Update On Wayland and X11 Support

Phoronix was at the Linux Foundation Collaboration Summit and has two articles on the status of Wayland and X11 integration. The second talk was about the current status of Wayland, and its impending release (version 1.0 is due this summer). The developers also have an experimental GNOME-Shell working on Wayland. There's a (kind of shaky) video of this talk (attached, and at youtube for those wanting the html5 version). The first talk (by Keith Packard) covered X11 support on Wayland. It's basically ready to go, but window management is implemented only as a hack right now. The next year could be quite exciting for GNU/Linux and BSD users as distributions begin including Wayland as an alternative to X.org.

Comments

Why?

[bytor4232]

Someone enlighten me, why Wayland?

Re:Why?

[Hatta]

Because it wasn't enough to crapify just the desktop environment.

Re:Why?

[Picass0]

IIRC X11 has a lot of redundancy with stuff that's already in the Linux kernel. Weyland is a compositing window manager that removes those redundancies for better performance and code maintainability. Weyland loads X11 legacy support only as needed on an app by app basis. Benefits include Compiz working much better by bypassing X.

Re:Why?

[Anonymous Coward]

Because Everyone wants shiny eyecandy GUIs made of pixmaps, and X11 is rather more built for vector graphics than pixmaps, particularly where hardware support (drawing to textures, compositing, etc.) is concerned. There's all manner of hacks to make it work on X11, but one can expect better performance and easier programming with a dedicated pixmap-based rendering server designed for modern graphics cards.

Whether the benefit is worth the cost of a new platform, and of some specific design decisions in Wayland, is not as obvious as proponents like to think, particularly if, like me, you prefer to avoid the eyecandy shit that Everyone must have.

Re:Why?

[localman57]

They should get you to write their web site. I clicked the link, and after reading the summary on the first page, I still didn't have any idea WTF it was for...

Re:Why?

[jedidiah]

Compiz is fairly worthless when compared to more advanced video driver features and the ability to remain interoperable with other Unixen.

If I wanted MacOS, I would dust of my Mac and switch it on.

Re:Why?

[Microlith]

Then use TWM. Or hush and go use your dusty old Mac. Some people want eyecandy and the ability to have a dynamic desktop without feeding money into Apple's DRM-happy coffers.

Re:Why?

[Hatta]

Benefits include Compiz working much better by bypassing X.

Those who would surrender basic functionality for eye candy deserve neither.

Re:Why?

[Anonymous Coward]

Instead of compiz one can say "any compositor whatsoever," where you should understand that virtually everyone using X on Linux currently uses a compositor.

Compiz does not implement MacOS.

This is a long-time pain point but it is at a low enough level that you might totally fail to understand what it's for if your understanding of X is that it provides a dock or panel

Re:Why?

[makomk]

Downsides include not having any kind of network transparency or remote desktop support. Even supporting something like VNC on top of it would require a lot of internal changes and probably kernel-level support code too, and the developers basically consider this Someone Else's Problem.

Re:Why?

[zdzichu]

Thats OK, apparently Wayland is not for you. It is for distribution makers to integrate into OS. So dont worry, you should not care about Wayland. Leave it to technically knowledgable people to decide when it should be made available to casual users.

Re:Why?

[Microlith]

He won't be able to use twm if you wayland faggots have anything to say about it. Can't you keep your stupid shit corralled in arch linux?

Then step the fuck up and maintain X11 and the X11 compatibility layers for the toolkits that are out there. Don't tell other people they can't move on to newer things because it would inconvenience your lazy ass.

Re:Why?

[Microlith]

By your logic we should be happy with TWM and a bunch of X Terminals. Anything beyond that is purely eye candy, of course. Especially any sort of OpenGL, which doesn't work over the network.

Re:Why?

[Anonymous Coward]

That's not the point. There are people who WANT to understand it, and getting information from the source is usually better than second hand. That's the GP's point.

Re:Why?

[Anonymous Coward]

Then step the fuck up and maintain X11 and the X11 compatibility layers for the toolkits that are out there. Don't tell other people they can't move on to newer things because it would inconvenience your lazy ass.

Good luck moving on when the hardware manufacturers that matter have already said they have no plans to support Wayland. Oh and good luck bringing people in with this attitude.

Re:Why?

[Anonymous Coward]

I have the distinct feeling all of this has more to do with cellphones and tablets than overall graphics support for linux.

Now I'm not arguing that isn't a valid way to go, especially has linux exploded into those markets, not so much with tablets yet, but as a user whose primary desktop OS is linux, compiz does nothing for me either for work, or fun. It was great for 15 minutes when I set it up to see how it worked, but that was it. The actual usefullness of compiz, in my particular working environment of managing IT infrastructure, is close to if not 0.

Guess I'm still missing the bigger picture, even though I've been following this since, oh, 2009!

Re:Why?

[Gothmolly]

You are correct - anything else IS eye candy. And many of us ARE happy with TWM and X Terminals - because those are the tools we need to do our jobs.

Re:Why?

[cupantae]

As I understand it, X11 is horribly out of date.
It should be clear to anyone that there are some stone-age omissions in X:
- Basic transparency is either an ugly hack that glitches all the time (xcompmgr) or else a fairly extravagant system like compiz.
- Smooth fonts is an extra.
- Multiple displays is an extra.
- There are serious security flaws in areas such as screen locking. See this great series of points by the author of xscreensaver.
Then, there's all this support for legacy technologies that don't affect 90% of Linux/BSD systems, let alone the potential general market.
All in all, X11 has fundamental weaknesses, doesn't reflect modern usage and is really too big to fix. Legacy support and compatibility are so important that problems can't be fixed.

I haven't been following Wayland too closely, but my understanding is that it will address these issues as fundamentals of the system. I hope it does address each of these issues, because they are important. Performance could certainly improve on X. Let's hope it does with Wayland.
I'm not that familiar with how display servers work either, so correct me if I'm wrong about anything. No need to rage, I'm not trollling anyone.

My prediction is that it will become fashionable to whine about Wayland, lots of people will resist it for a while, but in the end it will be the most suitable alternative and only the truly stubborn and those who need X for some reason will avoid it. See the history of PulseAudio for reference.

Sorry about the identical AC post below. I forgot to sign in.

Re:Why?

[Anonymous Coward]

Hey, guys, it's called evolution. We keep figuring out how to do things better and better (for the most part).

Go fast or get out of the fast lane.

Re:Why?

[TheGratefulNet]

you've just described me. I run fvwm 1.4 and I open a term window (whatever is the current default from the distro I have) and I run things. env DISPLAY has stuff and if I run an X app, a window comes up. the window mgr lets me move windows and iconify them. what the hell else do you want/need?

I believe in things being as lean as possible and having to prove or justify any excess. since the old DECwindows days, I've run essentially the same style of window mgr and 'xterms' work basically the same as they did 30yrs ago. you type in them, you can mouse in/out of them, you can scroll them. the paradigm has not changed, really, at all, in all these years.

I cannot justify a 'desktop'. I run a window mgr and windows come up as terms or apps and that's that. runs very fast and bug-free and stays up for months (and years) at a time.

I don't quite get the need to have to add complexity to what does not need it.

(oblig GOML)

Re:Why?

[tibit]

Good luck running any modern CAD system without hardware accelerated 3D. X terminals are useless for much of 3D CAD work. The services available on X screens don't offer even the most basic primitives needed to efficiently draw 3D content. To draw quite rudimentary 3D stuff you at least need 2D triangles with interpolated color (gradients). You'd think XRender would offer that, but sadly it doesn't. Never mind Z buffering, texture mapping on triangles, etc.

Re:Why?

[Anonymous Coward]

I did read it and basically understood it, but agree entirely they are writing like a bunch of arrogant nerds trying to look brilliant. I've lunched with a fair number of Nobel Laureates over the years -- my take away point from those hours is this -- really smart people understand the importance of starting with a simple explanation.

Re:Why?

[TheGratefulNet]

con't confuse 'better' with 'uses lots of new libraries and is so kewl!'

I've lived decades (literally) without desktop clutter. clutter in processes, in screen RE, in library dependance and complex building procedures. you kind of have to have a desktop on the commercial OSs like win and mac, but on linux, you can live really well without any desktop at all. I'm not talking about server mode, either; but a user having a display, mouse and doing local things on it but just not incurring the overhead of what is today called 'desktop'.

as fast as cpus are, they are so much snappier when a zillion other procs dont' have to chat with each other and compete for cpu time.

its not evolution, its complexity because some people seek that out and want to add more. go figure, but a lot of 'engineering' is done for no good reason at all other than boredom or the challenge or the misguided idea that we need to add more complexity and be 'modern'.

some tools just don't need to be re-invented. many mechanical tools have not gone thru change in centuries. why do we ASSUME that software HAS to 'evolve' and change? that's absurd.

Re:Why?

[Anonymous Coward]

IIRC X11 has a lot of redundancy with stuff that's already in the Linux kernel.

Great. How about those of us that run the X server on Mac OS X and BSD boxen? Or run clients from Solaris and AIX servers?

X11 has many deficiencies, but I loved being able to SSH into a box and just run a program without having to futz around with VNC/NX servers and clients: things just showed up like any other program on my desktop. It could get laggy at times, but it was nice to be able to run NetWorker consoles on IRIX machines in another city over the corporate WAN.

Hopefully, if anything comes after X11, it will have similar remove capabilities.

Re:Why?

[Anomalyst]

How does wearing a pair of khaki slacks and penny loafers correlate with ones candidacy for software availability?

Re:Why?

[Anonymous Coward]

Total and utter bullshit. OS X supports X11. Windows can support X11 via small, free add on programs. The same can and will be done for weyland.

Re:Why?

[Anonymous Coward]

One problem with X.org is that it has been stagnating. The multiple monitor support has actually gotten worse over the years if you can believe it.

Windows and OS X can seamlessly blend all my monitors and graphics cards together, all running at full acceleration, and all features available. X.org can't do it and has to rely on hacks like Xinerama that break all sorts of stuff so that you end up with a crippled machine.

Then there is stuff like resolution switching, etc. that are all hacks in X.org.

Now it's true all of this stuff could be fixed in X.org if someone would just sit down and do it but the people with the time are kids and kids want new and shiny. Plus they would need someone with skill and experience to do the design and planning but all of those type of people are stuck in the past.

Re:Why?

[Mad Merlin]

Especially any sort of OpenGL, which doesn't work over the network.

Wrong! OpenGL works over the network, in fact it was designed to.

Re:Why?

[Anonymous Coward]

Thing is, X on weyland will be relatively trivial to implement, so there's really no reason to fear it.

Re:Why?

[loufoque]

OpenGL can work over the network, but that support is just not well integrated in linux distributions.

Re:Why?

[manicpop]

There's a lot in Compiz that is "useless" eye candy, sure... but there are also a few task switchers and window organizers that I find useful.

Re:Why?

[nine-times]

Why should I care about "basic functionality" that I don't use?

Re:Why?

[Anonymous Coward]

No, it *is* for us.
It is for us to muck around endlessly trying to make the &^%$#@ thing work. Just like Pulseaudio.

The arrogant tone of your answer leads me to believe that you are part of the problem^H^H^H project.

Re:Why?

[cupantae]

Unixen

Germanic plural for Unix?
I am very much in favour of this.

Re:Why?

[cupantae]

Could you elaborate on what the basic functionality being surrendered is?

Re:Why?

[KiloByte]

... Compiz ...

Those who would surrender basic functionality for eye candy deserve neither.

Compiz is pretty damn usable, with both XFCE and Gnome2. I guess you're thinking about Gnome3.

For example, on ancient machines with a semi-decent graphics card, Compiz with superfluous eyecandy turned off runs circles around regular window managers, because windows don't have to be repainted over and over every time you switch windows and/or desktops. You can see the contents, read something, focus your eyes, move the mouse to whatever you want to click, etc, giving the program inside a couple of seconds to swap itself in/etc, making you not even notice that the system is thrashing. On a regular window manager, you'll instead see an empty broken window.

Re:Why?

[Anonymous Coward]

"Good luck moving on when the hardware manufacturers that matter have already said they have no plans to support Wayland. Oh and good luck bringing people in with this attitude."

Um, the slides in the presentation were from INTEL.

I guess they don't matter?

Re:Why?

[Anonymous Coward]

Frickin nerd.

Re:Why?

[cupantae]

why do we ASSUME that software HAS to 'evolve' and change? that's absurd.

A better question might be why you assume you know what other people are thinking.
Your comment is full of strawmen, and really doesn't add much to the discussion.

Re:Why?

[Anonymous Coward]

Hmmmm curious. I just popped open my task manager, I see 96 processes and 0% cpu usage on an older core 2 duo. In fact it only ever went up to 1% occasionally as the task manager refreshed the list. Doesn't seem like they are competing with each other at all. Maybe it's time to join the rest of us in 2012?

Re:Why?

[Aguazul2]

Do you think that graphics on Linux is in a good state? Do you like tearing on all your animations? Are you happy about all your graphics going through layers and layers of libraries to get to the screen? To me, brought up on the Amiga, Linux graphics is just plain terrible. If Wayland can put all the dross of X into libraries or a daemon that start only when necessary, I'd be very very happy. I'd also like to be able to run my terminals full-screen with 8-bit graphics and my desktop at 32-bit, but that may be asking for too much. The console hasn't been the same since the demise of SVGATextMode.

Re:Why?

[Alex Belits]

Modern X servers support accelerated 3D just fine.
There are even multiple ways to implement it, with 3D being implemented either local to application or local to the display.

Re:Why?

[Alex Belits]

The above is completely wrong.

Re:Why?

[Alex Belits]

Oh yes, it is!

Re:Why?

[Alex Belits]

The way how OS X and Windows "support" X11, I wouldn't wish on anything to be supported anywhere.

Re:Why?

[Alex Belits]

Congratulation, you are severely misinformed, and everything you said is wrong.

X11 has a lot of cruft that is no longer used because it was moved to different layers -- for example, it supports display-side (server-side in X terminology) bitmap font rendering while modern toolkits use application-side (client-side for X) rendering that implements antialiased fonts. While this is a design flaw, the solution is to either not to use the old system (what all modern software already does) or add a new font rendering system if it is warranted (what may be a good idea but hardly a priority for anything). Same applies to most of X cruft.

Separately from anything stands X protocol with its excessive "chattiness". But all "alternatives to X" propose to drop the protocol completely and instead use direct function calls -- the chattiest thing that was ever created! After such a change, it would be impossible to add any remote functionality other than crude "hey, I am drew something, let's update the whole bitmap, and oh, compositing is already applied using MY screen layout!" mechanism. It's a huge step back.

Re:Why?

[Anonymous Coward]

Unless they simplify mutiseat configuration I don't give a rats ass.

Re:Why?

[Alex Belits]

Don't copy-paste other posts, especially ones that are completely wrong.

Re:Why?

[Alex Belits]

You are an idiot.
"Seamless blending" is the part that works "perfectly" in X already. The problem is with configurations that treat multiple monitors as separate screens, or combine incompatible hardware each with its own proprietary drivers. X users simply have higher demands and longer history of sophisticated multi-monitor suppport than Windows users.

Re:Why?

[Anonymous Coward]

Bingo! Welcome to the "New" Linux, like new Coke and just as much made of suck, where the buggier and shittier it is the better it is as it makes those unwashed look at Linux as a buggy pile of ass and only those who prove themselves "worthy" by fighting with the software for days or weeks to make it work deserve the "blessings" of the great RMS. See PukeAudio as a recent example.

Re:Why?

[tibit]

3D local to the application is otherwise known as VNC. Render however and whatever, then ship the bitmap to the screen. You don't need any of X infrastructure for that. Even if you ignore the overhead of shipping the bitmap, there's still the overhead of getting the stuff rendered by the local GPU from its memory to main RAM. Such overhead is not present when the GPU renders directly to the framebuffer.

3D local to the display requires that the display and the application (X server and client) are on the same host. Otherwise the OpenGL calls are routed through a remote procedure call system. This would be fine if not for the fact that the OpenGL API horribly suffers from network latency if it's being executed using RPC. It was never designed to be used that way. Never mind that in dense datasets (say complex CAD assemblies), the geometry may well be orders of magnitude larger than the rendered framebuffer, it then makes no sense to do remote rendering on the X screen; doing it locally and shipping the resulting bitmap will take less network bandwidth.

Re:Why?

