable wireless networks.
And if you've used GTK2 and other widget libraries in the last 10 years, you'll notice their network performance is pretty terrible. Network transparency is kind of useless if none of the software is designed around it. There are too many round-trip messages going on these days to deal with the broken parts of X protocol and in giving fancy user interaction.

VNC is slow, but it's what most people use because it works in a way that is convenient. NoMachine/NX/FreeNX is a technically better alternative, although it's not very popular. (being proprietary doesn't help)

Re:But is Wayland better?

[hublan]

You do realize that most of the folks that started Wayland were originally long-time X hackers, right? There has to be a reason why they gave up trying to get X11 to behave properly, besides "because it's bloated".

Education: https://www.youtube.com/watch?...

Re:But is Wayland better?

[caseih]

Nothing about RDP limits it to full desktops. RDP can remote a single window as well as a full desktop. MS doesn't normally use it that way, but it can be done.

Furthermore, Xrdp can run in rootless mode, if I'm not mistaken. A single X11 apps could connect to it.

Re:But is Wayland better?

[slack_justyb]

I'm going to start where a lot of people don't usually start. The actual people who maintain X11. They hate the code base, they just simply don't want to deal with the tangled mess that it is. Seriously go look at a dependency graph of just the xserver or a slightly higher level view of the state of things. Point, no one wants to maintain this mess. Anyone feeling frisky in doing so is strongly encouraged to do so, but the majority of developers who have worked on this in the heyday have long since left the building. The sheer pool size of people working on X is low and fresh blood in the development pool is best described as anemic. Fewer developers working on one project and more on another project pretty much seals the deal on the direction. Arguments of X being better falls on non-existent ears. You want to talk to an X developer? Head over to Wayland, that's where you'll find a lot of them.

Next in line is that X is ineffective at one of the things that it's suppose to do, draw stuff on your screen. (Not even going to touch multi-monitor, sleep, touch input, etc all which have had extensive hacking to get it working and thus resulting in patches of code with serious bus factor one issues.) X11 lacks pretty much everything we take for granted in a modern GUI. Want anti-alias text? Well X11 doesn't do that. Want the concept of an alpha-channel? Not present in X11. Quite literally, X11 does nothing in the way of anything that say KDE, GNOME, Unity, Cinnamon, or whoever wants. Instead, your chosen toolkit is using a library that builds in memory the bits that need to be drawn and if your xserver supports RENDER, your toolkit just gives a stream of bits over to X11 via that method, and X just forwards it on to either the card or to a compositor, which by the way X11 doesn't have a concept of, hence the reason you need one external to the xserver. At some point someone said, if every toolkit is just building bits by themselves and then having X forward it on, why not just cut out the middle man? Why have this extra layer that we keep having to build ad-hoc extensions for? (RENDER, XDamage, RANDR, XFixes **yes literally an extension to fix stuff but mostlly to turn a lot of old X11 stuff off.) All of these wonderful extensions are in reality short circuiting old cruft in a code-ugly fashion. Add in new complexities being added to video cards, functionality that's difficult to eventually get working, and yeah everyone is ready to put the old girl out to pasture. X11's lack of so many things is a roadblock to tapping your card's fully ability, which is why most of the time we're happily ignorant of all of the by-passing of huge parts of the core of an xserver, with the prolific set of extensions that come automatically built into your distro. (which is why a lot of folks never notice and just think that this is the way X was built, but nothing further from the truth could be said. Try building an xserver from source.)

Now let me move on to your points

Network transparency. X11 has it. Wayland doesn't.

If you are using X11 over ssh, you aren't using X11's network transparency. What you are doing is streaming pixels across ssh, but you aren't using anything remotely looking like core X11 protocol. On the remote side, Cario, Qt, Mutter, or someone is drawing pixels and then that gets wrapped into a generic X11 package and sent to you to open up and then have your computer decide what to do with the newly received pixels. There's no commands like "Window A is currently at location x,y. It has a button at rx, ry relative to the top-left corner of the parent widget, blah blah blah." Nope, it's just "here's pixel one, here's pixel two, here's pixel three..." There's no distinction in X between a button in an application running on a remote server and a picture

So what makes Ubuntu different from Fedora?

[Anonymous Coward]

With Ubuntu's switching to PulseAudio, to systemd, to GNOME 3, and now to Wayland, what is it that makes Ubuntu different from Fedora?

The only difference I can think of is where an installation ISO would be downloaded from, and typing "apt-get" instead of "dnf" to install packages.

Those are really minor differences.

So what's the point of using Ubuntu if it uses the same kernel, the same init system, the same windowing system, the same desktop environment, the same sound system, and pretty much all of the same userland software that Fedora does?

At least things like Unity and Mir made Ubuntu somewhat unique. But now Ubuntu has basically become Fedora with just a different name. Why would anyone even bother using Ubuntu now?

Re:So what makes Ubuntu different from Fedora?

