Slashdot Mirror
Wayland 1.0 Released, Not Yet Ready To Replace X11
An anonymous reader writes "After being talked about for four years, Wayland 1.0 was released today. The Wayland 1.0 release doesn't mark it yet as being ready for Linux desktop usage but just being API/protocol stable for future expansion. Wayland will now maintain backwards compatibility going forward, but how much longer will it take to replace X11 on the Linux desktop? Quite a while seems likely."
Perhaps this is going to be ready when the IPv6 flag day is?
This sig no verb.
I guess all of the X11 developers that are the ones developing Wayland don't understand X11 that well. You should send them a link to the classes you're teaching.
Oh, we'll call it 1.0, but I'm sure distros won't start shipping it by default because we said it's not ready. Why are you pulling that face? Worked for KDE didn't it?
I did, at least, skim the article, and I still don't know. Didn't X11 just (as in last few years) get replaced with x.org? This is another replacement already? Ok, before posting, I google and see this:
... [An X server has] a tremendous amount of functionality that you must support to claim to speak the X protocol, yet nobody will ever use this. ... This includes code tables, glyph rasterization and caching, XLFDs (seriously, XLFDs!) Also, the entire core rendering API that lets you draw stippled lines, polygons, wide arcs and many more state-of-the-1980s style graphics primitives. For many things we've been able to keep the X.org server modern by adding extension such as XRandR, XRender and COMPOSITE ... With Wayland we can move the X server and all its legacy technology to an optional code path. Getting to a point where the X server is a compatibility option instead of the core rendering system will take a while, but we'll never get there if [we] don't plan for it.
What's different now is that a lot of infrastructure has moved from the X server into the kernel (memory management, command scheduling, mode setting) or libraries (cairo, pixman, freetype, fontconfig, pango etc) and there is very little left that has to happen in a central server process.
which bored me to tears, so I'm no longer interested, but for those who are....
Free the Quark 3 from asymptotic confinement! Bring your charm! Don't get down! All colours and flavours welcome!
Actually it sort of does. Theres so much legacy garbage in the current X protocol that just wastes resources and makes it nearly impossible to maintain.
As I don't actively maintain X I have no stake in this, but if it helps to motivate new generations of developers to dive into thats great.
I can also note that I think the model Wayland is employing seems more sane, given what we've learned over the years of model and X driver development (DRI anyone?).
Not you again. Wayland is being built by X developers, and several other prominent X developers have said it is a good thing.
Would you say that you understand X11 A) better, or B) worse, than the people who actually write code for Xorg?
I've used various Linux distros for 13+ years, and have always been disappointed in the horrible desktop graphics. Compared to a Windows machine, a comparable Linux rig is normally a miserable graphics experience. I was shocked the first time I got my hands on a Xoom tablet - the graphics are by FAR one of the best features of the Android platform. I can't for the life of me understand how there can be such an extreme difference between graphics in desktop linux distros compared to the gorgeous and snappy graphics under Android. And I'm not even running garbage graphic cards - one of my machines has a very modern nVidia card that plays all the latest games under Windows, but still handles poorly on Linux. From a technical perspective, I can't understand how there could be such an absurd difference, other than that the Android kernel is a fork, and clearly the Android kernel developers are far more concerned about creating a pleasant and useful graphical experience?
Imagine recreating xlib so it doesn't communicate with an X server but directly draws things on the local screen, maybe in a multithreaded fashion. In such a scenario, the ability to share a display between many programs would be lost or alternatively a badly behaviored program could disrupt the other's windows. What kind of increased performance would be obtained (if any) by replacing IPC, as used in X11 (and in wayland too?) for drawing and use in process fonctions instead?
Try never. Yes, I know that it should be possible to write a Wayland client that provides X11 server capability, but in that case, it is the Wayland client that is replacing X11, not Wayland.
Seriously, though, the Wayland effort appears to be throwing out every advantage the X11 display had over the Windows display for a replacement that will probably never be quite as good as a Windows. I just hope that developers of programs which currently support X11 continue to support X11, or my life will get much more difficult. In fact, for much of what I do, without X11 support (and only Wayland display supported), I would probably be better off with a Windows desktop instead of a Linux desktop.
The real "Libtards" are the Libertarians!
He already said years ago that he intends for Ubuntu to use it as soon as it is ready. I expect Ubuntu will likely be the first major distro to do so.
http://blindscribblings.com - Tasty pop-culture in conceptual fashion.