[TheRealGrogan]

The day I can't have a simple system is the day I stop using this. It absolutely sickens me what the major distros are doing these days. That's not "freedom"

The day I can't run a kernel, a minimal set of daemons, a shell, an X11 window manager started only when I choose to, and a bunch of scripts that tie it all together is the day I might as well just use Windows.

I already have to keep that around to boot into for my games anyway, since development of Linux ports has lost momentum. The last good game I bought for Linux was Quake 4. Unreal Tournament 3 was a wake up call for me... they lied and procrastinated. It was when I resigned myself to buying Windows games.

I'm afraid I'm skeptical about this Wayland. I'm not sure I want things to change. We'll see when the time comes.

Re:Why?

[timothyf]

Why, exactly, should we take your word for it?

Re:Why?

[shish]

I believe in things being as lean as possible and having to prove or justify any excess

This is pretty much the reason *for* Wayland -- whether you use Xorg or Wayland you have support for all the excess eyecandy (not that you have to use it), but Xorg is an exponentially growing pile of add-ons while Wayland has the features planned from day one, plus it's even smaller than that, because it lets the kernel do the heavy lifting of driver support.

Re:Why?

[Alex Belits]

No. 3D local to application is VirtualGL, and it works JUST FINE.

Re:Why?

[Anonymous Coward]

Not to a lot of us.

Re:Why?

[Alex Belits]

Because I used it for five years to run VariCAD on a server with no local video acceleration, displaying on client with Nvidia card, both running Debian.

Re:Why?

[Anonymous Coward]

Network transparency is the big one. I know, not everyone runs X applications from other machines regularly, but I do, and I'd miss that quite a bit.

Re:Why?

[allo]

german plural would be Unixe

Re:Why?

[Anonymous Coward]

german plural would be Unixe

Irony: A Grammar Nazi who doesn't know the difference between 'German' and 'Germanic'.

Re:Why?

[spitzak]

No, wayland uses a socket to communicate between the client and the compositor. It is not function calls. Many calls to the client library can be (and are) consolidated into a single block message sent to the compositor. Wayland was designed for this, by avoiding return values from the calls, so they can be asynchronous (this is where Windows fails badly, and Xlib (though not X itself) also did in many cases, because they failed to realize that some calls they thought would be rarely used and thus ok to stick a return value on are used extensively).

Re:Why?

[PreparationH67]

I recall an old tale about a pot and a kettle. Assuming you are the same Anonymous Coward. Cunt.

Re:Why?

[Seta]

I think anyone that has watched the creation of new X extensions and deprecation of old ones (usually too slowly) over the years can agree that there's a lot of cruft in X. Nobody is denying that. If you wanted to, you could probably get Keith Packard himself to point you at some cruft that they're trying to remove. They've been trying to clean up for years.

That being said though, I can hardly read your last few sentences as anything but exaggerated nonsense. Wayland, at the very least, has had a system for tracking surface damage for a while now. The notion that the compositor will require a complete surface copy for some arbitrary reason that you didn't specify is rubbish. Additionally, in a remote access context, insinuating that a compete screen copy would be required each time something has changed (for whatever arbitrary reason you also didn't specify) is also rubbish, especially since you already have every client notifying you of surface damage.

Step back for a second, Wayland is a *protocol* and Weston is a work in progress for crying out loud. They haven't even finished hammering out the details of local screens and clients yet, let alone stapling together a recommendation for remote desktop access. The fact that you're trying to draw a conclusion at all at this point is ridiculous. Heck, input is still being pulled in raw from evdev devices isn't it?

The grandparent may have not been so well versed in his Xorg, but that doesn't really warrant the kind of knee-jerk, half-assumed, hostile response you just gave.

Re:Why?

[Seta]

Aye, this is the problem that XCB tried to solve, but adoption was stalled, maintainers were few, and, for a while there, the closest thing you could get to documentation was the Doxygen API reference and the protocol/extension documentation. There were too few people that thought XCB mattered, so it had somewhat fallen by the wayside. Even today, updates to Xlib far outpace additions to the XCB library. I do believe that Xinput is still behind quite a bit.

Re:Why?

[Anonymous Coward]

The day I can't have a simple system is the day I stop using this. It absolutely sickens me what the major distros are doing these days. That's not "freedom"

Don't be a jerk. That is "freedom." You can do what you want, and no one is holding you back. Just don't make others do it for you. You are "free" to build your own distro, or program your own OS. You may not like it, but not liking something is not the same as not "freedom." If you go back to Windows, you legally can't build your own distro. That's not "freedom."

Re:Why?

[dkf]

By your logic we should be happy with TWM and a bunch of X Terminals. Anything beyond that is purely eye candy, of course.

Ah, I remember TWM. I used that for years, and it worked great. (Well, it did once you reprogrammed all the event bindings; the default set were so awful that coding your own was a rite of passage.) I also remember people at the time moaning about how bloated and full of eye candy it was...

Re:Why?

[Alex Belits]

This all can be done by extending X without breaking it.

Re:Why?

[Alex Belits]

That being said though, I can hardly read your last few sentences as anything but exaggerated nonsense. Wayland, at the very least, has had a system for tracking surface damage for a while now.

And without having this accessible over the network, network transparency goes to Hell.

The notion that the compositor will require a complete surface copy for some arbitrary reason that you didn't specify is rubbish.

VNC operates in terms of the root window, and thus is completely unusable for this purpose. If Wayland developers designed their own remote protocol (even if it was primarily bitmap-based) and window manager interface, it would at least make their efforts somehow legitimate -- maybe even add support for X on top of this for compatibility. But now it's "we will draw pretty pictures, dirty people who need remote applications should use VNC!" -- that's completely unacceptable.

Additionally, in a remote access context, insinuating that a compete screen copy would be required each time something has changed (for whatever arbitrary reason you also didn't specify) is also rubbish, especially since you already have every client notifying you of surface damage.

I said no such thing.

Step back for a second, Wayland is a *protocol* and Weston is a work in progress for crying out loud.

it's not a protocol if you can't use it across the network. And even if it was, there would be still no reason to use it if it is not at least as efficient as X when used for such purpose. There is absolutely no reason for making a WORSE system.

The fact that you're trying to draw a conclusion at all at this point is ridiculous. Heck, input is still being pulled in raw from evdev devices isn't it?

You are ignorant. X uses input drivers and its own event system. There wasn't even evdev subsystem for most of the time X existed (and there still isn't on systems other than Linux).

Re:Why?

[Anonymous Coward]

X11 was terribly out of date in the 90's

Re:Why?

[Teun]

some tools just don't need to be re-invented. many mechanical tools have not gone thru change in centuries. why do we ASSUME that software HAS to 'evolve' and change? that's absurd.

Such evolution also happened to the many mechanical tools you reference, it was around the same period software is now in, it's infancy.

Re:Why?

[Daniel Phillips]

How about my word? I just tried glxgears running over ssh, it works fine, but slow. At 1920x1200 it's about .5 frames/second. I do not doubt it can be speed up considerably. This is over a dodgy wifi link.

Re:Why?

[CharlyFoxtrot]

Not really "new" it's been that way forever. It's basically like that Kurt Vonnegut quote : "Beware of the man who works hard to learn something, learns it, and finds himself no wiser than before. He is full of murderous resentment of people who are ignorant without having come by their ignorance the hard way."

Re:Why?

[jbolden]

Wayland offers X11 support, Linux post Wayland will be when it switches roughly as interoperable as it is now. The problem will be in the other direction as apps start to migrate to Wayland.

Re:Why?

[CharlyFoxtrot]

Dutch plural would be "Unixen" and that's a Germanic language too.

Re:Why?

[jbolden]

Wayland is actually simpler than X. That being said, I think it will be a very very long time, like a generation minimum, till X.org doesn't have an X for Linux. As far as games Wayland is a pure advantage.

Re:Why?

[jbolden]

VNC is going to work the same way it does with Windows or MacOSX which doesn't involve low level support. VNC is an alternative to Network transparency.

Re:Why?

[jbolden]

Huh? X11 works fine with OSX. You get a window manager that speaks Aqua so even cut and paste works. Heck if some application wanted to they could even implement a remote drag and drop. You get a access to apps and a reasonable menu. Finally of course you can run a Unix Window manager in full screen mode.

So what exactly is the problem?

Re:Why?

[Daniel Phillips]

I would have to call your screed uninformed. I am running accelerated 3D right at the moment, using the open source Radeon driver on GLX. No screen tearing, I assure you. In this configuration, OpenGL does not go through "layers and layers", it uploads 3D data directly to the card via DMA like any other high performance 3D setup.

Re:Why?

[jbolden]

You are absolutely right that stuff could be added to X and compatibility maintained if that were a high priority. And there is no reason the invention of Wayland prevents that from happening. X will be free to mature if there remains an application market that needs network transparency.

Wayland is mainly a step back, in that it discards network transparency but at this point even phones do their own rendering. Server side renders mostly don't seem to matter and for client server the web seems to good enough.

Re:Why?

[jbolden]

For applications that are used remotely they just stay on X11. Wayland will have an X11 the way OSX does now. For applications written for Wayland yeah you likely will need a VNC style solution.

Re:Why?

[Anonymous Coward]

Wrong! OpenGL works over the network, in fact it was designed to.

Wrong! You're interpreting the fact that it has been done as "it was designed for it". But it wasn't. The GL API originally did not even acknowledge the existence of such a thing as a network, being completely oriented towards local graphics. And although I say "originally", I don't think it has any network support today either.

You can take a lot of APIs with no notion of networking and make them function over a network connection by inserting a man in the middle between the application and API implementation, adding a few sideband APIs for network connection setup and maintenance. GL-over-networks is an example of that.

I'd also like to note that common early GL usage patterns were almost actively hostile to networking. Immediate mode GL is terrible for network efficiency, because the program submits one vertex to the graphics pipeline per GL API call. This was a supported and common design pattern, and that puts a lot of holes in the idea that GL was designed for the network. If it really was, the original API would've been oriented towards batch submission of data, and hostile to the 1-vertex-at-a-time model.

Re:Why?

[sl3xd]

VNC is going to work the same way it does with Windows or MacOSX which doesn't involve low level support. VNC is an alternative to Network transparency.

Excrement is an alternative to food. Gotcha.

The only redeemable thing about VNC is that it's better than nothing. That's a pitifully low bar. VNC is marginally better than having somebody take a picture of their screen with their camera, and text message it to you.

VNC has a horrible user experience, it's horrifically slow, and universally ugly - the only way to get it to "work" even half-reasonably is to compress the bitmaps so severely that the artifacts make text illegible.

If you increase the bandwidth used to the point VNC can operate with anything other than "please kill me" as its user experience, X11 will run rings around it.

If I had a dime for every time I've been bitten by not being able to copy & paste because of someone's VNC implementations, I'd be able to buy a yacht.

There's no reason that *real* network transparency shouldn't be provided.

X is long in the tooth; the X protocol can certainly be improved upon. Throwing out network transparency is a huge step backwards.

Re:Why?

[Daniel Phillips]

Downsides include not having any kind of network transparency or remote desktop support. Even supporting something like VNC on top of it would require a lot of internal changes and probably kernel-level support code too, and the developers basically consider this Someone Else's Problem.

Network transparency is essential and it is most definitely the Wayland project's problem to demonstrate clearly how it is to be achieved.

Posted from an Konqueror session running on my home server over a Wifi link to my desktop. Works great. Is essential. Until this works in Wayland, Wayland is broken and should not be deployed except to developers.

Re:Why?

[jbolden]

The claim of the GP is that VNC would work at all. Your claim is that genuine network transparency is better than VNC. I'm neutral. I think X11 needs to have setting for higher latency and better handling of standard libraries.

But you can't do the sorts of low level hardware things that GDI and Quartz do and be network transparent. It is a real choice between divergent feature sets. It is not a step backwards, it is a step in a different direction.

Wayland still offers X11, so applications that need X11 just don't move over to the new standard. But the fact is that for 2 decades Linux has offered network transparency and there is far less of it used now then when I used it all the time in the early-mid 1990s. Client server with client side rendering has been a vastly more effective paradigm.

Re:Why?

[Aaron B Lingwood]

Someone enlighten me, why Wayland?

There are many reasons to adopt Wayland.
The reason I have been following Wayland so closely is to ditch X.
X is the Linux Desktop's largest security hole.

::further reading::

Re:Why?

[Seta]

And without having this accessible over the network, network transparency goes to Hell.

Again, we're talking about a protocol for a combined compositor/window manager (they merged them, remember?) talking to clients on the local machine. All they are currently making is A.) the Wayland protocol, B.) A combined windowing system and compositor to demo it with and C.) some demo clients. They aren't trying to make some swiss army protocol that magically solves the problems of local displays and remote displays at the same time. They specifically said that's not what they want to do in the first place. Wayland, in any form, couldn't do that and X11 doesn't do it either. If you think that X11 is some kind of godsend then fine, I personally don't care, but they're trying to make something that performs very well locally right now and X11 is mediocre at doing both of those jobs. Just mediocre, and it won't be doing any better in it's current form. Even X11 forwarding, your most cherished toy, slows down horrendously when having to deal with a lot of difficult to compress image data in addition to dealing with input events. I try to drag a rectangular object around inside another window and... what's that? The rectangle can't keep up with the cursor? It's so slow that I might as well just use Spice, a protocol made for remote desktop (and remote desktop alone) network communication instead? My gosh! Who would have thought!

Now, that being said, now that both compositing and window management duty has been combined and simplified, isn't this where you would want to start writing that remote desktop implementation? One that doesn't tie you to tens of years of legacy that forgoes latency hiding and efficiency (Xlib) or one that's so bare metal that you cry every time you look at how much code you've had to write to do anything (XCB)? One that was written for remote desktop sharing in the first place and not a rehash of the same mediocre mess we've been dealing with untangling for years?

VNC operates in terms of the root window, and thus is completely unusable for this purpose. If Wayland developers designed their own remote protocol (even if it was primarily bitmap-based) and window manager interface, it would at least make their efforts somehow legitimate

So if they don't re-invent not only the wheel, but the drive shaft as well, you won't be satisfied? They aren't trying to do the whole job for you. We (the programmers) are supposed to be doing something besides sitting in the peanut gallery and yelling "do more for us! we don't care about your goals!".

-- maybe even add support for X on top of this for compatibility. But now it's "we will draw pretty pictures, dirty people who need remote applications should use VNC!" -- that's completely unacceptable.

Wait, so you missed the part where they're working to run X within Wayland so that you can still run your X applications if you'd like and still be able to use your X11 forwarding if you so pleased? While they aren't building it into the protocol (again, there's no reason they should), isn't that exactly what you just asked for? I know you aren't supposed to read the article here on Slashdot, but I didn't think you'd go so far as to not read up on the topic at all. That is impressive. You've earned your low number.

I said no such thing.

Exactly. You said nothing at all about what actually triggered the ficticious event you referenced in your fake quote. Not a damn thing. You gave absolutely no context. You just made a vague statement, probably in hopes that nobody would call you on it.

it's not a protocol if you can't use it across the network.

...did you actually say this statement? Wow, never thought I'd hear that. I suppose a sandwich isn't a sandwich if it's not made for eating then? Or doesn't use bread? Please consult a d

Re:Why?

[sl3xd]

Wayland still offers X11, so applications that need X11 just don't move over to the new standard.

It has nothing to do with the *application* needing X11 transparency. It's what the *user* needs. It doesn't matter that X11 support exists if the app I need to use over the network is wayland only.

It's virtually certain that many Linux distributions (especially ubuntu) will have have compiled out X11 entirely - meaning that Qt at GTK applications that should have no problem using X11 (and operating over the network transparently) will be entirely unable to.

Re:Why?

[jbolden]

I'd doubt that any distributions are going to compile out X11 support from Wayland. Why wouldn't they want their distribution to be able to run apps that require X11 that aren't designed for Wayland? Maybe in something like 20 years if essentially all apps have gone to Wayland,... But even then I doubt it. There is always going to be the Linux server culture with remote management and network transparency and running the X11 client is cheap and....

However, I suspect that QT and GTK will within 5 years have calls that offer Wayland only features and the applications will be compiled for those features inside of 10 years and...

So it will come down to applications IMHO.

Re:Why?

[tibit]

What I mean by "otherwise known as VNC" is that it is local, and you're transferring bitmaps to the display. It doesn't then matter how the rendering is done (GPU, CPU, black magic, or otherwise).

Re:Why?

[Alex Belits]

So if they don't re-invent not only the wheel, but the drive shaft as well, you won't be satisfied?