[arth1]

With Ubuntu's switching to PulseAudio, to systemd, to GNOME 3, and now to Wayland, what is it that makes Ubuntu different from Fedora?

Gratuitous privilege escalation enabled by default.

Re:So what makes Ubuntu different from Fedora?

[Eravnrekaree]

First of all, most distros have always had the same window system, X Windows. The reason for this is that since all applications and window managers which are GUI, have to talk to the WIndow System, its important to have standardization around the same API. Otherwise you end up with a MESS of an app that works on one distro not being able to run on another distro or having to run 10 different windowing systems, because each application ends up being tied down to one or the other. You also have to have video hardware drivers and those have to plugin to the window system as well. If we are going to change the core window system, all of the distros had better agree to it or else we will end up with a fractured ecosystem like above. Now, because of X's design of leaving look and feel to the Window Manager, you can completely change the look of the user interface by changing the window manager, which does not affect applications. This is why you can use the same apps regardless of what WM you use. Wayland should and will continue this philosophy.

All of what I said also applies to the sound server, as well, so it was important to standardize around pulseaudio if we are going to have a sound server, which is a good idea. The alternative to a sound server would be to incorporate that kind of functionality into the kernel, its better to have it in a user process rather than to add further complicated code to the kernel, as with X.

As for systemd, rather than to rehash all that here: please read this: http://0pointer.de/blog/projects/the-biggest-myths.html. Basically, systemd is a big improvement over what we had before and the criticisms are mostly myths.

Well

[JWW]

Ubuntu ditched the bad idea that was Unity. Time to ditch the bad idea that is systemd....

Re:X also has stuff!

[serviscope_minor]

Not to mention window managers. One thing we'll sadly lose is the richness of the X window manager ecosystem. That's not a techincal argument against Wayland.

Re:X11 SUCKS

[Billly Gates]

Quotes from the Unix Haters Handbook here.

Let's desconstruct here your arguments X11 myths:

Myth: X Demonstrates the Power of Client/Server Computing

Fact: "The database client/server model (the server machine stores all the data, and the clients beseech it for data) makes sense. The computation client/server model (where the server is a very expensive or experimental supercomputer, and the client is a desktop workstation or portable computer) makes sense. But a graphical client/server model that slies the interface down some arbitrary middle is like Solomon following through with his child-sharing strategy. The legs, heart, and left eye end up on the server, the arms and lungs go to the client, the head is left rolling around on the floor, and blood spurts everywhere.

The fundamental problem with X's notion of client/server is that the proper division of labor between the client and the server can only be decided on an application-by-application basis. Some applications (like a flight simulator) require that all mouse movement be sent to the application. Others need only mouse clicks. Still others need a sophisticated combination of the two, depending on the program's state or the region of the screen where the mouse happens to be. Some programs need to update meters or widgets on the screen every second. Other programs just want to display clocks; the server could just as well do the updating, provided that there was some way to tell it to do so.

The right graphical client/server model is to have an extensible server. Application programs on remote machines can download their own special extension on demand and share libraries in the server. Downloaded code can draw windows, track input eents, provide fast interactive feedback, and minimize network traffic by communicating with the application using a dynamic, high-level protocol.

As an example, imagine a CAD application built on top of such an extensible server. The application could download a program to draw an IC and associate it with a name. From then on, the client could draw the IC anywhere on the screen simply by sending the name and a pair of coordinates. Better yet, the client an download programs and data structures to draw the whole schematic, which are called automatically to refresh and scroll the window, without bothering the client. The user can drag an IC around smoothly, without any network traffic or context switching, and the server sends a single message to the client when the interaction is complete. This makes it possible to run interactive clients over low-speed (that is, slow-bandwidth) communication lines."

Other fun tidbits that made me chuckle

" How to make a 50-MIPS Workstation Run Like a 4.77MHz IBM PC

If the designers of X-Windows built cars, there would be no fewer than five steering wheels hidden about the cockpit, none of which followed the same principles -- but you'd be able to shift gears with your car stereo. Useful feature, that.
- Marus J. Ranum, Digital Equipment Corporation

X-Windows is the Iran-Contra of graphical user interfaces: a tragedy of political compromises, entangled alliances, marketing hype, and just plain greed. X-Windows is to memory as Ronald Reagan was to money. Years of "Voodoo Ergonomics" have resulted in an unprecedented memory deficit of gargantuan proportions. Divisive dependencies, distributed deadlocks, and partisan protocols have tightened gridlocks, aggravated race conditions, and promulgated double standards.

X has had its share of $5,000 toilet seats -- like Sun's Open Look clock tool, which gobbles up 1.4 megabytes of real memory! If you sacrificed all the RAM from 22 Commodore 64s to clock tool, it still wouldn't have enough to tell you the time. Even the vanilla X11R4 "xclock" utility consumed 656K to run. And X's memory usage is increasing."

Dude if there ever was a case f