Yep, the issue with constantly pushing forwards and looking for the next new thing means that you can periodically make a bad choice. Shuttleworth, while some of his descisions haven't been the best, has been instrumental in pushing the Linux desktop to where it is today. Linux has never enjoyed so many desktop users. That brings good and bad, but its still an overall positive.
Unfortunately, developers are the second worst group for understanding what users need. The only part of X11 that people need is network transparent remote display, but that's the one part that the developers are absolutely hell bent on removing.
When you consider how fast the switch to x.org from xfree86 took place in three Linux world, any clearly superior x-like implementation with fully compatible APIs and without unacceptable license encumbrances could be adopted in very short order. If the functionality requires every program and every library to be reworked, then it will probably never happen.
Not having ANY real Wayland knowledge, I can only hope it is not another change in Linux for change's sake.
Sig Battery depleted. Reverting to safe mode.
The important feature about X Windows was Network Transparency - You could run an application on one computer with its screen output and keyboard and mouse input on a different computer. Sure, there are other ways to do it - lots of ssh sessions, or web browsing (especially with AJAX etc.), or competing window systems like NeWS, or screen emulators like VNC and Windows Remote Desktop - but fundamentally it's a lot cleaner to have some kind of network-transparent window system than to have an application need to drive a "screen" on its own machine.
25 years later, do we still need this? Yes! Virtual machines are taking over the computer business, so you can't expect the application to be running on your desktop (even if it _is_ running in a VM on top of your desktop), screens are a wide range of different sizes and capabilities (laptops, tablets, big monitors, etc., which often don't resemble the machine the app is running on), web browsers are getting used in increasingly complex ways because Windows didn't have a convenient X interface, and there's more and more ugliness around, and more waste of resources trying to emulate things that X did adequately well.
There are lots of good reasons to replace X, but Network Transparency is still the core feature, even if you want the application to have more control over the screen and its associated hardware than we had back in the 1980s, or if you want to move processing functions to different points between the client and the server (e.g. NeWS and NeXT's Display Postscript did some things differently, and Plan 9 and its successors had their own opinions about how to implement everything), but if Wayland doesn't offer Network Transparency yet, it's not an adequate X replacement.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
I can also note that I think the model Wayland is employing seems more sane, given what we've learned over the years of model and X driver development
Well I can note that I think client-side decorations is so much bollocks.
Wayland seems to abdicate responsibility on a lot of stuff that really needs to work across the desktop consistently, and that X was solving commendably.
Well, it certainly doesn't need replacing with another X server. What it needs replacing with is something a little less complex and wonderful.
+1. No remote display, no dice.
M.any years ago they were talking about Enlightenment wanting to replace X11
While E has gone to E17(still beta), it's not replacing X11 yet.
I dunno much about the project that's the topic of TFA, so I won't know how successful it would be in replacing X11
Muchas Gracias, Señor Edward Snowden !
A)
no question. I use the code, they make it. car analogy: who understands a car better (what it should do and how): the engineer or the driver?
AB HOC POSSUM VIDERE DOMUM TUUM
It's native to the windowing system on both the local and remote machine and there is absolutely nothing I need to do to either system to pop up a remote display other than insert a 'DISPLAY=remote-host:0.0' in front of the command line.
I guess all of the X11 developers that are the ones developing Wayland don't understand X11 that well.
"Don't understand X11 that well", I don't know (though I'd doubt they don't understand X). As for whether X11 needs replacing; not here, no.
I do worry that they're ignoring the ideals of X, and focusing too much on cellphone/*pad interfaces, as opposed to the original X ideal (any device *should* be able to handle it, regardless of what your particular prejudices might be).
Does Fluxbox run on Wayland? If so, I'll try it.
"Tongue tied and twisted, just an Earth bound misfit
X11 remote display can be added via a daemon. Like it is with OS X, Windows, etc.
I run: Windows, OS X, Linux, FreeBSD. Just because you have a hammer, doesn't mean everything is a nail.
So.. X11 needs to be replaced with something shiny because of...? Seems the response most often stated is the code base is a mess. Why not just clean that up instead? And has others have noted here, remote usage is still important and no, VNC is not the same thing.
Theres so much legacy garbage in the current X protocol that just wastes resources and makes it nearly impossible to maintain.
I've heard much the same thing said about rxvt vs. xterm. I like xterm, and its devs have done brilliant stuff over the years keeping it relevant, fast, and working. rxvt does 99.999% of what xterm does, and does it well (except for corner cases).