It doesn't really matter how they implement it as long as it works. There are good engineering practices. This feature-stripping and breaking existing use models isn't one of them.

Wait, so you missed the part where they're working to run X within Wayland so that you can still run your X applications if you'd like and still be able to use your X11 forwarding if you so pleased?

I don't care about X applications per se. I care about making it impossible to produce an application that will not work remotely. With X11 I have this -- anything that works locally will also work across the network as long as latency is reasonable for its use. With "oh, we will support X11 because we heard, someone wrote X11 server in Java, so just use that", I won't.

Re:Why?

[Alex Belits]

Except the whole point Wayland developers are claiming is to avoid supporting X in any form, relegating it to a ghetto of "compatibility mode" that new application won't use. That makes Wayland applications non-network-transparent except through crude kludges that have nothing to do with X, don't work with window management, and lack other features that X currenly supports. I don't think, they would even think of implementing X window management for Wayland applications, so window management would be fundamentally broken.

Re:Why?

[TheRealGrogan]

I do exercise my freedom. I don't use that shit. If I had to, I'd go back to Windows because it would be pointless. I've been using this stuff long enough to know what my options are.

I do put in the time to have exactly, and only what I want. ... and I probably shouldn't have said "the major distros" though there is a lot that I don't like about some of the most mainstream ones (bloat, complexity and dependencies on things I mostly wouldn't want etc.). I'm thinking mostly about the one that has the most marketing force and the one that new users are most likely to encounter. I've often thought "I don't want people to think this is what "Linux" is like".

Re:Why?

[ogdenk]

LOL I remember early versions of Slackware in 1994 (maybe 95?) requiring intense muckery just to get it to f**kin boot much less get drives partitioned and ready to go.

I remember installing NetBSD 1.2 on some architechtures requiring booting a kernel with a dismal little ramdisk (or sometimes via the network requiring tftp and NFS) and hand unpacking the archives that comprised your root filesystem AFTER you pulled your hair out understanding the funky disk partitioning scheme along with various quirks for that architecture. Then manually hand-editing half of the /etc directory with vi to configure the system. Then getting pissed and realizing you had to create files like /etc/mygate to set a default gateway on boot. After 2 days of setting it by hand with ifconfig. Then finally figuring out how to properly make the HD bootable so you can quit f**king netbooting just to get a bootloader prompt. When dual-head meant plugging in another VT220 until X worked.

OH! And no adduser scripts..... you had to create your own users by hand, including their home directories, populating it with login scripts, editing files in /etc to define user and group, changing permissions and finally.......finally.....setting a password and logging in to test and see where you made a typo.

I don't have much FOND nostalgia of the early free UNIX landscape. It was a learning experience and you felt triumphant getting a working system running and doing something useful and it would stay that way for AGES......was so lightweight and fast........ but it was grueling hell getting to that mountaintop.

In the end, sometimes a little initial hell is worth the real-world game-changing progress that ensues. Most of you would have thought that we were stupid going through that much effort now but if we didn't, you think Linux or BSD would be here today? You drive progress as much as developers. Submit bug reports.

Re:Why?

[justforgetme]

In the end, sometimes a little initial hell is worth the real-world game-changing progress that ensues.

Yep. Completely, absolutely, irrefutably correct. The only reason why the FOSS culture and the free desktop (as in the free OS) have gone such a long way since the early days of slackware is because there were people who wanted to get their hands dirty and muck about for days on a toy OS in order to get somewhere (usually the goal was understanding, more than achieving an actual OS). Those people later on became integral parts of the development progress, highlighting developers mistakes and decisive disadvantages of various decisions made and providing solutions and custom scripts to automate stuff in a non breaking way. Not to forget the part well versed early adopters played in community support.

The truth is that if FOSS developers were to wait for a perfect product until they released we would still be trying to init the GNU userland on HURD, or not even that one since almost the whole body of the GNU userland came out as incomplete alphas at first and those alphas then drove interest and development.

Wayland development seems have to made leaps since the last time I looked at it last year. Hopefully in the next 18 months we will be actually getting some workable prototype OS with it.

Re:Why?

[ogdenk]

The only reason why the FOSS culture and the free desktop (as in the free OS) have gone such a long way since the early days of slackware is because there were people who wanted to get their hands dirty and muck about for days on a toy OS in order to get somewhere (usually the goal was understanding, more than achieving an actual OS).

I just wanted to run UNIX on a low-end machine without paying a gazillion dollars, bang out some terrible code and experiment with all the wondrous world beyond the '$' prompt. I wanted a DEC, Sun or SGI machine but they were out of reach. I wanted root dammit. I was a broke teenager. In the end I got that little machine to do some amazing things. The effort and learning involved was also responsible for developing skills that earned me lots of paychecks.

Re:Why?

[jbolden]

Wayland developers have no intention of support X but they can bundle it in their window manager. The same way that Apple bundles an X11/quartz-wm with Aqua. Apple only supports the compatibility part, major advances in X11 pass through to Apple because mostly they just compile off the shelf x.org X11. Wayland AFAIK will do the same thing.

As far as applications for Wayland.
a) Those that need to be network transparent will use X11 standards.
b) Those that need performance will use the Wayland standards.
that leaves about 90% of the applications. Obviously in the early years those apps will be X11 however if Wayland is unsuccessful then they will likely become Wayland apps as the libraries move over, i.e. for no particular reason.

If you are saying that Wayland applications won't be usable with an X11 Server running directly against the OS, i.e. X11 not running inside Wayland, yes absolutely that's the intention. Wayland applications will not have X11 compatibility the same way GDI or Aqua applications can't run directly against X11. That's the intention. Wayland aims to replace X11 for desktop usage. There is no question the goal is to retire X11.

We can imagine a world, say in 2030 where Wayland has been successful. Libraries moved over to first start Wayland and later GUI libraries introduced Wayland only features that couldn't be implemented in X11. So applications start casually using features that prevent effective network transparency even those that would benefit from network transparency. In which case I'd assume those applications become client / server the way they are on Windows and OSX today. I certainly remember using network transparency a lot 1988-1995 or so, it worked wonderfully for dumb terminals. As you've mentioned X11 itself and X11 applications don't handle latency well and so using network transparency over a WAN has been less than pleasant. I like it on LANs but still don't use it all that much. What I haven't heard a good argument for is why client / server architectures don't replace the need for network transparency.

Re:Why?

[Brian Feldman]

Sorry, but there's no reason the NOTION of network "transparency" needs to be embedded within the design like X does in order to be done well. Break things down to a serialization protocol, certainly, but the typical usage of X is not for remote display and the layering of protocol and design should reflect that.

Re:Why?

[Anonymous Coward]

I have always been bemused by how the same people who are die-hard defenser on the so-called Unix philosophie can defend X11 against Wayland which is a far more elegant solution only for the sake of network transparency which shouldn't be there in the first place.

Those who pass value judgement about software evolution without contributing deserve to be let rot in their corner and slowly be forgotten as the cavemen they are.

Re:Why?

[Anonymous Coward]

For things to become more convenient and easy ?
I am always astonished by the number of Linux users who are actually opposed to progress ans mostly devoid of intellectual curiosity.

Re:Why?

[Anonymous Coward]

> It's a huge step back.

No, it's what is done everywhere.
Drawing protocol can't deal with the complexity of modern environnement.
And no, using something that look like TWM is not an option. It's freaking ugly and yes visual matters.

Re:Why?

[Anonymous Coward]

A modern OS that can't make use of my $400 video card is a PoS. Linux don't make it on the desktop until it catches up to modern graphics support. Don't need pretty pixel shaders every where, but it should support full HW accel compositing.

P.S. I do enjoy your reference :P

Re:Why?

[Alex Belits]

Network transparency is not a requirement per application, it's a requirement per use case. If ANY application that user has to use remotely is not usable over the network, then ALL of them are unusable over the network, and the user is forced to use "remote desktop" to run all of them in a pseudo-desktop window.

My whole point is that we should not allow development of applications with sabotaged network transparency. The author won't notice if he used a new whiz-bang toolkit and his application is now "local only", but the user will, and I do not want to be faced with applications built with this kind of sabotaged libraries in toolkits in my work, just because someone used a nice-looking library while being unaware how it hurts users.

Re:Why?

[Alex Belits]

As you've mentioned X11 itself and X11 applications don't handle latency well and so using network transparency over a WAN has been less than pleasant.

I said no such thing. Current X11 implementation doesn't work well on high-latency connections -- this is usually not a requirement in the first place (they work great over gigabit LAN), and it can be fixed without breaking X.

Re:Why?

[Anonymous Coward]

There is a comment somewhere in the Phoronix forums (don't remember where) from an X.Org developer who points out that X's network transparency only works for apps that use the native X toolkit - of which there are virtually none anymore. Anything that uses GTK, etc. IS rendered on the server and sent over the network with poor compression. Anyone who thinks that the image is being rendered on the client side is incorrect. X.Org developer says anything like VNC or RDP would be much more efficient because it could use compression. Also, everyone says that adding VNC-ish something would not require internal changes to the framework.

Re:Why?

[Anonymous Coward]

Not sure if you were kidding or not, but I like using Compiz for its functionality... Nothing beats its zoom module, for example.

Re:Why?

[DaVince21]

The actual usefullness of compiz, in my particular working environment of managing IT infrastructure, is close to if not 0.

I think this is the case. There are some mighty useful plugins for Compiz which can be rather handy depending on what you do with your computer. For example, there's a zoom functionality with which you can zoom in the entire screen. I've found this useful on numerous occasions, for example, when operating from a TV screen that's a few metres away. There's also plugins that allow you to draw on the screen, which is useful for presentations and such.

So yeah. Compiz *does* actually add functionality next to the eye candy. It's just stuff that not everybody will need.

Re:Why?

[jbolden]

High latency connections are absolutely a requirement. The whole direction for the last 2 decades in terms of outsourcing the data center and now cloud computing has been to deliver servers via. a WAN. MPLS reduce jitter and latency but still not enough to get LAN like performance. IPv6 will likely reduce latency but it will still be years before IPv6 can be assumed. An application solution that doesn't work well with WAN is unacceptable.

This is one of the reasons that network transparency did not become a killer feature for Linux. Even Linux apps have had to use a Client / Server approach.

Re:Why?

[jbolden]

If ANY application that user has to use remotely is not usable over the network, then ALL of them are unusable over the network, and the user is forced to use "remote desktop" to run all of them in a pseudo-desktop window.

That's not true. You can use network transparency for some applications and a virtual desktop for others, just have more than one connection. Moreover these may not be happening at the same time.

My whole point is that we should not allow development of applications with sabotaged network transparency. The author won't notice if he used a new whiz-bang toolkit and his application is now "local only", but the user will, and I do not want to be faced with applications built with this kind of sabotaged libraries in toolkits in my work, just because someone used a nice-looking library while being unaware how it hurts users.

First off I can't imagine using many if any applications where the choice of client architecture wasn't decided rather deliberately. So I don't agree applications authors will be unaware of this choice to move towards local only, I think it will be a quite deliberate choice to take advantage of features only available on the local libraries.

I understand your point, you consider network transparency vital. But if you want to convince people, you have to explain why network transparency is so lightly used. Why is it that Citrix (remote desktop) seems to be the remote system of choice and not Unix based solutions? Why is it that most applications are designed for client / server either thick client or web? OSX has been in what is pretty close to the post Wayland world now for ten years with xquartz included in the OS. And yet I can't think of a single major application that considers X11 a value add. When Unix apps are ported they generally port in via. X11 and then almost immediately create an Aqua interface. Remote applications that can't be delivered via. Aqua or thick clients use Flash, Java, AJAX. If network transparency were genuinely highly valuable over remote desktop we would see more of it being used in practice.

You and I are on a discussion board with a high number of Linux users and essentially everyone on this board is technology knowledgeable enough to use X11 on windows. The developers have pushed towards additional functionality they haven't chosen to ever offer a thick client with network transparent access. Why not?

On LInux where everyone's native GUI has full support for network transparency the overwhelming amount of application usage is applications running with the same machine being the XServer and XClient, that is not using transparency. Even Linux shops don't make much use of it anymore. The fact is that Unixes have had network transparency since the mid 1980s. End users just don't use it very much. It hasn't turned out to be a killer feature, I thought in the 1990s that it would be when I compared networking on Unixes to technologies like Windows for Workgroups. But X isn't all that much better 18 years later and Windows / OSX networking is. The fact is that X11 is optimized for hardware X servers, and those don't exist anymore. X11 simply hasn't been able to migrate to new technologies and as a result server side rendering has had no advantages that are used in practice in any large scale way.

End users aren't meaningfully harmed by dropping network transparency.

I can think of a lot of reasons X11 failed:
a) The fragmentation between the major unix vendors
b) The fact that high speed client libraries like NeWS / 4sight never got fully integrated.
c) The conservatism of the x86free management people, with respect to feature enhancements and a lost year or two.
d) The hostility towards QT which prevent QT from being directly integrated.
e) And probably most importantly the move towards a more secure internet. The days when rlogin, rcp and rsh were common and xservers ran with low security so you could throw windows where ever you wanted are over.

Re:Why?

[Anonymous Coward]

TL;DR: Xorg is a terrible sprawling mess, Wayland is a modern replacement.

Re:Why?

[hobarrera]

Downside is, we'll have a more fragmented *nix world.
As soon as wayland-native applications start coming out, non-wayland OSs/distros won't be ablte to run them.
The solution would be to have X11 applications run on wayland, but that just means we still have X11 applications going around, and the issue you mentions is not fixed.

Aditionally, if the linux kernel already has redundant code: why not remove it from there? Linux isn't the only user of X11 you know.

Re:Why?

[hobarrera]

Even now, in 2012, I still can't get PA to work with VLC when I play a movie in 5.1 with my stereo speakers. There's lots of peoplewith the same issue on the internet, no solution.
ALSA works fine. It always will, probably.

I can expect the same from X11/Wayland.

Re:Why?

[hobarrera]

It's not just eye candy, compiz does add functionality.
Alt+tabbing with 7 firefox windows open isn't the same if you have previews of each or if you don't (it's way faster than reading the title).
Being able to preview all my desktops while I switching from one to the other help me remember what's on each.

The "drag-to-border-to-dock-as-if-compiz-where-a-tiling-wm" feature is pretty good too.

I don't really use any of the eye-candy of compiz anymore, but I sure as hell do use a lot of functional stuff on it.

Re:Why?

[hobarrera]

Exactly! An even more fragmented *nix world!
Instead of tweaking basic stuff to port between distros/bsds/etc, not It'll be a huge task, just as if you were porting to win32 or mac.

Re:Why?

[Alex Belits]

I understand your point, you consider network transparency vital. But if you want to convince people, you have to explain why network transparency is so lightly used.

it's not ME who has to convince people. I am not the person who is insisting on replacing a working system with some vaporware based on the arguments from The Unix-Hater Handbook. It's people like Wayland developers who should defend their proposals to sabotage the functionality that was present, and in wide use on all Unix-like systems for three decades. All I have to do is to inform ignorant newbies about this.

Re:Why?

[jbolden]

Linux has been the desktop leader on *nix since the 1990s. While Wayland itself is Linux only there is no reason other *nix can't move to via. a porting effort. So I would expect to see something like Wayland-Quartz just as there is XQuartz today. That is if Wayland takes off it will become the dominant desktop system for all the *nix IMHO. There obviously will be something a transition bump, but there will likely be little reason to run the X11 Server over the Wayland Client, so I assume most distributions will move quickly. That is because Wayland is backwards compatible the move on the client side is easy.

It shouldn't be bad.

Re:Why?

[jbolden]

If Wayland remains vaporware it isn't a problem. The issue is when it becomes a real system. And the counter argument to the real system is network transparency. So yeah you do have to defend why network transparency is worth the cost, given it not becoming a successful distributed application paradigm.

No one disagrees X11 has been the dominant system for 3 decades. What they are disagreeing about, and started disagreeing about even by the early 1990s was whether that was a good thing.

Re:Why?

[justforgetme]

paychecks

ftw

Re:Why?

[Alex Belits]

Why should I talk to ignorant and arrogant people who don't understand basic features and use models of the system they are trying to replace?

I would rather INFORM everyone else how harmful the "replacement" is so it will never get off the ground.

Re:Why?

[jbolden]

It is off the ground. The alpha already handles Clutter and EFL are done. X, GTK, QT have partially working versions, buggy as heck but running apps. Ubuntu and Fedora are planning to switch. KDE is working on a direct to Wayland compile. Compiz has already isolated the X based code. The system is still swamped with bugs and thus might fail but it is well past getting off the ground.

And I urge you to consider, the people you consider ignorant or arrogant are the people who care about the display server. If you can't make a valid argument to them you aren't going to be able to make an argument to others. So do what you want. But if argument is going to matter at all, most likely is going to be network transparency is so important that distributions shouldn't move over to Wayland even though it offers advantages. If there is an argument there you probably have less than 3 years to make it.

Again you can do what you want. But the questions I raised in this thread are the ones anyone who cares enough about the issue are going to raise.

Re:Why?

[Alex Belits]

Do you know, how many times a similar project was in exactly the same position, and ended up scrapped?

My words are not directed toward Wayland developers, those people are headed for either massive failure or massive redesign, they are directed to people who consider "porting" good projects (gtk, qt, kde, compiz) to the clusterfuck in the making, thus giving it mindshare.

Re:Why?

[jbolden]

Do you know, how many times a similar project was in exactly the same position, and ended up scrapped?

None, I don't remember any similar projects to replace X with this level of support in the last 2 decades. The closest I can think of is the move to change leadership i.e. from xfree86 to x.org.

they are directed to people who consider "porting" good projects (gtk, qt, kde, compiz) to the clusterfuck in the making, thus giving it mindshare.

Those people aren't considering they doing this, rather they at the 1/2 way mark more or less in terms of implementation of allowing Wayland versions that are equally functional, a porting library. That part is going to be done soon. The big step for them is offering Wayland versions that have features not present in the X version. And to convince them not to do that you are going to need reasons not name calling.

Nokia (QT) focus is on cell phones. Being able to offer developers the ability to bring their cell phone applications directly to Windows (GDI) with similar high performance routines. Your feeling is that network transparency is more valuable though you can't articulate a why.

GTK is run out of Gnome whose focus is moving down market on hardware. So similar to QT.

Compiz you don't stand a chance. Wayland is a pure plus for them, besides they are owned by Canonical that is committed to Wayland.

So if you are going to aim at considering it would be the application developer level who have to decide whether to port to Wayland or worse from your perspective, include feature only available in the Wayland library versions, you are going to need to provide them a reason. KDE is more interesting here. First off KDE doesn't make meaningful use of network transparency, so that's not a huge loss, and again that's something your argument to deal with. However, breaking compatibility with X11 would be a huge step for KDE. The whole "fragmenting the Unix market..." argument that is also being raised is going to have some level of inherent support. But at the same time better compatibility with the native window systems of OSX and Windows is a huge plus and that will have inherent support. I think the later fragmentation is worse for them than the former. Moreover I suspect that is KDE and Gnome move to Wayland all the other Unixes will follow them. So ultimately the only counter argument is going to be that if KDE and Gnome move away from X11 it will become a legacy technology and network transparency lost for a generation.

Which means you need to address the importance of network transparency. Because you are saying to those library and GUI providers they should make major sacrifices in their strategic goals to preserve network transparency. For that argument to carry weight you need to be able to articulate why network transparency is vital. And in particular, since it is obviously not vital now, why you would expect a reversal of the trend away from using network transparency.

Wayland vs X

[sick_em]

Okay so I understand the whole desire to toss out X and it's extreme amount of legacy code, but Wayland to me seems like even at version 1 will be crippled compared to X. The no network transparency I can handle (just barely), but no apps that require full OpenGL? I've tinkered around with OpenGL ES in the past and it does not seem like an acceptable substitute when you need full OpenGL. Why are distros planning on adopting it so quickly? Are these flaws that normal users would not notice or care about?

Re:Wayland vs X

[Hatta]

I can't deal with the lack of network transparency. Sure, you can run X on top of Wayland for now, but once people start writing Wayland only apps that's not going to be a viable way to work anymore.

Here's hoping Wayland flops.

Re:Wayland vs X

[Burdell]

Okay so I understand the whole desire to toss out X and it's extreme amount of legacy code

Why? What part of "legacy code" automatically means "toss [it] out"? The Linux kernel is over 20 years old, and the core BSD code is older than that. Do you also want to just throw them out and start over from scratch because they're old? Now, I agree with you that something that is less functional than its predecessor should not be adopted as its replacement, but I hate the assumption that "old == replace from scratch" that seems to be common in software development (especially in the open source/free software community).

I have tools in my toolbox that belonged to my grandfather, and my father has tools from his grandfather (100+ years old). We still use them because they work. We have other tools that we've bought to perform other tasks, but the old tools are not replaced just because they are old.

Re:Wayland vs X

[AvitarX]

Isn't gtk working on remote run ? I imagine similar will happen for qt.

This has the disadvantage of making those projects harder to do, but the advantage that the widgets know more about what's going on than x does, so it could be done more efficiently.

I've found x's network transparency to consistently be letter useful than vnc

Re:Wayland vs X

[Anonymous Coward]

I can't deal with the lack of network transparency.

Nor can I. Wayland seems to be an attempt to take most of Windows' worst mistakes and push them onto Linux.

Re:Wayland vs X

[Urban Garlic]

This.

Network transparency is huge -- I often need to run graphical installers for commercial software, or run graphical diagnostic software, on systems that live in my machine room, and have fairly basic video cards, and in any case don't routinely have graphical displays attached.

The really-right answer to this, of course, is to separate the display from the the installer and/or diagnostic executable, and connect over a network socket or something, but in practice this doesn't happen. In practice, I connect via SSH with the X session forwarded, and run the graphical app that way.

The loss of network transparency makes remote access much more complicated. It's not lethal, there are things you can do with remote desktop viewers that work, but you end up hacking together a rickety, insecure new solution to what used to be a solved problem -- it really doesn't feel like progress.

Re:Wayland vs X

[dmbasso]

Don't you prefer VNC? Every time I had to run a graphical application remotely I first tried through X but ended up using VNC.

Btw, last week I tried to run firefox from a remote machine (ssh -X) and it seemed to have opened a new instance of the local running firefox (I'm not really sure, I was just trying to do something quickly and didn't have time to verify wtf happened)... if that was the case, was that the expected behavior?

Re:Wayland vs X

[dmbasso]

The loss of network transparency makes remote access much more complicated. It's not lethal, there are things you can do with remote desktop viewers that work, but you end up hacking together a rickety, insecure new solution to what used to be a solved problem -- it really doesn't feel like progress.

I don't think running a VNC server bound to 127.0.0.1 with port forwarded through a ssh tunnel (ssh -L5900:localhost:5900) is much more complicated neither insecure.

Re:Wayland vs X

[Anonymous Coward]

VNC sucks ass. The only benefit it provides is that the latency is lower than X, but there are other alternatives for forwarding X sessions which don't suffer from the same latency problems.

And yes, Firefox sucks donkey ass for opening a new copy locally on my machine when I try to start it on a remote machine. Whoever designed in that behaviour was a fscking retard; if I'm trying to start Firefox on a remote machine it's because I want to run the fscking thing on the remote machine, not locally.

Re:Wayland vs X

[jythie]

People like being associated with the new shiney, since all the core work was done for X11 years ago. It is the same phenomena with politicians writing laws that cover things that are already illegal.. attach your name to a gripe as having come up (or being part of) the solution. Though I suspect part of it is over the years the flavor of the month programming has changed enough that a lot of people look at older code and go 'hey, this does not look like the twue programming we did in school! it must be wrongz!', so the fact X11 was written before the current fads is enough that some people will want to rewrite it for that alone.

Re:Wayland vs X

[Anonymous Coward]

VNC on unix is an X server. No X apps, no VNC

Re:Wayland vs X

[Anonymous Coward]

Throw them out? Yes, some people do. They have for a while. You can take the success of their intentions for either a demonstration of their task being unwanted or a sad commentary of entrenched resources. And no, I do not agree that something that is less functional than its predecessor is a less suitable replacement.

Sometimes simplicity and streamlined is better. Your analogy to tools is misplaced, the old tools work, and hopefully have few obstacles. But imagine if you had some three flanged hammer, but just needed one or two. Taking off the third might be worthwhile.

Re:Wayland vs X

[characterZer0]

X was designed for how video cards worked and what they could do decades ago. Support for newer features is hacked on. The kernel may be 20 years old, but the video drivers have been changing all the time. Wayland is targeted at modern hardware, and aims to be simpler to use and perform better.

Re:Wayland vs X

[Anonymous Coward]

A hammer is not the same as a codebase. Over time the code will grow and things will not stay clean, people leave and those taking care of it will not know how parts work so they code around it. After awhile some projects become a nightmare to work on.

Re:Wayland vs X

[Anonymous Coward]

It's not like there's only one VNC server. There's x11vnc, which connects to a running X server and serves the entire display over VNC. I'm not certain that will work with Wayland's X11 compatibility layer, but it should (giving you access to any Wayland-native apps), and if it doesn't... well then it's not really X11 compatible, is it?

Re:Wayland vs X

[makomk]

As I understand it, you can't run Wayland at all without having a recent GPU that supports it right now because it relies on hardware OpenGL support in order to actually be able to send graphics from the Wayland applications to the compositor. So VNC is probably a no-go too.

Re:Wayland vs X

[Anonymous Coward]

Whoever designed in that behaviour was a fscking retard; if I'm trying to start Firefox on a remote machine it's because I want to run the fscking thing on the remote machine, not locally.

Or, you know, you could just do a "firefox --help", and find out there's a "-no-remote" switch...

Re:Wayland vs X

[Hatta]

Really? That's not any more complicated than 'ssh -X'? It works anywhere and everywhere, without any additional software.

Re:Wayland vs X

[Hatta]

Why can't we do this without losing important features? Features that actually do something useful instead of just look pretty?

Re:Wayland vs X

[Anonymous Coward]

What you don't understand is that Linux might be 20 years old, but what you call "Linux" today is almost incomprehensible compared to the original Linux other than both being mostly UNIX-like. The same cannot be said of X which is still adhering to a basic protocol that is over 20 years old and was designed long before anything that remotely resembles a modern GPU had come into existence. That's why large portions of the X server are *already being bypassed* by modern programs to give you halfway decent performance. Did you know that the large majority of graphically intensive applications (like everything using Cairo) are basically just shoving bitmap images across the X server now instead of trying to use the ancient and convoluted drawing commands that were cool in the 1980's but haven't been useful in a long time? X needs to go away.

    Oh... and don't get me started about network transparency either. X's network transparency SUCKS over anything other than a gigabit LAN *and* you can't disconnect from X and resume the same GUI application on another device like any decent remote GUI system should allow (e.g. why the hell can't I start a remote GUI program on my PC then shift the same program it to my smartphone and keep working on it? The reason is that X's network protocol is *not* the be-all end-all that many people think it is). If X's network transparency was so awesome then NX would never have come into existence.

Re:Wayland vs X

[nine-times]

I think that when people talk about "legacy code", the word "legacy" isn't just supposed to mean "old", but also, "outdated or poorly written". The concept is that old programs may have been written to support hardware or situations that don't exist anymore, or else were anticipated to exist when the software was written but never materialized. Or they have features/functions that never really worked as hoped, or were not properly written or optimized at the time. This "legacy" code isn't necessarily imagined to be the oldest code, but just code that, if you were writing it again from scratch, you wouldn't want to include.

So for some people, yes, there's an "out with the old, in with the new" mentality. For others, there's just the idea, whether it's a reality or a misconception, that there is probably old code in there somewhere that should be removed or replaced.

Re:Wayland vs X

[grumbel]

Sure, you can run X on top of Wayland for now, but once people start writing Wayland only apps that's not going to be a viable way to work anymore.

Hardly anybody is writing X apps and hardly anybody will be writing Wayland apps, instead people will continue to write Gtk+, QT, HTML or SDL applications and network transparency can happen somewhere between those and Wayland.

Also network transparency in X sucks. It's far to slow, cumbersome and inflexible for todays world. It works ok in simple situations where you have dedicated X terminals with a limited set of apps, it doesn't work for situation where you might want to move a videoplayer or game from your desktop to your TV (can't move windows between displays, video streaming is far to slow).

Re:Wayland vs X

[MoellerPlesset2]

Why? What part of "legacy code" automatically means "toss [it] out"?

He didn't say everything that's legacy should "automatically" be thrown out. But X has a huge amount of cruft nobody uses anymore. Nobody actually writes towards Xlib, they use a toolkit. Nobody uses the orignal font functionality and descriptors, the bitmap fonts, the pixel-based rendering primitives, the image system that has no less than three different ways of storing an image (ximage, xpixbuf, xpixmap), that distinguish drawable and non-drawable images, depending on where they're stored. Et cetera. It's not thread-safe either.

Nobody is using the core X functionality, it's all outdated and largely replaced. The one redeeming feature of X - the network transparency, isn't that 'transparent' (again, the API distinguishes server-side and client-side stuff). Nor does it support modern stuff like drag-and-drop, and cut-and-paste has always been inconsistent (highlight-middle-click not being the same as the desktop or application's cut-and-paste buffer). Since nobody's using the core libraries anymore, the network transparency in X mostly consists of it passing events and bitmaps back and forth, something a simpler protocol like VNC can do just as well if not better.

In short, people don't need any of the things that are unique to X, and the things people actually use X for can be done better without it. It's a big load of cruft that exists for backwards-compatibility purposes only. Which is why it's entirely the correct decision to dump X11 and relegate X11 support to a compatbility library, so we don't need to have stuff held back and complicated by these legacy designs.

Re:Wayland vs X

[agwadude]

I don't think running a VNC server bound to 127.0.0.1 with port forwarded through a ssh tunnel (ssh -L5900:localhost:5900) is much more complicated neither insecure.

Is this a joke? Here are some of the missing steps in the VNC "solution":

• Starting the VNC server, with all the right arguments, on the remote end
• Making sure applications on the remote end will display on the VNC server (e.g. setting your DISPLAY variable)
• Starting the VNC client on the local end, with all the right arguments
• Determining what port number to use - if there's another VNC server running already on 5900 (on either end) you would conflict - this would definitely happen in practice if you have ssh sessions to several systems open at once
• Securing the VNC server against unauthorized access if there are other users on either the remote or the local end

Re:Wayland vs X

[marcello_dl]

firefox --no-remote will open a new instance. use -ProfileManager if the existing profile gives you probs. Handy for double login on the same webapp.

Back to topic, the trend is the dumbing down of the PC. Not the interface. the PC as a whole. Wayland is handy to further that objective, so it will be a success even if it kinda sucks. Or so my model of the universe predicts.

Re:Wayland vs X

[MrHanky]

There's no reason why the networking protocol needs to be built into the display server, and no reason why a display server that does not have its own networking protocol has to do without network transparency. From what I've seen, Wayland is flexible enough to allow for networking in multiple ways.

Re:Wayland vs X

[Anonymous Coward]

The network transparency argument is a CANARD. X *is already crappy at network transparency*. Try using X over even a decent broadband connection and prepare for agony. Oh.. try running a GUI application remoted over X on one device like your PC and then ship the same application over to your cellphone or notebook without having to close & restart the application.. YOU CAN'T. Just because Wayland is primarily directed to actually letting Linux take advantage of modern GPU hardware does not mean that X's mediocre at best network transparency needs to hijack all future development. Wayland can implement network transparency *on top of a good display system* instead of trying to hack a display system *on top of sub-par network transparency*.

Before I hear the litany of responses about how wonderful X's remote features are: Why does NX even exist if X is already perfect???

Re:Wayland vs X

[ZorinLynx]

I think a better solution would be for commercial software developers to stop assuming every system you install their software on has a graphical display available.

It's even more insulting when you're installing a command line application, and the installer is graphical. WTF? And many times it's implemented in a very bloated manner, with a JVM that takes 30-40 seconds to start.

CLI based installers are the best. Not to mention they're likely a lot easier to write. The best of both worlds, yet commercial developers don't go there. Arrgh.

Re:Wayland vs X

[Anonymous Coward]

> "Linux" today is almost incomprehensible compared to the original Linux

I do not think that word means what you think it means.

You probably meaned incomparable.

Re:Wayland vs X

[Anonymous Coward]

No I meant incomprehensible... just use your time machine to pull Torvalds 1991 into today and have him stare at the 3.3 kernel code sources ;-)

Re:Wayland vs X

[Anonymous Coward]

Okay so I understand the whole desire to toss out X and it's extreme amount of legacy code

Why? What part of "legacy code" automatically means "toss [it] out"? The Linux kernel is over 20 years old, and the core BSD code is older than that. Do you also want to just throw them out and start over from scratch because they're old? Now, I agree with you that something that is less functional than its predecessor should not be adopted as its replacement, but I hate the assumption that "old == replace from scratch" that seems to be common in software development (especially in the open source/free software community).