Why do I feel like I'm stirring entrails?
"Tongue tied and twisted, just an Earth bound misfit
...network abstraction layers to their GUIs, specifically to enable people to mirror and share their desktops efficiently.
The mundane, non-vertically integrated *nix world still doesn't have this ability after all these years. X11 can't mirror/share, its also laggy over broadband connections, VNC is primitive and slow, etc. The NoMachine people have claimed they can support screen sharing in NX, but I haven't seen a working example yet (and those features are in the proprietary version anyway); otherwise they did a good job of making X11 usably network transparent for use cases not contained within a single LAN (i.e. most situations).
Personally, I'm tired of seeing all the hand-waving about X11's network transparency. It doesn't help in the vast majority of instances where people want to share an app or screen during a teleconference. X11 is not advanced in this respect -- just sadly out of touch. It mainly addresses the rather outdated use case where you have a handful of engineer types who open a CLI and type an ssh command, possibly fiddle with the display variable, then type in the desired app as a command so they can run one expensive, customized app on a server in a specially cooled room 3 floors down.
Or, we could just keep X. Much better.
Remote X is a pig once ping times increase due to the number of round trips required. VNC is far superior in that situation especially when it can be notified of changes efficiently.
Remote vanilla X11 over even today's internet is fucking slow as balls and about as painful as being kicked in them. Even on a lan it can be pain if your client is generating too many unneeded events (try eclipse over ssh x11 forwarding some time). X11 does facilitate the use of, say, NX and VNC though to mitigate the problem. Does Wayland? If so then bring it on I say. As long as the end result is the same or better, how is it a bad thing?
Unfortunately, this is not what people are objecting to. The problem is not that you cannot map X11 on-top of Wayland windows and run remote applications displaying on Wayland, the problem is that you cannot transparently display Wayland-windows on another machine.
On OS X and on Windows, it is not possible to display the native apps on a remote machine using X11.
The lack of network transparency will lead to every developer needing to ship two different applications, one linked to Wayland and the other to X11, then everyone need to download the correct app and install the one depending on their machine, i.e. on the machine that will be accessed remotely, install the X11 version on others install the Wayland version. Sure, we all know that you usually use GTK/Qt for applications, and while some of the issues could be solved by the GTK and Qt developers by cleaver dynamic linking to the display manager, often this works so-so and some apps will not be compatible with this, so you still need to run double testing for all your applications.
They should have started with defining the Wayland system in a way that was network-transparent, they could even have updated the protocols and made it more suitable for modern applications (i.e instead of sending delta-bitmaps, send the scenegraph updates for the window instead).
In a word, money.
Desktop Linux generates very little on its own. Red Hat manages to rake in about a billion dollars in revenue, but Canonical only manages $30 million (wikipedia). Android doesn't have to make money per se as it has other avenues of monetization (e.g. advertising). It also lives or dies by its UI.
You can also contrast this to the embedded Linux, server, and HPC markets, for which I don't have numbers, but Linux has a dominant position.
Linux is more often optimized for workflow than graphics. The former is generally excellent and highly configurable. While Linux advocates like to tout the number of individual contributors to the kernel, a large number of these are corporate contributors wishing for anonymity. There's not a lot of ROI in shiny buttons. Also, low-end graphics run on more platforms.
That said, I like the way my desktop looks (earth image with weather and sunlight, updates every 1/2 hour). Dunno what you object to with yours, but it can probably be remedied if you have some free time.
Those who advocate genocide deserve every protection afforded by law, and none afforded by common human decency.
Somewhere in that supposed .0001% is things like, perform faster under heavy system load. I've tried dozens of terminals over the years and it never takes more than a couple weeks to send screaming back to xterm for one reason or another.
As an X11 user (developer of client apps) I would say I understand the X11 features that I love better than X.org developers.
Note that X.org rewrote the implementation of X11, but they aren't the ones that designed it, and re-implementing well doesn't automatically mean they will have insights into the API design, or meet my needs better with the new one.
All of that may be true, but for me I'm way more likely to open on a remote display when both hosts are on the same LAN. And why would I want to slow it down with SSH when it is all local?
An X11-aware app like emacs even has it in the default GUI menu to open a "frame" on a remote X. It is nice, I can have the same buffer open on two computers at opposite sides of the room, whichever end of the room I am at I can type code. No syncing, no fuss.
I don't really care whatever fancy new what-the-what they have that they love, I'll switch when there is no longer an X11 that compiles on a modern *nix. And that will be, oh, never.