Yep, it's what jwz called Cascade of Attention-Deficit Teenagers development model. And it's not the whole F/OSS community -- this disease is specific to the Johnny-come-lately Linux community. It's the flip side of CatB, and it's not pretty -- I go back and forth on whether it's worth it all or not.

Re:Wayland vs X

[ebh]

I work remotely almost all the time, and I routinely access remote systems in all the normal ways (VNC, remote X, RDS, etc.)

> Starting the VNC server, with all the right arguments, on the remote end

$ vncserver -geometry WxH

> Making sure applications on the remote end will display on the VNC server (e.g. setting your DISPLAY variable)

VNC servers are pretty good about starting GNOME (or whatever) sessions; $DISPLAY is set correctly by default.

> Starting the VNC client on the local end, with all the right arguments
> Determining what port number to use

All you need is the right server/port number, which you're given when you start the VNC server.

> Securing the VNC server against unauthorized access if there are other users on either the remote or the local end

You can set a password for your VNC session that's different from your Unix (or whatever) password.

IOW, this is no harder than RDS, and over WANs/VPNs, you tend to get much better performance than pointing $DISPLAY back to your local machine.

Re:Wayland vs X

[CAIMLAS]

Network transparency is huge -- I often need to run graphical installers for commercial software, or run graphical diagnostic software, on systems that live in my machine room, and have fairly basic video cards, and in any case don't routinely have graphical displays attached.

I can run an X server on Windows and do this using Putty just as I can under Linux while running X 'natively'. I don't see how this would be a real problem.

On the other hand, Wayland will make Linux desktop use more capable. It'll be faster, and being able to do something like RDP will be more tenable. Additionally, Wayland is better suited to serve as a general display manager for all ranges of Linux, not just the desktop - this is something which X fails at due to its size and complexity, neither of which is needed or desirable on, say, Linux. The lack of a display manager like Wayland is why GTK did a frame buffer at one point for embedded use and why Android doesn't use X or its toolkits at all (in part), IIRC.

Re:Wayland vs X

[SwedishPenguin]

Right.. Get back to me when I can do this with VNC:
1. Run matlab -nodesktop in screen on remote server
2. Let it run some calculations for a few days or so
3. Connect with ssh -X and set the DISPLAY variable
4. Plot the results

I can with X11 network transparency.. VNC is all or nothing - either I connect to an entire desktop session and waste bandwidth on sending pictures of a bunch of characters or I don't get any graphics at all.

Re:Wayland vs X

[Anonymous Coward]

Because I think that with 1+ GHz multi-core computers being the norm, I ought to be able to flip the "on" switch on my car, boat, thermostat, whatever and have the thing INSTANTLY cold boot and be ready for use. I don't think I should have to wait for the system to spin up 100 services I don't care about, probe ISA bus, the PCI bus, etc. or a myriad of other things a "stock" runs everywhere including on some retarded 1997 Compaq portable distribution has to do.

There's a lot of value to be had by rearchitecting software for what today's hardware offers and throwing out the legacy hooks that make software harder to maintain, debug and optimize.

If your distro offers cold-boot to functioning GUI (i.e. you can drive, change the station, turn on the AC, whatever) in less than 5 seconds, than ignore me. But if it doesn't, then you advocating making millions of people wait for their device just so you can be happy that your software runs on some old piece of shit hardware no one but you and two other nerds care about.

Leave legacy support to old distributions. Let's lighten the load.

Re:Wayland vs X

[Hatta]

If it's provided by the display server, you get network transparency for free. If you expect application developers to implement network transparency, it's going to be spotty at best. If you use 3rd party tools it doesn't integrate well at all.

We don't need several crappy ways to run apps over the network. We need one that "just works".

Re:Wayland vs X

[EyelessFade]

If you'd bothered reading anything from their faq, you'd known that they plan on using OpenGL and was going to, but they can't yet because it draws inn all X11 libs. AKA GLX. And network transparency will come, they even say so in TFV. "It's trivial"

Re:Wayland vs X

[Viol8]

"Try using X over even a decent broadband connection and prepare for agony"

I do it often , never had a problem. Hell , I even used to run xterms over a modem link FFS!

"he same application over to your cellphone or notebook without having to close & restart the application"

Not a problem if the application is written to support it. There's no limit to the number of X servers an app can link to at any given time.

"Why does NX even exist if X is already perfect???"

No one is claiming X is perfect - but why would I would to exchange something thats imperfect for something that apart from eye candy potential had very reduced functionality and is clearly broken from a network point of view.

Re:Wayland vs X

[ppc_digger]

x11vnc is not limited to X. It can also remote control a console session on a Linux framebuffer device, and a few other things. Also, from TFA:

- Jesse has been playing around a bit with some remote Wayland support using libvncserver. He's apparently had some success with this and expects to push some code upstream soon.

So there are already people working on it.

Re:Wayland vs X

[Anonymous Coward]

ssh -X is doing a bit of work behind the scenes -- to get the same ease of invocation, it's not horribly hard to make an equivalent script to do the grunt work of patching together a ssh tunnel, a VNC server, and a VNC client, though of course it's not "just there" wherever openssh is.

Anyway, it's pretty straightforward to do the functional equivalent with VNC, and there's no need for the all-or-nothing dichotomy you say. What you must do is exactly what you do with ssh/x11 -- separate the tty session from the X session. This can either be:

1. Run matlab -nodesktop in screen on remote server
2. Let it run some calculations for a few days or so
3. Connect with script and set the DISPLAY variable (Where the script sshes into the remote host with port forwarding to launch a VNC X server, and launches a VNC client on the local host.)
4. Plot the results

or, my preference:

1. Leave a persistent X server running on remote server (Xvfb or Xdummy)
2. Run matlab -nodesktop in screen on remote server, and set the DISPLAY variable
3. Let it run some calculations for a few days or so
4. Plot the result (if you're scripting matlab, you can put this in the script -- or call it interactively, no difference)
5. Connect with different script, which uses x11vnc to connect to the existing X session, same port forwarding and local client. (You can do this before you plot it, too, but I put this order to emphasize you can.)

I prefer this method, both because I can toss the plot(s) in the script, including plotting intermediate results, and because the persistent X server gives my graphical session most of the benefits screen gives my tty session, (i.e. disconnect immunity, multiple viewers, but not multiplexing), with a dead-simple script* -- and these are benefits you can't get with ssh -X!

* If you run a VNC X server, your script has to check if the VNC server is already up so as not to start a new one, which gets messy when you might have other VNC sessions for other purposes, but you can have x11vnc die with the connection, and start a new one every time.

Re:Wayland vs X

[Alex Belits]

GTK and Qt don't and never will have any network transparency on their own -- they use X for this purpose, and it works JUST FINE. If any project will replace X, it will also have to replicate network transparency provided by it, and no, "let's copy bitmaps of a virtual screen" won't cut it, applications must obey window management on the user's box, and should not produce more network traffic than notoriously "chatty" X.

Re:Wayland vs X

[MrHanky]

X11 doesn't give you network transparency for free, it does it at the cost of being a crummy and overly complex set of ancient technologies that no one uses any more. And the network transparency itself is based on how a window system was expected to work 20 years ago, meaning it's extremely inefficient with modern applications (GTK, Qt, etc.). Ever tried running Firefox across a LAN with X11? It's not a pleasant experience. Perhaps NX improves it. Sticking with X11 simply because it has one outstanding feature would be fine only if that one outstanding feature wasn't implemented so poorly.

Re:Wayland vs X

[Lussarn]

Those programs work the way they do becuase X11 is what it is. If there where no network transparancy the installer would probably be comandline. Likewise diagnosicsprograms would probably be better by using a simple webserver and serve to your browser. No change comes with zero flaws, but in the end X11 network transparency is used only by a few (I have definitely used it also for these examples so I know whar you are saying). But I can live without it, and X isn't really going away in 5 minutes anyway, there will be time to adapt. Most of the time even VNC would be adequate today. Also remember that Wayland is made to make desktop Linux/Unix better, the examples I hear about why X11 and it's network transparency are great are always about the server.

Re:Wayland vs X

[Anonymous Coward]

quote: Not a problem if the application is written to support it. There's no limit to the number of X servers an app can link to at any given time.

That is complete and utter BS. Name one single application that I can start on my PC and then simply run a command and have the same application *including all the program state of the running application* simply pop up on another device (doesn't even have to be my smartphone). I have tried to do that and with the exception of a 10 year old research project that never went anywhere there is NO SUPPORT for this functionality and I've been insulted several times by the so-called X "experts" for even suggesting that such functionality could even be useful. Meanwhile many of the same "experts" blather on about how screen is such a great utility but daring to suggest that similar functionality in a GUI application would be useful is a mark of stupidity.

Re:Wayland vs X

[Anonymous Coward]

You can do OpenGL to a bitmap... at some point OpenGL goes to raster... That raster does not have to actually be shipped to a hardware display (even if 90% of the OpenGL stuff happens in the GPU... you've never heard of something like CUDA or GPGPU?? People actually use GPUs for things having nothing to do with display), it can be taken from the raster memory and shipped over the network with whatever technology works. With the right combination of compression and latency reduction techniques (just don't do RTs stupid) you can get pretty satisfactory results. After this, you can get a little fancier. Yes, the naive method may suck for remoting certain things (like FMV), but solutions exist for that too.

Getting better performance than hinted raster transfer for 3D primitives is extremely difficult unless you can make some strong assumptions about remote client hardware. Companies like Onlive demonstrate that ofttimes it's not even worth the attempt.

Basically, you need to get a clue.

Re:Wayland vs X

[Anonymous Coward]

But X has a huge amount of cruft nobody uses anymore.

So end-of-life it. Not hard. Far superior to a big-bang rewrite with less functionality and major incompatibilities.

If they're not willing to do an incremental upgrade/rewrite of X, something X was designed for, then they shouldn't be doing it at all.

Re:Wayland vs X

[simcop2387]

Why hope it flops? There's currently projects to do network transparent wayland. In fact it should likely end up working much better than doing X over the network since it will be able to do things that and actually compress the data properly to get better transfer times and won't need so many round trips just to draw something (last i looked X needs about 4 round trips for putting something on the screen, minimum, all synchronous). Along with that you've got the rather annoying problems where X11 doesn't provide any good ways to render things in any decent way that actually works among all the drivers: Xrender [held up by nvidia still for their fencing code], XAA [being dropped since most other drivers have moved on], EXA [i think this is the latest one], UXA [done by intel that's very similar to EXA but uses GEM instead of TTM for memory on the graphics card], and then a number of others that are in embedded graphics DDXs that are only supported by the manufacturer because they don't want to give out code.

The situation with X11 is not as full of rainbows as people seem to think. Because of all the stuff above, most toolkits and libraries have stopped using X11 for doing anything other than pushing bitmaps, they're doing all font rendering in application, all widget drawing in application, all scrolling in application, the only thing the X server gets used for by the applications anymore is to display images and get input. That's what wayland is being written for, to replace the rather horrific design issues of the X11 server for handling the hardware.

Wayland aims to fix all this by instead using the kernel interfaces KMS and indirectly DRI2 (through Mesa). This leaves one interface to talk to hardware to maintain, far less code for bugs to be found in, and makes it possible to support things that X11 can't do right now like supporting hot plugging graphics cards, NVIDIA Optimus (the "support" in linux right now isn't really support. It's running two X servers and copying one of them to the other to make it look like there's only one).

Re:Wayland vs X

[AvitarX]

If X works "just fine" why does it run worse than loading a window through VNC?

As for the toolkits, I don't see why they couldn't implement network transparency (or close to it), sending widgets rather than drawing commands. I don't really know these things, but in my head that would be the first step to getting something as efficient as citrix working (though obviously only for a subset of apps).

I've seen a GTK backend that went to a terminal, why couldn't one be made that went over the network?

Re:Wayland vs X

[snadrus]

GTK offers HTML5 as a backend. Network transparency at the toolkit level is what X11 intended (with it as the toolkit). You can't get much less chatty than an HTML5-forwarded app in terms of existing standards.

Re:Wayland vs X

[Alex Belits]

If X works "just fine" why does it run worse than loading a window through VNC?

Because you have very high latency, and X protocol is not optimized for that, what is fixable. On the other hand, VNC does not let the user to combine remote and local applications on one desktop with full window manager, and this is pretty much unfixable without a serious redesign.

Re:Wayland vs X

[Hatta]

X11 doesn't give you network transparency for free

Can you actually demonstrate that network transparency is what slows X down? Every analysis I've seen points to heavy toolkits, which are going to be present on Wayland as well. As I understand it, X uses shared memory to communicate with clients on a local system, which is going to be as fast as anything else.

I've never seen any evidence supporting this notion that network transparency is inherently costly.

Re:Wayland vs X

[Alex Belits]

Yes, and mplayer offers two solutions to display video on VT100.

Re:Wayland vs X

[allo]

inconceivable!

Re:Wayland vs X

[Anonymous Coward]

Some of us use the two cut and paste buffers differently and don't wish them to be 'consistent'.

Re:Wayland vs X

[Anonymous Coward]

http://en.wikipedia.org/wiki/Legacy_code

Re:Wayland vs X

[MrHanky]

What I was saying was that X11 does not give you network transparency for free, not that network transparency in X11 is inherently costly in itself. X11 delivers network transparency at the cost of being a massive patchwork of ancient crud. As for shared memory magically being 'as fast as anything else', that's simply untrue.

Re:Wayland vs X

[spitzak]

Nor does it support modern stuff like drag-and-drop

Actually DnD support was added and works quite well. Much better than the cut & paste because they scrapped a lot of stupid complexity that nobody uses.

and cut-and-paste has always been inconsistent (highlight-middle-click not being the same as the desktop or application's cut-and-paste buffer)

Actually this is a requirement so that selecting text should not interfere with the most recent cut or copy command. But what X gets wrong is failing to unify this selection with DnD. In fact a design goal should be that middle-click does the same action as dragging the selection and dropping it at the point the middle-click was done. The original X clipboard design was actually equivalent to DnD, not cut & paste, but too many intermediate users and developers failed to realize this and made the current mess which is only now being cleaned up.

I believe Wayland is a chance to fix this. Last I looked they have only worked on DnD and there is no clipboard support yet. It would be good to point out that selection + middle click should be unified with the DnD, not the clipboard.

Re:Wayland vs X

[takeda64]

I would say that it depends on complexity, there are some programs that are simple and short you can compare them to tools. Majority of programs though are complicated and they grow. You could compare them to a house. When you built it it is fresh, everything works fine, and it has latest goodies. But with time, regulations and requirements change. Maybe you need to tear down a wall, make another room, build another floor. With time no matter how well you took care of it it shows sign of its age. You might not even be able to make some changes because regulations forbid you. For similar reason some changes might require fundamental changes and will be very expensive, so it will be more feasible to just tore it down and build a new house. I'm not sure whether you're involved in software development but if you will look at source code of some programs that were there for some time you can see how complicated they become. X11 is one of the well known for its complexity.

Re:Wayland vs X

[takeda64]

Because it is starting from scratch, it doesn't mean it won't incorporate other features later on.

Re:Wayland vs X

[Anonymous Coward]

Some of us jerk off to furry midget gay ASCII porn, too. That's not a good argument for not designing for the most common scenario.

Re:Wayland vs X

[Hatta]

Great, then there's no reason not to make Wayland network transparent. If they'd do that, a lot of cynical bastards would shut the hell up.

Re:Wayland vs X

[dkf]

The really-right answer to this, of course, is to separate the display from the the installer and/or diagnostic executable, and connect over a network socket or something, but in practice this doesn't happen.

I see where this is going. You'll have a graphical installer just for the server-side component of the network-independent installer. (I've seen things at least as bone-headed.)

I've no problem with Weyland, so long as it allows X11 clients to connect. There's buttloads of them out there, and it will take a hugely long time to persuade them to switch. Some simply won't until the other server platforms do; there will need to be deployed Weyland "server" implementations for far more than Linux for X11 to cease to be relevant. In particular, there's X11 servers for both Windows and OSX, and the way they work might point the way for X11 and Weyland to coexist (core and "native" apps using Weyland; X11 for others). Sure, the X11 apps won't have access to all the snazziness of Weyland but the apps will still be largely usable, and that's what really matters.