Never say never :)
I hear you, honestly, but do keep in mind that sometimes change can be a good thing too. If they do it right, and it works everywhere, and gives you everything you had before but more.. why hold back? Sure, stay with what works for you for now but don't write off your options forever-more just because they're not there yet. I remember a time when I wouldn't own any phone but a Nokia, and wouldn't use a laptop because they were too underpowered and the screens were terrible. I wouldn't use anything that wasn't x86 and 32bit due to compatibility and stability. I wouldn't use wifi when ethernet was available because it was way slower and unreliable. I wouldn't consider using the internet on a phone due to cost and performance. All that's changed now and I couldn't be happier that it has
To anyone who complains about not using network transparency of X server, I do use all the time and I do need it. And so do thousands of system administrators and power users.
I hope Wayland has some kind of network transparency support. It doesn't have to be X protocol, it could be something new and improved, but there must be a network transparency support.
--Coder
It's native to the windowing system on both the local and remote machine
I thought that modern linux toolkits (GTK and Qt) and windows managers (eg. kwin and metacity) render windows entirely offscreen and just tell X to draw the resulting pixmap. Remote X just sends that bitmap over the network to the remote host. There seem to be little efficiency to be gained by using remote X over VNC in this case
there is absolutely nothing I need to do to either system to pop up a remote display other than insert a 'DISPLAY=remote-host:0.0' in front of the command line.
You need to have a command line to start with. If you're starting from nothing, opening a VNC session is not really harder than opening an SSH session.
I had a sig once. It was lost in the great storm of '09.
In my experience this network transparency of X11 is not really usable over the internet anyways - it's just too slow. I'm using x2go for linux remote desktop. Maybe it's useful over LAN, but is there many people who use it? Do you use it yourself?
Hmm... they believe huge libraries of obsolete and unused code is a good use of limited storage space available?
For one thing, I get just the application window, not a whole extra desktop in a window. I can freely cut and paste between the remote application's UI and a local one.
Good comment. As a sidenote, when I looked at your list, I think right now SSD is a technology that is soonly becoming viable. It has been unreliable and expensive, but that is changing. There are probably other technologies too.
Odd, because I have been managing remote machines for many years (> 10) remotely. Including over 1Mb ADSL, ISDN, even dial up. The point is not that it works as well as a local desktop, but that it works remotely when that is what you need. Our standard demo to make Windows admins jaws drop used to be running something like the Linux version of ET over the network - there you go, full remote 3D display. Trickery required: none.
It seems to me that there's nothing to gain by this. You say that nobody renders like this any more, but I believe you're talking about network transparency. Most detractors of X11 do. But this doesn't seem to be any problem at all.
Look at WoW. Runs faster on Linux than on Windows.
The renderer can't be bad, can it. The mapping of API to game has to be pretty good to manage that, especially since it was written for a different renderer (if, indeed, your proposition about "nobody does it this way" is right) and a different OS.
It seems that still people are looking at X11, seeing network transparency and going "Oh, it HAS to be going slow if it can render over the network!".
If Weyland implements X11 then you're not going to get rid of any of the code that they complain is never going to get used, so no gain there. And if the code is never going to get used, it doesn't slow anything down.
I can cut and paste OK from VNC to other windows. I use a mix of both VNC and remote X depending on where I am working and where I need to connect. On laptops, when moving, remote X is unusable because of the lost connection problem.
http://slashdot.org/comments.pl?sid=3202975&cid=41738551
Everything I write is lies, read between the lines.
VNC is one to one, but with X you can have a pile of windows open from a pile of hosts. It may not be important to many here but it's the entire reason why my workplace has linux on the desktop machines and why the MS windows machines all have X as well.
The funny thing is despite all the "cloud" hype people are forgetting one of the major advantages of being able to use the resources of more than one computer via networking. VNC is the dumb terminal approach, which is fine for many tasks, but it has limitations.
Yep, we will lose this bragging point, but does it really matter? Not for me, Windows admins jaws - what are they good for? Yes, it is great future on paper, it works, but it's slow. Maybe latency is the main factor here - my ping times to my servers rarely less than 400 ms. Anyway, this functionality is needed by low percentage of geeks, I'm not even talking about average users. Managing servers is much more convenient by ssh - I do not need remote desktop for it. When I do need remote desktop - I need it to be responsive and support adjustable levels of lossy compression. I found that x2go is currently the best and easiest solution for my needs.