Re:Wayland vs X

[dkf]

Nobody actually writes towards Xlib, they use a toolkit.

Toolkit authors write towards Xlib (or sometimes even raw X protocol). Or were you assuming that toolkits just happened by magic?

Re:Wayland vs X

[Burdell]

I suggest you go read Linus' original announcement for the Linux kernel. I'm betting you aren't running a single-processor i386 system, and you probably don't have any MFM hard drives. Support evolves over time without replacing the whole thing. The problem with "targeted at modern hardware" is that there's a lot of what you might not consider "modern hardware" still in use.

Re:Wayland vs X

[Burdell]

"Nobody actually writes towards Xlib" - so GNOME, KDE, etc. are "nobody"? X clients have used toolkits since almost the beginning; Xt, Xaw, Motif, etc. existed a long time ago.

If nobody uses the original font functionality, then deprecate it and move on. When something is obsolete in the Linux or BSD kernels, the code is retired; they don't throw the whole kernel out and start over.

VNC is completely different from X and inferior in almost every way. For one thing, rather than just run an application on a system and export its display, you have to run a virtual X server (with all of its overhead such as a window manager) and then screen-scrape the virtual screen and export it. It is impossible for VNC to have any of the drag-and-drop or cut-and-paste functionality you criticize in X.

Re:Wayland vs X

[jbolden]

It will get OpenGL rather quickly, likely long before the mainstream distributions switch. Remember the GUI libraries have to switch and then the GUIs before the distributions would even consider it.

Re:Wayland vs X

[jbolden]

There won't be any impact on applications designed for network transparency they will just stay with X11. Wayland supports X11 apps, X11 won't support Wayland apps.

Re:Wayland vs X

[jbolden]

Matlab is going to have an X11 version for a very long time. That will still be possible. The only effect is on applications that drop X11 support.

Re:Wayland vs X

[jbolden]

but why would I would to exchange something thats imperfect for something that apart from eye candy potential had very reduced functionality and is clearly broken from a network point of view.

Why would I want to trade in my boat to ride on roads when it works well on rivers? Wayland is not designed for remote applications for those applications there will be a downgrade. If this is a big deal to the applications you use they wont' switch away from X11 and the whole thing won't matter.

Re:Wayland vs X

[jbolden]

It isn't costly in the way the GP said it was. Where it is costly is things have to be designed to be network transparency safe and have extra layers of abstraction. The kinds of behaviors where OS libraries pass information two ways from the GPU and CPU (like in Quartz or GDI) aren't possible. The demo that Steve Jobs did of 32 video streams being hit with graphic effects would be beyond impossible on X11.

Re:Wayland vs X

[Daniel Phillips]

Wayland developers have done a poor job of communicating their intentions regarding network transparency. They need to communicate a clear and credible plan that does not attempt to push the problem off to somebody else's project. The plan must not rely on application or per-desktop transparency, but transparency at a common layer that all graphical applications use. And the plan must be executed, mature and reliable before Ubuntu makes the mistake yet again of pushing out something new and shiny before it is ready.

Until I see such a plan it would be foolish for me personally to have a lot of confidence in Wayland. That said, it is about time to take a closer look at the innards and from an opinion that is not second hand about the relative design quality of Wayland vs X. In truth, producing a better design than X is in no way a high bar. But X is mature, it is working now in many different context, and very well. That is a high bar indeed. If there is a credible plan for Wayland to meet that standard I have not seen it. Maybe that is just a failure of communication.

Re:Wayland vs X

[Daniel Phillips]

X11 doesn't give you network transparency for free, it does it at the cost of being a crummy and overly complex set of ancient technologies that no one uses any more....

Run glxinfo. See where it says "direct rendering: Yes"? That means you are talking nonsense.

Keep in mind that Keith Packard, perhaps the lead dev for Wayland, was also instrumental in making OpenGL direct rendering a reality on Linux. So whatever mistrust I may have about Wayland's design directions, they are tempered by the fact that Keith Packard is involved. Mind you, they need to be tempered a lot more before I stop perceiving this project as mainly a way to interfere with my future productivity.

Re:Wayland vs X

[Anonymous Coward]

Really? That's not any more complicated than 'ssh -X'? It works anywhere and everywhere, without any additional software.

Heh. Kids these days.

(By which I mean this: At one time it was also complicated to do the equivalent of 'ssh -X'. More complicated, if I remember correctly. The things you had to do in order to tunnel X11 got compressed into the '-X' switch because lots of people wanted to do them without having to remember all the magic incantations. If lots of people start demanding easy VNC over SSH, that'll get simplified into a single ssh switch too.)

Re:Wayland vs X

[Conley Index]

I don't think running a VNC server bound to 127.0.0.1 with port forwarded through a ssh tunnel (ssh -L5900:localhost:5900) is much more complicated neither insecure.

Is this a joke? Here are some of the missing steps in the VNC "solution":

• Starting the VNC server, with all the right arguments, on the remote end
• Making sure applications on the remote end will display on the VNC server (e.g. setting your DISPLAY variable)
• Starting the VNC client on the local end, with all the right arguments
• Determining what port number to use - if there's another VNC server running already on 5900 (on either end) you would conflict - this would definitely happen in practice if you have ssh sessions to several systems open at once
• Securing the VNC server against unauthorized access if there are other users on either the remote or the local end

ssvnc does all this good enough for many applications. (It has some bugs, especially the Windows version, but nothing you cannot work around, if you are familiar with the *nix world.)

Re:Wayland vs X

[Richard_J_N]

Actually, it can be quite useful having 2 different clipboards, if used thoughtfully. Otherwise, klipper or parcellite can unify them for you if you want.

Re:Wayland vs X

[Timmmm]

Well indeed. VNC is a hassle. But so is remote X! Have you ever tried to connect to an existing X session remotely (VNC-style)? I'm fairly sure it is not possible.

That is one of the many reasons why a replacement for X11 is needed.

Re:Wayland vs X

[Anonymous Coward]

Some of us LIKE the separate pastebuffers. I hate that Firefox, for example, merges them.

Re:Wayland vs X

[Kymermosst]

Your idea of pretty is not the same as my idea of pretty.

I might add that I hate all of the stupid animations that GUI creators like to throw in. For instance, when I click a button that is supposed to make the windo go away, I want it to just fscking disappear. I don't want it to shrink down and slide away, I want it to instantly vanish. And so help me if it has a stupid sound effect...

Re:Wayland vs X

[dbIII]

A lot of that "cruft" is still used by some applications, including expensive commercial ones. Sadly that's why the open source implementations of X on MS Windows fail to deliver because they are cutting out or failing to implement the "cruft" without checking to see if it is used or not.
Taking the same approach on a different platform is going to break things there as well.

Re:Wayland vs X

[Viol8]

"That is complete and utter BS."

man XOpenDisplay

Fscking idiot.

Re:Wayland vs X

[Brian Feldman]

You're not a developer. Obviously, the statement went over your head. Writing to Xlib means writing to Xlib. It does not meaning writing to GTK+ or Qt that uses Xlib internally when on the Unix platform.

Re:Wayland vs X

[MrHanky]

Lololwhut? You think direct rendering of OpenGL somehow reduces the complexity of X.org? Idiot.

Re:Wayland vs X

[Anonymous Coward]

Except Linux works well whereas resizing a window without a barely supportable amount of tearing on X11 is impossible.

Re:Wayland vs X

[Anonymous Coward]

I hope Linux flops. Any OS that has a user base that actively discourages gamers via hamstringing modern graphics while b*tching about why no one uses it for a desktop, deserves to disappear into nothingness.

/s

Re:Wayland vs X

[badkarmadayaccount]

VLC has an ascii-art backend as well.

Re:Wayland vs X

[badkarmadayaccount]

RDP

Re:Wayland vs X

[Barabul]

I do not think that word means what you think it means.

Re:Wayland vs X

[renoX]

> Wayland is targeted at modern hardware, and aims to be simpler to use and perform better.

Sorry, but that's false: Wayland aim to be simpler to *develop for*, being a smaller codebase, that's all.
As both are low level layers below GUI, there should be very little difference in usage, except that Wayland isn't currently network transparent.

And in performance Wayland should be very similar to X given that X has a DRI2 extension which is very similar to Wayland..

Source links?

[Anonymous Coward]

Any non-Phoronix links available for this information? Say, the Linux Foundation site maybe?

Wayland is a huge step backwards.

[Anonymous Coward]

To start with: the best wayland-related comment of all time appeared right here on slashdot:
http://tech.slashdot.org/comments.pl?sid=2699657&cid=39198273

Second, please note that all comments in this post are regarding unix or linux as a desktop operating system as opposed to HPC or server-based usage.

Now to move on: Currently the only wayland compositor implementation is weston. Weston requires kms, which makes it pragmatically linux-only, and it also requires udev, which makes it *actually* linux-only. To compound the problem, the developers are talking about integrating it with systemd. When asked how that will affect porting efforts, developer response was along the lines of "just port systemd to bsd."

Wayland is meant to replace x.org entirely, but it simply replaces the overcomplicated xorg ecosystem with an overcomplicated mess of build-time dependencies while removing all of the features that have kept x11 on top. To make the whole situation an absolute joke, x11 integration is regarded as the most important part of wayland's code.

Fedora and Ubuntu want to switch to Wayland entirely. This is another in a series of awful interface decisions that have lead to things like Linux Mint and Scientific Linux creeping up and taking userbase away from what have been powerhouses. What I can't figure out is why no overarching community has arisen among the 'conservative' linux users. There's clearly a lot of backlash to the ridiculous things going on at freedesktop.org -- xcb, wayland, systemd, journald, et al -- but there seems to be no alternative standards committee coalescing... and that will be the death of linux as an alternative for power users.

Re:Wayland is a huge step backwards.

[Hatta]

What I can't figure out is why no overarching community has arisen among the 'conservative' linux users

I see what you did there.

Re:Wayland is a huge step backwards.

[Anonymous Coward]

So what you're saying is that the "innovative" open-source community should throw up firewalls to resist change at all costs... especially when it comes to changing what is undeniably one of the weakest elements in the Linux desktop experience? All in the name of sub-par "network transparency" that is worse than a sick joke over WAN connections and that has fallen behind other network transparency solutions? Wow... I want to live in your brave new world of waiting 10 seconds for responses from remoted X-terms and steadfastly refusing to even consider using modern graphics hardware in an efficient manner! You sir are a visionary!

Re:Wayland is a huge step backwards.

[Eivind Eklund]

To start with: the best wayland-related comment of all time appeared right here on slashdot:
http://tech.slashdot.org/comments.pl?sid=2699657&cid=39198273

Second, please note that all comments in this post are regarding unix or linux as a desktop operating system as opposed to HPC or server-based usage.

Now to move on: Currently the only wayland compositor implementation is weston. Weston requires kms, which makes it pragmatically linux-only, and it also requires udev, which makes it *actually* linux-only. To compound the problem, the developers are talking about integrating it with systemd. When asked how that will affect porting efforts, developer response was along the lines of "just port systemd to bsd."

This is a prime example of Linux developers doing "Embrace, extend, extinguish" on Unix. "We're dominant - let's lock everything down to our solution, and force anybody else to play catchup."

It may not be particularly intentional to damage other systems - having Linux blinders on, it's easy to see "All the world's a Linux machine", just like it used to be "All the world's a VAX" and "All the world's a Windows machine" - but it does the same kind of damage as if it was intentional.

Eivind.

Re:Wayland is a huge step backwards.

[Chemisor]

So what? If you are using BSD, you are perfectly welcome to continue using X.org on it. Wayland in no way "damages" it, it merely aims to replace it on Linux. Are you seriously going to cry "embrace, extend, extinguish", just because Linux developers are not interested in porting their work to your BSD?

Re:Wayland is a huge step backwards.

[Anonymous Coward]

What I can't figure out is why no overarching community has arisen among the 'conservative' linux users.

It's called "Debian".

Re:Wayland is a huge step backwards.

[snadrus]

Doesn't the "Extinguish" involve breaking support for while not providing source code? That doesn't apply
What about "Extend", which protocol is getting extended here, exactly?

Re:Wayland is a huge step backwards.

[goruka]

And who cares about that?, we are stillt talking about Open Source. if this helps make development faster by reusing more code It's certainly welcome. If *BSD users want this functionality, they can write a compatibility layer, the same way they did with ALSA, then they have the source code for Wayland available.
Otherwise, they are giving Wayand devs the burden of mantaining all the ports and even their own video drivers (like in X11)

Re:Wayland is a huge step backwards.

[Eivind Eklund]

I'm saying "Embrace, extend, extinguish" because there are developers that very specifically tries to tie things to Linux. Wayland specifically damages it: It creates a new path which is likely to be used by significant amounts of applications, and which means those applications won't work except with Wayland.

And no, it will not be feasible to keep using old versions of things if the world moves along. It is e.g. not feasible to use an old web browser, both due to the web itself evolving and the now known security issues. You're effectively forced to upgrade a lot of things; staying with the old does not mean keeping the old features, it means losing features.

Re:Wayland is a huge step backwards.

[Eivind Eklund]

Doesn't the "Extinguish" involve breaking support for while not providing source code? That doesn't apply

Source code has nothing to do with it. It involves breaking support for your competitors.

What about "Extend", which protocol is getting extended here, exactly?

Unix API space.

Re:Wayland is a huge step backwards.

[Eivind Eklund]

And who cares about that?, we are stillt talking about Open Source. if this helps make development faster by reusing more code It's certainly welcome. If *BSD users want this functionality, they can write a compatibility layer, the same way they did with ALSA, then they have the source code for Wayland available.

Otherwise, they are giving Wayand devs the burden of mantaining all the ports and even their own video drivers (like in X11)

"Do not tightly depend on Linux-specific features" and "maintain your own ports" are different. There's significant differences between writing reasonably portable code (which I think it is fine to ask for) and to fully deal with creating all kinds of infrastructure (which I don't think it is reasonable to ask for.) systemd is a seemingly OK solution to a problem that's been solved a number of different ways, with intentional direct Linux dependency, created by somebody (Lennart Poettering) that works for a Linux company (Red Hat) and would have advantages to blocking off other ways of dealing with things. This has also been stopped systemd from being used by some other Linux distributions, but the author is pushing for direct dependency on systemd. I interpret that as partially ego-driven (he'd like his software used), and partially personally philosophically based (he's a consistent GPL-user) with even more than normal Linux blinders on.

Re:Wayland is a huge step backwards.

[Brian Feldman]

If necessary, I'm sure we can reimplement systemd portably and, if it is a general improvement in other ways, that would take over that niche in the Wayland ecosystem organically.

Re:Wayland is a huge step backwards.

[Anonymous Coward]

> why no overarching community has arisen among the 'conservative' linux users

Because they are mostly morons stuck twenty years ago who can only complain and are unable to produce anything remotely useful ?
Seriously, "conservative" users should just stop bitching, group together on the site of their choice and be forgotten by the rest of the internet.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Comment removed

[account_deleted]

Comment removed based on user account deletion

Giant Step Backwards

[agwadude]

One of the features that always distinguished X from other display systems like Mac and Windows has been network transparency. You can ssh to another Linux system, start an X application, and that X application will appear on the system you ssh'd from. This is immensely useful and evidence of a well-thought-out design, but it's an afterthought to Wayland. They say they might be able to render to a VNC server, but VNC works like crap and is full-desktop forwarding rather than individual window forwarding.

It's extremely ironic that when X was created in the 80s they recognized the importance of distributed systems and network transparency, but now it's 2012, the Internet and the cloud is king, yet network transparency isn't a core feature.

All this because you can't cross-fade when switching VTs in X or have a "rotating cube" animation (see "Is wayland replacing the X server").

Re:Giant Step Backwards

[Anonymous Coward]

You can run seamless apps on RDP and ICA, and have been able to do so for at least 10 years now.

Re:Giant Step Backwards

[CAIMLAS]

I don't understand how or why you think Wayland will prevent you from running X apps remotely. You can do this in Windows using Xming without a problem. What it might allow you to do is use the full desktop remotely without full bitmap transfer (ala RDP).

Re:Giant Step Backwards

[Viol8]

"That feature is not a requirement for the GUI in my car or boat."

Who gives a shit about your car or boat? We're talking about desktops and servers , not some mickey mouse Busybox system in your toyota.

"Wayland gives them a way to composite multiple OpenGL apps without the overhead and legacy crap of X."

If all you want is something that draws straight to the framebuffer then there's nothing stopping you doing it yourself.

Re:Giant Step Backwards

[Chemisor]

While IT professionals at work run X over a network all the time because they have a lot of computers to manage, it is something 99.999% of computer users never do. Home users don't even know what ssh is and even if they did, they'd have no reason to use it. If you are a sysadmin gloating over your domain displayed in a 100-window screen session, Wayland is obviously not for you. It's for the rest of us, who are perfectly willing to give up network transparency to get a lighter, faster windowing system.

Re:Giant Step Backwards

[Alex Belits]

That feature is not a requirement for the GUI in my car or boat.

Oh yes, it is. Automotive companies are working on switching from CAN to Ethernet (with realtime extensions), so there is no good reason to prevent people from implementing remote user interface.

Re:Giant Step Backwards

[Alex Belits]

Windows using Xming without a problem

lol

Re:Giant Step Backwards

[Anonymous Coward]

Home users aren't using a Linux distro in the first fucking place. Cutting out all the things that it does genuinely better than Windows and MacOS is not a path to success.

Re:Giant Step Backwards

[manicpop]

Why can't we have a lighter, faster windowing system that supports network transparency? Network transparency might seem like something only "power users" need, but it can be useful in many real world situations. My girlfriend couldn't tell you what X11 is, but she does know she can check her e-mail on her laptop even though Thunderbird is installed on our desktop computer. She just has to click on an icon that runs a script I wrote that sshes to the desktop computer and runs Thunderbird. Sure, "most people" don't do this, but "most people" would find it useful if they knew how to do it.

Re:Giant Step Backwards

[sl3xd]

If you are a sysadmin gloating over your domain displayed in a 100-window screen session, Wayland is obviously not for you. It's for the rest of us, who are perfectly willing to give up network transparency to get a lighter, faster windowing system.

Hate to tell you this, but you're wrong on a couple of points:

1.) A lot of people absolutely *require* network transparency. By dumping network transparency, you're alienating the admins and developers who've brought Linux to the point it is today, and continue to support it. Network transparency costs nothing in terms of runtime speed or resources. If you're not using network transparency, no extra cycles are consumed. This has been true with X for ages. Alienating the admins who have been promoting Linux on the desktop at organizations worldwide isn't a good thing - you'll force them and their organizations to use something else (likely *BSD).

2.) "X is bloated and slow" is honestly the biggest load of crap I've seen in my years in computing. Have you ever looked to see the kind of resources required by "modern" technologies? Quartz uses a few times more resources than X11, and the Windows GUI is even worse. I really don't see how anyone can claim that X11 is hard on system resources. It may have been true twenty years ago, but twenty years ago, a megabyte was a huge amount of memory for a computer to have.

X uses considerably fewer resources than my web browser. X is using less than 0.5% of my three-year old system's memory. What, pray tell, is that extra fraction of a percent of memory going to buy *anything* in terms of 'lightness' or 'speed'? Nothing at all.

I'm willing to bet Wayland+Weston will use more RAM and more CPU cycles than X11 does. Given the "screw you community" view the project has to anything other than Linux, I really don't see how Wayland will succeed. Few projects do well when they alienate entire communities of developers.

Re:Giant Step Backwards

[Chemisor]

A lot of people absolutely *require* network transparency. By dumping network transparency, you're alienating the admins

And I'm sick and tired of admins thinking they are the world. Yes, sysadmins have very specific requirements for their work environment, but if you are an admin, you're a minority in a minority. There are so few of you, it boggles the mind that you think everybody must design every application to accomodate you. Here we have Wayland, which is not designed for you. It gives the vast majority of us, who care about a lightweght GUI, what we want. We don't care that it currently does not support network transparency because outside your sysadmin world pretty much nobody uses it. Why not? Because most of us only have one computer. But what do you care about the majority of computer users? This fancy new UI does not cater to your needs so it must be crap, no matter what anybody else thinks. Grow up, admins. You can still use X.org. Let the rest of us have something we like once in a while.

Alienating the admins who have been promoting Linux on the desktop at organizations worldwide isn't a good thing - you'll force them and their organizations to use something else (likely *BSD).

Yeah, sure you will. Wayland is not Linux, and neither is X.org. Just because we want to use Wayland, doesn't mean you can't keep using X.org if that's what you like. Gees...

"X is bloated and slow" is honestly the biggest load of crap I've seen in my years in computing. Have you ever looked to see the kind of resources required by "modern" technologies? Quartz uses a few times more resources than X11, and the Windows GUI is even worse.

Being fat is unhealthy, no matter how many friends you have who are fatter than you. No, X.org is not the fattest UI out there, but that doesn't mean there is no room for improvement.

X uses considerably fewer resources than my web browser. X is using less than 0.5% of my three-year old system's memory.

X uses enough resources to be incapable of starting instantly. Most people still don't run off an SSD, so loading times are quite noticeable. And resources are not the only thing wrong with X; lack of frame boundaries in the protocol has to be the single reason why X looks so bad and requires each and every app to jump through a boatload of hoops just to refresh the screen without artifacts.

Given the "screw you community" view the project has to anything other than Linux, I really don't see how Wayland will succeed. Few projects do well when they alienate entire communities of developers.

If Linux is rare on the desktop, the rest of the "community", by which you must mean BSD, is truly miniscule. Vanishingly miniscule. As in, you'll be lucky to find a dozen guys using BSD on the desktop. Oh, yeah, you must be a sysadmin, comprising 99% of the BSD using community, and, as I mentioned above, utterly disparaging of anything not written specifically for you.

Grow up. Most people are not sysadmins. People who are not sysadmins and in a unix environment are most likely using Linux. (Yes, there are Mac fans, but they have MacOS and are not interested in using anything else.) These people don't need network transparency. These people are much more interested in having a smoothly drawn desktop, and Wayland can give them that, while using less resources. So stop whining and keep using X.org. Really. Move along. Nothing for you to see here in Wayland.

Re:Giant Step Backwards

[agwadude]

This is getting out hand.

First, X network transparency is not used just by sysadmins. At my university it's used every day by students who ssh into lab computers and run X software. They're not sysadmins. Many aren't even power users. They just know that if they ssh from their laptops (be it Mac OS X or Windows with an X server installed), they can simply run an app and it works. They don't care if it's slightly more smoothy drawn; they just want to get their work done. This guy's girlfriend (who doesn't even know what X is) X forwards Thunderbird from her desktop to her laptop. One of the libraries in the Australian Museum in Sydney has (or had, I haven't been there in a while) a row of thin clients which X forward web browsers from a server in some back room. These are real world examples of non-sysadmins using X forwarding.

But I don't really care if it's sysadmins or non-sysadmins using this. The fact is, the people behind Wayland are removing functionality from a very core part of a Linux system, largely so they can provide slightly prettier graphics. This is not a tradeoff that should be made for any feature, whether it's used by sysadmins, or scientists, or writers, or artists.

And it's not true that X will always be there. X is being kept around solely to help the transition. What do all these people do once apps start becoming Wayland-only?

Re:Giant Step Backwards

[sl3xd]

Grow up. Most people are not sysadmins. People who are not sysadmins and in a unix environment are most likely using Linux.

You seem to be more than happy to claim (with absolutely nothing to back up your claims) that only a tiny number of users actually care about network transparency, about the number of xBSD users using it as a desktop, and so on.

You are no more capable to speak for any group of people than I; making gradiose claims as if you speak for "the majority" does not give the impression of maturity.

You rant about sysadmins, but you know what? I'm not a sysadmin. I'm a normal Linux user; and every other "normal" Linux user I know - every single one - needs network transparent windowing for their day to day work. Is it a case of selection bias? Quite possibly. I'm only qualified to state my own observation - that nearly every one of the hundreds of Linux users I've ever interacted with want or need network transparency in the windowing system. Only a couple are sysadmins.

As a result, I obviously find it laughable to see anyone claim that only a fraction of a percent of desktop Linux users want or would find network transparency useful. I'd personally guess it's somewhere between a quarter and half of desktop Linux users.

As far as "just keep using X.org"... orphaned software isn't terribly useful. Distributions stop providing it, and orphaned software often won't even compile it in as little as a year, depending on the march of the toolchains. The forward looking thing to do is to petition the Wayland developers (many of whom are also Xorg developers) to provide network transparency. It's not an unreasonable request; since they are also Xorg developers, they're far better suited to designing an efficient protocol than nearly anybody else on the planet.

The penalty placed on those who don't use network transparency is exactly nothing. Network transparency will not in any way hurt those that don't need it; that's the entire point of dynamic libraries.

Re:Giant Step Backwards

[dbIII]

And I'm sick and tired of admins thinking they are the world

You've got it wrong, they are just more aware of what the world is using than a single guy behind a single screen that doesn't know what anyone else around him is using.

If Linux is rare on the desktop, the rest of the "community", by which you must mean BSD, is truly miniscule

X also runs on Microsoft Windows so your petty little bagging of a small group because they are small is irrelevant.

It's amusing that as an isolated individual you are criticising a group because they are "small". I can see here that the subject matter is irrelevant and you could be saying this about any topic and dragging in an imaginary "majority" that you pretend support your viewpoint. Please act like an adult and put your own views forward as your own and treat others as human beings instead of mere strawmen to insult.
I think the point the poster above was trying to get across (but if it isn't it's my view anyway), is that without network transparency there's not really a lot of reason to use linux at all, you just get the cut down MS Windows experience of being limited to a single machine with your window on the world being a web browser, VNC or similar limited little thing instead of being able to run anything anywhere (within the limits of latency, bandwidth etc). Even in the age of examples such as Siri and the "cloud" hype you are talking about a step backwards to the 1980s and single user non-networked systems in my opinion while the rest of the world is finally noticing that network transparency is a good idea.

Re:Giant Step Backwards

[Anonymous Coward]

What it might allow you to do is use the full desktop remotely without full bitmap transfer (ala RDP).

Haha, you're funny. You think RDP is bitmap based like VNC.

Re:Giant Step Backwards

[hawk]

Perhaps Xflash Xorgon will save us?

hawk

Re:Giant Step Backwards

[Anonymous Coward]

If a gtk application is compiled for wayland, will it stilll work as an X client?

Is the same true for other toolkits?

UGH.

[b5bartender]

A shaky 15-minute amateur video of a powerpoint presentation? WHY

Re:UGH.

[fahrbot-bot]

For a more authentic "Blair Witch" / "Cloverfield" documentary vibe.
No comment on whether either film is an apt metaphor for the projects themselves.

We're NOT talking

[Anonymous Coward]

the ordinary VNC protocol here. (I wish people would stop saying VNC in this context.) The trivial remoting solution for Wayland is VNC-like, that is "screen scraping" (or window scraping) -- just like Remote Desktop which is the only way of doing remoting which doesn't suck utterly.

Re:We're NOT talking

[Electricity Likes Me]

Remote Desktop on Windows seems to be somewhat more intelligent then just "screen scraping".

Similarly, VNC on Linux can be like this.

But by far the best solution I've encountered is x2go, which is based on NX, which notably just caches and removes some of the overhead of straight networked X.

Surely, the better answer here would be to tackle that problem - implement a well behaved network protocol.

Re:We're NOT talking

[Anomalyst]

The x2go web page (http://www.x2go.org/doku.php/start) is a mess, the download page link is broken as well as numerous others e.g. http://plugin.x2go.org/ I hope their codebase is better maintained (doubtful, its not in git).

Re:We're NOT talking

[DarwinSurvivor]

Except that it requires the x server (or wayland/etc) to be RUNNING on the remote machine. With remote X, I can cold-boot a machine that doesn't even have x11 *installed*, and then ssh -X into it and run any gui application I want, without EVER having to install OR RUN x11 on the remote machine. THAT is the true power of remote X.

Re:We're NOT talking

[Anonymous Coward]

http://www.x2go.org/doku.php/wiki:development:git

Of course, Mercurial is nice too...

Re:We're NOT talking

[jbolden]

No you can't. You need the X11 client installed to talk the X protocol. You don't need to be running a desktop (X11 server) but you absolutely need X11.

Re:We're NOT talking

[Electricity Likes Me]

That big message at the top that they're currently trying to consolidate the page might be notable. x2go is in the Ubuntu reps, and the Windows client works great. Also the download links from the normal site http://www.x2go.org work fine - I don't actually know how you navigated otherwise. Some metaphor about don't judge a book by it's cover would also be appropriate here.

I highly recommend giving it a go if you have Linux servers to maintain, since it's connections work like straight SSH connections (no additional ports or authentication mechanisms - which more or less seamlessly fixes problems with both VNC and the original NX) and it's very responsive and usable.

Re:We're NOT talking

[DarwinSurvivor]

Not all of it, just the portion of the libraries used by the clients.

Re:We're NOT talking

[bmullan]

Documentation always suffers for most software and x2go is no different I guess.
x2go's website has been both in migration to a new site, a new platform all while the x2go server/client architecture has been under a rocketing development of new features & capabilites in the past 12 months or so.
Best advice is to join either the x2go user or x2go dev mailer's until the new website gets more updates.
- x2go-user@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/x2go-user
- x2go-dev@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/x2go-dev

neckbeard rage!

[not already in use]

I'm always entertained by diehard linux folks who simultaneously desire greater adoption of linux on the desktop while wildly opposing any actual steps to make it a reality.

Re:neckbeard rage!

[Anonymous Coward]

I'm always entertained by diehard linux folks who simultaneously desire greater adoption of linux on the desktop while wildly opposing any actual steps to make it a reality.

You are talking about two totally distinct groups as if they were one and the same.

I don't care about Linux adoption by the general desktop or phone using public, I just want a powerful OS that puts me in control of my machine.If NeckbeardOS gives me more control, I will happily use that.

Re:I don't understand

[eobanb]

In Unix/X parlance, the 'window manager' is distinct from, and higher-level than, the 'windowing system'. XFree86 and X.org are display servers like Wayland but have also taken over the job of being a windowing system. In contrast, examples of window managers are twm, openbox, compiz, etc.

thoughts

[Anonymous Coward]

If anyone actually read up on wayland, they would see that you can plug RDP or VNC or whatever you want in for remote access. SSH in an forward VNC over the session, or RDP over the session. Same effect. In my opinion, this effectively negates the argument about network transparency. I find that X over a WAN connection is completely useless anyway.

X has a lot of legacy code, and a lot of layers. A new display server/interface is what is needed to get the linux desktop up to performance par with Windows7 and OSX. I'm not trying to start a thread war here, but X without a compositor feels very old and X with a compositor feels very laggy. Wayland is supposed to solve this and still provide for backwards compatibility with apps requiring X.

Re:thoughts

[nurb432]

VNC or RDP does not replace the network transparency of X. it would just be a terrible kludge to get around lost core functionality.

Re:thoughts

[DarwinSurvivor]

The big reason remote X is useful (which will be 100% impossible with wayland) is that you can install an X11-compatible application WITHOUT installing X11 and still run that application via remote X. This means that headless servers that don't even *have* a graphics card can still have gui tools installed and all you need is a client with an X server that can run "ssh -X".

Re:thoughts

[snadrus]

I thought X11 network transparency (now that toolkits forward whole bitmaps to X) basically are VNC.

Re:thoughts

[JesseMcDonald]

VNC or RDP does not replace the network transparency of X. it would just be a terrible kludge to get around lost core functionality.

That "core functionality" is something which no modern toolkit actually uses. Remote rendering through X11 these days consists almost entirely of clients rendering into bitmaps and sending them, uncompressed, over the network to the server. VNC or RDP offers the same thing, but with less cruft and better compression. The XRender extension (which is not part of the core protocol) offers some server-side rendering, but its capabilities are limited.

I agree that true remote rendering might be able to beat VNC/RDP-style image transfer, but for a general solution that would require a proper remote interface to Gallium3D or OpenGL, which could be adapted to work as well (or better—less overhead) with Wayland as with X11.

Re:thoughts

[spitzak]

I think you will find that unless you install Xlib and glx and Qt and a bunch of other libraries, your headless server will not run that remote display.

It certainly would be nice if a wayland client would look at an environment variable and set up a remote connection to a wayland server without having to pre-run any software on the client end. But considering this is *exactly* what Xlib clients do, I don't see this as being impossible to solve.

Re:thoughts

[Anonymous Coward]

That "core functionality" is something which no modern toolkit actually uses. Remote rendering through X11 these days consists almost entirely of clients rendering into bitmaps and sending them, uncompressed, over the network to the server. VNC or RDP offers the same thing, but with less cruft and better compression.