The article keeps saying Linux.. I'm a bit worried about compatibility with other systems such as BSD. And even Windows, will there be Wayland servers for Windows just like there's a couple of X servers now? Maybe we'll still use X for the networked stuff, and Linux will be just like Windows and Mac, with a compatibility layer to display X11 applications.
No. The reason we get these discussions about Wayland is that it's a framebuffer system and getting the graphics sent over a network is not considered to be important and is not part of the design. All we've got is speculation about screen scraping and exporting the graphics with something else looking at Wayland's framebuffer from the outside. From what I've seen it's a SVGAlib thing designed to run with newer hardware - a good idea but no replacement for X by any stretch because it does other stuff. Six months ago it didn't even have a window manager (so one full screen app at a time) and some guy here was going on about how it was "faster" despite not actually being able to run any sort of test, but I'm not sure where it is now. I wish them luck but the time for claims is when they can be proven.
Actually some of them are (http://en.wikipedia.org/wiki/Keith_Packard)
But yeah, it would be great if they would keep network transparency in Wayland but made a new protocol for it, fast enough for normal usage over the internet, supporting adjustable lossy compression, making coffee, BJ on demand and add a few other useful futures. Just wait, let me raise my army of Atomic Mutant Coders, it will be their second task.
Agreed ; it's slow, even over a LAN. It's fine as a stop-gap measure (like for VirtualBox, where the GUI VM configurator is is much easier to use than the command line), but it chugs.
For serious remote desktop usage on Linux, the only thing I've tried that's actually any good is NX, although that is the only thing I've tried. Microsoft's RDP protocol is excellent. NX is the only thing that comes close to it.
I suppose I might be making some mistake with how I've got it set up, but I've never seen applications run over ssh -X perform any other way but clunky and slow.
X11 is a bottleneck in Unix security. It is not safe to run applications from multiple users on the same desktop if you require isolation between them. The applications can listen in on the keystrokes and (I think) read the content of the other applications' windows. Desktop OSes are in some sense trying to catch up to mobile OSes when it comes to application-level sandboxing, so this isolation an important feature.
It would be extremely exciting if Wayland had a sane isolation model, either based on SELinux or on standard user accounts. I wouldn't have to look at things like Qubes OS (which admittedly is much more secure than my Fedora will ever be). I did RTFA and browsed some of the links, and there is no mention of it, so I'm not holding my breath.
Warning, sliders spotted! In WHICH universe is xterm fast? Try "time cat " on xterm and something sane like Konsole and you'll see for yourself that xterm is multiple times slower.
send screaming back to xterm for one reason or another.
No you don't. Didn't you read all the posts up there? Xterm uses "legacy" protocols which are full of "cruft" and "slow" and "legacy". And noone uses "legacy" stuff anymore. So clearly you don't use xterm because of legacy.
And if you do (and you don't because noone uses "legacy" API calls any more) then you're in the rare 1% and you can go jump because "normal users" (whoever the hell they are) don't need "legacy" things. Oh and those legacy cruft ridden bits of code which you don't use and are paged out onto disk are actually taking up positively kilobytes of RAM which make my Sun 3/60 really REALLY slow.
$ ps -ef | grep xterm | wc -l
54
SJW n. One who posts facts.
That's effectively what's happening with Android, but I think it's a bit of a backward step since it throws away most of the applications that are the reason I use linux in the first place, at least for the moment. Also Win2k/XP and OS X showed that people really want to run a server OS on their computers anyway instead of a cut down piece of crap like Win95/ME.
For all the crowing about network transparency, it is not the killer feature to most. In terms of network management MS's RDP easily wins out. Reason is that it works well and supports low bandwidth situations very well since it isn't just a pixel-based protocol. However for ease of use and cross platform goodness, nothing beats VNC. That's what our Linux guru uses, in the situations where a command line isn't enough, not remote X.
Frankly, given the difference between local interconnects and network speed, I don't think network transparency is the kind of thing to worry about. Rather, make a good local display system, and then add the ability to serve it out over the network through whatever means are appropriate, much like Windows does with RDP. The focus should be on good local display, as that is really what matters.
X11 is a dinosaur and a millstone around the neck of Linux. Huge chunks of it are completely obsolete and what remains is just a bottleneck and extra context switches to drag down performance. The sooner it is dumped the better for Linux. People who really want remote desktops can run X11 as a client over Wayland though eventually Wayland will get a remote protocol which will do away with that requirement too.