Not if the main thing rendered over remote X is text. I often have X terms running on 20 different machines. And yes I am a system manager.
I think it would be unfortunate if one of the main tools I need to do my job was taken away from me.

Re:thoughts

[Daniel Phillips]

You are wrong about needing GLX and QT on the remote machine. Xlib, yes, it must be installed on the remote machine before you can run any X application remotely. Otherwise, to what library would the application link?

Re:thoughts

[Anonymous Coward]

All you need to install an X application are the nine billion X11 libraries and their myriad dependencies. Hardly less than the whole of X in fact... the server is tiny in comparison.

Re:thoughts

[DarwinSurvivor]

True, but you don't need to have the bloody thing RUNNING (hdd space is MUCH cheaper than memory)!

Re:thoughts

[Anonymous Coward]

Keep using xorg then, the other 99.99999999999999% of UNIX users will use wayland.

Comment removed

[account_deleted]

Comment removed based on user account deletion

The problem Wayland attempts to solve

[Ponder]

is to provide display management for linux devices that generally do not require network transparancy such as phones and tablets and which are resource constrained so the bloat of a full xorg stack is unacceptible. Clearly Ubuntu which has designs on becoming the tablet king is embracing this - Fedora also has an interest because it is the basis of the olpc, raspberry pi and other lightweight device spins. The obvious simple way to support network transparency is to run an X server as a Wayland app and this works fine so backward compatibility is easy to provide in fact Gnome is adding westin support into mutter so apps will use wayland if available and X if not. Going forward adding network transparency nativly to wayland is a fairly trivial and can be implemented more efficiently than X - according to the developers.

So:
Plusses: smaller leaner and simpler code base, backward compatibility for legacy X apps, possibilty of network transparency not based on what was state of the art 30 years ago. Tight integration into linux.
Minuses: linux only (possibly), some pain in the transition possible while support is added to distros. Developers currently focussed on solving specific problems for Tizen.

   

Re:The problem Wayland attempts to solve

[Anonymous Coward]

Phones and tablets /do/ require network transparency, much more than any other device!

Picture this: I take out a phone or tablet, ssh -X into a server somewhere, and run some high-performance program with a gui for monitoring. Now I'm able to use the full power of the server or HPC cluster from my tablet or phone.

If I can't do that then I won't buy a tablet or phone, it's simply not worth the effort if it just becomes a low-powered handheld pc.

Re:The problem Wayland attempts to solve

[Anonymous Coward]

Yeah, but you're looking at it backwards -- since Wayland has an X11 compatibility layer, it will do that just fine.

The opposite case -- running apps meant for your local display to a remote display -- is sometimes useful (e.g. working around a damaged screen, though VNC is really better for this use case), but not often.

(Not a big Wayland fan myself, just pointing out how your particular argument is wrong)

Re:The problem Wayland attempts to solve

[Ponder]

Possibly

But such a requirement is a niche and not the use case of 95% of users and as the other poster adds is easily accomodated by running a specialed X server on top of wayland.

Re:The problem Wayland attempts to solve

[Anonymous Coward]

On most desktop systems X is run with -nolisten tcp, so it's not only about "tablet" for quite a long time now.

Re:The problem Wayland attempts to solve

[Daniel Phillips]

Plusses: smaller leaner and simpler code base, backward compatibility for legacy X apps, possibilty of network transparency not based on what was state of the art 30 years ago. Tight integration into linux.

I want more than the possibility of network transparency, I want it to be a proven, reliable fact. And general plaudits about smaller and leaner are uninteresting, I want to know specifically what the design improvements are, not in terms of generalities, but specific, verifiable assertions.

Tight integration into Linux, what is that? Anything different from "actually works"?

Comment removed

[account_deleted]

Comment removed based on user account deletion

The conspiracy aspect

[Casandro]

Giving it a conspiracy look, one might interpret it as an attempt of the unwashed hordes trying to take over Linux. Instead of going the sane unix-like route of moving graphics into virtual filesystems, like it's done on Plan9, they want to essentially replicate X11, but without its good parts. Wayland is not network transparent. It still needs libraries to be linked with your software. What you get is a bit more gimicky graphics, but you loose a lot of important features.

It's perhaps not quite as bad as the proposed move to binary log files, but still it offers next to no advantage for quite a bit of cost.

Replication of functionality between the kernel and X11 could be elimnated easily, buy building an X11 server that accesses the kernel.

Re:The conspiracy aspect

[jbolden]

No one has any idea how to make an X11 server reliable enough and secure enough to have access to the kernel on the majority of machines. Further the kernel group has never expressed interest in allowing a change like that. Lets stick to discussing technology we know how to make, otherwise I can discuss my windowing system with built in teleportation to resolve the network transparency need.

Re:The conspiracy aspect

[Casandro]

Well, uhm, it currently has access to the hardware. If the hardware abstraction would move from the X-Server to the Kernel, it would actually probably be safer.

Nobody is talking about putting an X-Server into the Kernel, and even Plan9 with its filesystem based GUI doesn't. (On Plan9 normal programes can create virtual filesystem structures which can be mounted anywhere. It's the Plan9 way of providing an API)

Compatability mode

[Picass0]

My understanding is Wayland is not ready for prime-time just yet, but when distros start using it will have X11 legacy support to handle stuff like this.

By putting a 1.0 out there they are telling developers of X11 apps to start planning/testing/adapting.

Quite exciting times ahead

[nurb432]

That isn't the term i would use.

Loss of network transparencey

[nurb432]

Its more than more complex, its makes the entire thing totally useless and a REAL bad idea.

Great if they want to clean things up, but to go a completely different direction at this stage of the game is ludicrous.

X works. Don't break it.

tl;dr? 60 FPS on an embedded GPU!

[WilliamBaughman]

Discussions of Wayland on Slashdot tend to be all about a lack of network forwarding or missing features, so I think I'll share some of the positive things I expect to see from Wayland:

• Fewer CPU cycles spent in the graphics stack, shortening time to sleep
• Less memory used by the graphics stack
• More efficient compositing, meaning less memory bandwidth used in memcpy routines, lowering DRAM power and greatly improving speed in certain scenarios
• A graphics stack that's fast enough to get 60 FPS scrolling on an embedded GPU

I'm not aware of any X.org implementation that's gotten 60 FPS on an embedded GPU. That's not me trying to knock X.org, say anyone should stop using it, or say people need to "upgrade" to Wayland before it's feature-complete. That's me recognizing the reality of X.org not being "one size fits all" in a world where embedded or mobile Linux (think Android) outsells (and out-deploys) Linux on big core 10 or maybe 100-to-1.

Disclaimer: A big part of my job of performance optimization of applications on Linux running on mobile devices.

linux kernel need a rejig

[Anonymous Coward]

lot of comments here points to the fact that the entire linux kernel in current state is nearing expiry date and need a full rejig.

Re:linux kernel need a rejig

[Daniel Phillips]

Go away troll.

Sorry.

[Anonymous Coward]

With remote X, XLib or GTK or Qt or whatever still needs to be installed on the server for such things to work. (Hardly surpising if you think about it.)

Re:tl;dr? 60 FPS on an embedded GPU!

[Viol8]

"A graphics stack that's fast enough to get 60 FPS scrolling on an embedded GPU"

If you want that sort of performance on a limited power device just write to the framebuffer directly. X doesn't have a monopoly on talking to the graphics hardware.

Re:I don't understand

[Alex Belits]

Same as how X11 is "implemented" on Android. It's absolutely useless.

I have my Nokia N900 running X11-based Maemo for years -- with compositing, remote X, phone-friendly window management of remote and local applications, etc.

Re:I don't understand

[Alex Belits]

If their implementation was compliant with X standard, it would be possible to just re-use any X window manager. And yes, that would require Wayland window to be manageable from X -- otherwise it's a shit implementation similar to "We formally support POSIX, give us government contracts!" in Windows NT.

Actually it's easy

[snadrus]

A compositor plug-in can track changes & present this using VNC server protocol. It doesn't exist yet because it's not the focus, but it's fairly easy. Then the VNC Client under Wayland works like most any other client. Someone started on it for a Google Summer of Code project last year, but didn't finish.

RDP

[Anonymous Coward]

RDP wasn't much more intelligent than simple screen scraping at the time I was using it -- it may have changed since. It was still far better than X11 or VNC. It just avoided 1) being synchronous (as many programs on X11 are), and 2) the absurd JPG encoding which (AFAICT) is/was default in all the VNC implementations on Linux.

NX was a huge improvement over X11, but you couldn't take over an existing session (like in RDP), so that was basically useless for my use case.

But all of this is beside the point: The VNC-like protocol which is trivial to implement for Wayland is [b]fine[/b] for most cases (and [b]much[/b] better than the remote X11) and Wayland in no way prevents an even more efficient protocol from being implemented at the toolkit (or perhaps slightly lower) level.

I was just commenting on the error of assuming that we're talking "VNC" here.

Re:RDP

[Anonymous Coward]

I don't think RDP has ever been a screen scraper (unlike VNC). RDP was intercepting drawing commands at the GDI layer and sending them to the connected client to draw in their own buffer. Bitmap caching came along to support the DrawImage() series of API calls. And, btw, the first implementations of VNC used raw and RLE encoding - JPEG came along when idiot Windows developers with their analogue-VGA-cabled monitors started getting involved (they couldn't tell it looked shit because, well, everything looked shit on their screens already).

Re:I don't understand

[spitzak]

In Wayland the client draws the window borders (and before you blurt out "but they will be different from each other and 'confuse the user'" be aware that nothing on X makes anything other than the window borders draw identically, but somehow there are all these applications where they draw their own scrollbars and buttons and text editors and they all look alike and fail to "confuse the user".)

X ICCCM window management was invented before there were shared libraries, so now there is a huge messy complex api built on top of the limited atom+value and signalling that existed in X11. Having written a toolkit (fltk) I can confirm that talking to the window manager literally takes 10 to 20 times as much code as if I drew the window borders myself. And we still have horrific problems where various window managers disagree on whether a window should vanish when the app loses focus and how "stay on top" should work, and we can't put the most trivial buttons or items in the window pulldown menu, something Windows allows and is making it increasingly difficult to port applications from Windows.

Wayland wisely has decided to scrap this mess by having the clients draw the borders.

Emulating X is trivial if you just work with the windows the X client knows about and repliate them on Wayland. The problem is that all that ICCCM stuff will need to be interpreted by something and draw window borders as well. It is not clear if this should be done by a local "act like xlib" library, or by a fake X server, or by an actual X11 window manager running on the fake X server. That's why window management is currently a "hack" because this has not been decided.

Why no X12?

[manicpop]

http://www.x.org/wiki/Development/X12 Even this page on X.org lists a lot of great reasons why X11 is outdated and needs to replaced, yet I don't know of any serious project to create a new, modern X. All X development is focused around fixing and extending X11. I guess X11 has become so big and so universal that there's no real desire to tear it down and start from scratch.

Comment removed

[account_deleted]

Comment removed based on user account deletion

It was working in 2000, why not 2012?

[dbIII]

I was running OpenGL stuff from a IRIX box in the next building on a fairly low end linux box with a low end 3D card in 2000, and it worked better than you describe. Just about everything that runs X now, including any recent purpose built terminal, is a hell of a lot better than that now.
My desktop machine at work now is getting a bit old so I'm actually posting this using firefox on a fairly fast cluster node via X. While that's not 3D some of the geophysics stuff my co-workers use involves a lot of 3D content, and I've played with Blender on some nodes during setup to see how they handle it. That's with Nvidia's accelerated driver in nearly every case.
By the way, what's the "modern CAD system"? I'm only familiar with the old stuff (I'm putting the most recent AutoCAD in that category since it hasn't really changed in years and is low end anyway) so you may have a point I'm not aware of.

Re:It was working in 2000, why not 2012?

[tibit]

Any parametric CAD system. Alibre Design would be one example (it is good and cheap, although, sadly, it doesn't run on OpenGL). I have some assemblies that consist of a couple thousand parts. Those parts aren't washers either, even though a washer doesn't exactly have the most trivial geometry. I've been trying to get it to run on wine for the a bit a year or two ago, and I eventually coaxed it to get as far as trying to display some geometry, so I could see how many calls into wine's directx implementation it'd do, and how much data it'd try to push. It was a couple hundred megabytes over a couple tens of thousands of calls.

Never mind that remote screens don't provide any DRI, so if you're using Mesa it probably won't be accelerated even if you have a high-end 3D card on the display. I'd like to be wrong on that one, but I'd think that Mesa needs DRI to use 3D hardware.

Re:It was working in 2000, why not 2012?

[Unknown Lamer]

Indirect GLX rendering has been accelerated for something like ... five years now (Google for AIGLX). There shouldn't be too much overhead versus running locally -- bandwidth, obviously, is going to be a limiting factor for texture heavy applications, but once the commands etc. hit the X server everything goes through the same path as local rendering.

It used to be that OpenGL apps used DRI to directly send commands to the graphics card. Those were the bad old days when video games could crash the entire display system but hosing the graphics card... and you couldn't really run more than one DRI client at a time. It was a mess.

Re:It was working in 2000, why not 2012?

[tibit]

OK, so DRI is of historical relevance at this point. Good to know.

Re:It was working in 2000, why not 2012?

[Unknown Lamer]

Sort of, DRI is now DRI2, and still used... but the server itself is doing the direct rendering, and clients are doing GLX through the server. Which is much saner since you consolidate buffer management and only the X server is capable of hosing the entire system.

I find it amusing that the desire for fancy desktop effects got us working AIGLX ... you'd have thought folks doing CAD or something would. Oh well, I guess the compiz loving people did something good at least :)

A window manager is just another application to X

[dbIII]

You can also run window managers over X, remotely or locally. In 2001 I was running cygwin on a win2k or NT box and handling the windows related to X from a linux box running enlightenment (no shaped windows, but that was due to cygwin not having support for those at the time). Presumably Wayland should be able to call up the window manager in gnome, kde or whatever until they've got their own window manager that matches the way everything else in Wayland behaves.

Wayland has massive potential

[Trogre]

Wayland is an awesome project and shows immense promise to supplant the outdated X11. I especially like core concepts such as making each redraw pixel perfect and tight integration with vsync.

However all this will be for absolutely nothing until Wayland has integrated network transparency. Without that it's just another unusable toy.

Re:Wayland has massive potential

[Brian Feldman]

I've used Unix as my professional development platform daily for a decade. I hardly ever use X remotely, and when I do, it works very poorly because the protocol is not designed for the reality of modern rendering. It is extremely slow and ugly even on a "fast" link.

It happens again

[aetjeajewrj]

"Those who don't understand Unix are condemned to reinvent it, poorly." – Henry Spencer http://en.wikipedia.org/wiki/Unix_philosophy

Widelands and Wayland

[luk3Z]

Wayland - my 1st thought was another similar to Widelands game...

X or OpenGL is not the culprit then

[dbIII]

it is good and cheap, although, sadly, it doesn't run on OpenGL

Then it's obviously not designed for remote access of a 3D program and is about 15-20 years behind a "modern" system if it's designed to be used remotely.
There's hacks to get around that which are designed to get around having no 3D accelerated hardware at the desktop end but the best one (VirtualGL) involves putting decent graphics hardware on the server and you may as well just run a monitor extension cable over ethernet from the server room instead of screen scraping and sending it over the network, unless of course you need a few instances.

if you're using Mesa it probably won't be accelerated

If there's no OpenGL then Mesa is never going to be used at all let alone help one way or another. It comes down to the directx in Wine.

Sorry, but this is looking a bit like a deliberate strawman since you are writing about a program not designed for remote access and blaming that failure on X. If it's not compatible you need to either use it as designed (on the workstation) or work around the problem with hacks (accelarated graphics hardware on the server and efficient screen scraping).

I've been trying to get it to run on wine

Even worse, you appear to be blaming X for something never designed for remote access or even that platform at all and it's running through an imperfect compatibility layer. How do you know it's not Wine that's at fault? It doesn't work with every MS Windows program. Sorry, what exactly is your point because your example really has nothing at all to do with your comment above about the limitations of hardware acceleration on X. Your example is not even using hardware acceleration on X so has nothing at all to do with your comment about hardware acceleration on X above.

Germanic

[Anonymous Coward]

English is a Germanic language.

Easy to add

[Anonymous Coward]

According to TFV it is easy to add network transparency to wayland. They just haven't done it yet.

But the wayland FAQ says that it is a fairly complicated task
http://wayland.freedesktop.org/faq.html#heading_toc_j_8

Which is right and which is wrong?