Is it possible to test it right now? If so, what is the easiest way? I would try it on VM.
Like Martin, I'm worried about Client Side Decorations:
http://blog.martin-graesslin.com/blog/2010/05/open-letter-the-issues-with-client-side-window-decorations/
I'm worried as I'm not sure they've done anything to address this. I'm certain I will not be touching it unless the window manager can be set to be the one drawing them.
The entire point is that it doesn't. Pure 1980s single user without a network mentality. The bits you've seen about networking here are speculation about adding an external screen scraper in response to people complaining about the single user without a network mentality. It's not an answer, it's just fobbing off the question with a hope that somebody else will do something about it.
It's not a speedboat, it's a harbour wall with graffiti on it. They are hoping that somebody else will take a photo of the graffiti, put it on their speedboat, and then deliver it somewhere else. However nobody has the camera or the speedboat yet. It may not be that hard to do, I don't know, but it's a non-answer because nobody has done that for Wayland yet or even started trying.
It doesn't stop X11 from running as a client over Wayland for those who need a remote app and working in practically the same way as it does now. But it does mean the vast majority of people do not need to run it and do not incur the performance penalties from doing so.
So exactly why does Wayland preclude running remotely? The answer of course is it doesn't. Nor does it preclude somebody running X11 transparently over Wayland (just like X11 will run over Windows / OS X) if they need it but sparing the rest of us who don't need it.
*sighs*
I hope this doesn't end up like:
* Enlightenment 16 and 17
* Gimp and GEGL
* oss, alsa, jack, esd and pulseaudio
* KDE 3 to KDE 4
* Gnome 2 to Gnome 3
I.e. lots of promise and hype, then a looooong time before the new version becomes "useable", if at all.
I now tend to automatically roll my eyes whenever I hear an opensource community mention "rewrite" or "replacement".
The two main complaints I see discussed here appear to derive from some fundamental misunderstandings about what Wayland is.
Wayland is a Protocol and an optional helper library to implementing that protocol. This protocol says nothing about net work transparency, in both the sense of enabling or prohibiting it. It also says nothing about client decorations. The key points here is not to make a decision for or against any particular technology or methodology and then be stuck with that decision for the next 20 years, like we are with X.
How or if, either of these work is all down to the compositor. The reference compositor 'Weston' does not do network transparency and leaves window decoration to the client or its toolkit. However none of the big desktops, i.e end users, will be using this compositor. For example KDE will continue to use Kwin as their wayland compositor, and KDE have already clearly said that Kwin will be decorating their windows and not the clients!
As to network transparency, all windows are drawn to their own back buffers, and where these buffers will be eventually displayed is also the choice of the compositor, and it might well just decided to send them over a network connection. e.g. like what VNC does.
I think if you base your opinion on what other people say, including me, then maybe you shouldn't comment? All of this is discussed first hand on the Wayland web site and/or mailing list.
Unfortunately since I've posted a bit late, I doubt many will read this...
xpra does what you want.
Is that really X11, the network abstraction layer, though? Or is that another extension written to bypass X and go straight to opengl.
09F91102 no, 455FE104 nope, F190A1E8 uh-uh, 7A5F8A09 that's not it, C87294CE no. Ah! 452F6E403CDF10714E41DFAA257D313F.
I'm using eclipse running on another host in the LAN via X11 every day. It works fine. There's no appreciable difference to running it locally. Granted, this is on a fast LAN that's somewhat tailored to do this well.
Switch back to Slashdot's D1 system.
I.e. lots of promise and hype, then a looooong time before the new version becomes "useable", if at all. I now tend to automatically roll my eyes whenever I hear an opensource community mention "rewrite" or "replacement".
Well yes. I think the GEGL thing might have been worth it, but time will tell. PulseAudio seems to be stable now, bringing features that even FreeBSD had years ago...
Anyhow I agree. Basically, these things happen when the older generation goes away or fets bored. It's definitely a case of the latter. I've read things by Keith Packard who clearly knows a lot about X11, but he should be really ashamed of himself about the pure FUD he's been spewing about X11 and Wayland.
I also simply cannot understand how someone who spent so long on X11 could even consider design decisions like client side window decorations. It is utterly baffling to me, and seems to be supported by statetments that are outright unsupportable never mind FUDdy.
Basically as far as I can tell, they've got bored of hacking X11 and want to level the ground and start from scratch, like so many programmers want to do. That's great for them and huge fun, I'm sure, but their position allows them to cause serious damage along the way.
SJW n. One who posts facts.
I use it all the time over LAN and yes even the internet. I don't find it slow. But then i am not trying to play quake or anything like that.
If information wants to be free, why does my internet connection cost so much?
Not to mention that in new "cleaner" code, they have to go and rediscover all the edge cases again. Which of course means the code starts clean and then quickly gets messy, prompting calls to clean up the new code by rewriting everything again.
Wash, rinse, repeat.
PulseAudio seems to be stable now
What that really means is "The ALSA drivers and distributions' default configurations seem to be stable now", much thanks to the hard work of the PulseAudio developers. With good drivers and properly configured, PulseAudio has been stable for a long time.
My other account has a 3-digit UID.
We ran X on terminals far weaker than the average cell phone in the 90s.
Give me Classic Slashdot or give me death!
RDP can share a single application, rdesktop supports application sharing mode. You can run a single Outlook window over RDP on you linux box, works like a charm.
X11 does facilitate the use of, say, NX and VNC though to mitigate the problem. Does Wayland? If so then bring it on I say
Wayland doesn't mitigate anything. It solves the problem by completely eliminating the feature.
Give me Classic Slashdot or give me death!
Can I have a remote Wayland session over ssh, like I do with X? No? Then no, it's not going to replace X in an industrial-grade environment.
Finding God in a Dog
I love the concept of network transparency, but in many cases VNC-style screen scraping seems faster. I've exported the display of apps from our Linux farm to my desktop. However it is painfully slow. VNC on the other hand operates with barely any noticeable lag. Exporting the display of a Linux app from one machine in the farm to another seems to work fine though.
It seems that network transparency in X requires a very high bandwidth and very low latency network connection, where VNC does better with less bandwidth.
I hope they maintain network transparency in Wayland. But I also hope they can improve it so I don't need VNC anymore.
No, that's what the developers say all over the page but it doesn't answer the problem at all. Once applications are targeted for Wayland rather than targeted for X your X daemon inside Wayland is useless.
It's not that the legacy is crap.
It is just that the modern world looks different than it did all those years ago.
The choice for the number of bits an in IPv4 address wasn't wrong at the time, but we need mode bits if we want to Internet to keep growing.
It's the same with the GPUs of today, the best way to talk to them is not the way X11 works.
Compositing seems to be what the creators of Wayland want to do.
Sure you can do compositing with X11, but it isn't ideal. But compositing is what most systems already do. Compiz and KWin, GNOME3 and Unity all are doing compositing I believe.
I think i read newer versions of Windows (since Vista ?) and Mac OS X also do compositing.
If compositing is all you ever do, you don't need everything else from X11. Especially if that abstraction/code is in the "hotpath", the code paths which are the most performance critical.
New things are always on the horizon
I use it over a LAN. I have a thin client in my garage, on my workbench. It's a nice little low power, 12V fanless PC with a very bare bones Debian. I turn it on and a moment later I have a graphical login to my main desktop. Why do the Wayland developers want to take this away from me?
Also, I have used X over the internet in the past. It's a bit too slow IMHO when tunneled through SSH which is probably the most common way to secure it. I used to run openvpn though and it worked just fine through that. I only stopped using it because my router with the built in open vpn server died and I have not gotten around to replacing it. I intend to bring that back one of these days.
I don't even do that. I have a dedicated X terminal in which I have modified the rc scripts so that I go right to a remote X login screen. All I do is turn it on and in a moment I am connected to my remote X desktop.
But they are not doing it right, they are not giving us everything we had before. They are taking away one of the key features that made *nix plus X more useful than other alternatives.
I have read the FAQ several times. Concerning network transparency it is nothing but doubletalk. Their answer is to run X on top of Wayland? Great, so one can run X applications remotely. If new applications and new versions of existing applications are written for Wayland rather than X then that is not even remotely useful. Even if only certain applications which might not run well remotely are targeted towards Wayland and everything else stays X forever.. that is not good. One person's opinion of what is usable and what isn't varies from another. LAN speeds vary, etc... It's better if anyone who wants to can try to run any application remotely and make up their own mind about usability.
VNC and NX? First, unless those projects are modified to run Wayland Apps then that still only gets you X apps which Wayland aims to eventually deprecate. Even if they were modified to run Wayland apps... this is not the same. It is a window acting as a view into another computer with all remote windows locked inside of it, not a local window to a process on a remote computer.
I thought people needed memory management, window management, font management, text rendering capability, graphics drawing, backing store, OpenGL, composited desktops, the ability to sync sound and video, stuff like that. You must mean that network transparency is the one feature that sets X apart.
In fact it's the only part of X that's still relevant. KMS, DRI, cairo, pango, GTK or QT, Gnome or KDE libs etc have taken over almost every part of X. Wayland simply acknowledges this and aims for simplicity by allowing those other parts of a modern system to do their jobs and doing little else. Networked display will show up at some point via some method that's better than VNC too.
You're not thinking of the Berlin Consortium or something are you? I think they changed their name to something catchier later, but it escapes me
That was Fresco. It died sometime in 2004.
No, Wayland is throwing out ALL X11 compatibility
As long as there's a widely available, free X11 server that sits on top of Wayland, compatibility isn't thrown out. And seeing as X.Org X11 and Wayland share a lot of developers, I fully anticipate that such a server will be maintained. From the FAQ: "we now run [X11 applications] under a root-less X server that is itself a client of the Wayland server. [...] With Wayland we can move the X server and all its legacy technology to a optional code path." Do you fear that this code path will bit-rot due to lack of test coverage as GUI toolkits are ported to Wayland and desktop Linux distributions make the switch?
Oh and those legacy cruft ridden bits of code which you don't use and are paged out onto disk are actually taking up positively kilobytes of RAM which make my Sun 3/60 really REALLY slow.
[Is this sarcasm? Or am I even slower than I think I am this morning?] Well, there's your problem! Geez, Sun video has *always* sucked (I once owned a U30). When I was first playing around with Linux on an i486 peecee, then was introduced to a current Sun workstation, I couldn't believe how pathetic it was. Max eight colors, or something?!? Are you kidding me?!? The crappy video artifacts brought out by Linux & *BSD on Sun hardware was pathetic.
I must admit, I'm a bit confused reading this thread at the moment (it's early for me; I was up late). However, who could possibly care that a terminal is slow? I've never seen my CPU spike trying to dump chars to a screen or scroll its contents. The stuff I run from that CLI spikes the CPU; not *rxvt, not xterm, ...
For me, whether I use *rxvt or xterm depends on what I'm running in it, and how that looks in *rxvt vs. xterm, and no, I don't kanything, and gterm is just insulting (yeah, prejudiced wrt both, sorry).
More Scotch ..., er coffee.
"Tongue tied and twisted, just an Earth bound misfit
For serious remote desktop usage on Linux, the only thing I've tried that's actually any good is NX, although that is the only thing I've tried. Microsoft's RDP protocol is excellent. NX is the only thing that comes close to it.
Perhaps this project may interest you. I've tried all of the janky, half-assed remote desktop solutions that Linux has to offer (x2go, NX, X11 forwarding), & this one seems to work the best for me.
I can't for the life of me imagine that it's actually efficient to have a bunch of fancy animations and shadows and transparency and 3D effects
These effects help the user see how the various things displayed on the screen relate to one another. Shadows, for example, indicate to the user that a particular element is on top of another in the stack, which is why Mac OS has been doing some sort of shadow (even if faked through the shape of the window frame) since 1984. Animations of window A going into or out of icon B, such as a folder window coming out of an opened folder icon or a window minimizing down to a taskbar icon, indicate that the object represented by B is the parent to the object represented by A.
Remote display on Macs suck and is no replacement for X.
You should actually bother to use the things that you seek to copy.
A Pirate and a Puritan look the same on a balance sheet.
It's not that they are replacing the X protocol for networks. They are just trying to ignore it. Some people buy into the idea that you can "fix this later" but you really can't It took Microsoft about 20 years to do that.
MacOS just ignores the problem entirely.
It would be one thing if Wayland was addressing that use case rather than just trying to ignore it.
A Pirate and a Puritan look the same on a balance sheet.
I now tend to automatically roll my eyes whenever I hear an opensource community mention "rewrite" or "replacement".
They mean it doesn't look enough like Win* or Macs. Just be thankful you don't have to use it (as opposed to when the same's done on Win* & Macs). "apt-get install fluxbox ..."
"Tongue tied and twisted, just an Earth bound misfit
Wayland's 'market' will be adjusted to its capabilities
sooner than vice versa. Today's paradigm.
In truth, "Not yet ready" will be its middle name.
any device *should* be able to handle it
How can you tell that someone has never developed for cellphones/*pad? Guess...
How can you tell if someone's a Java developer? Guess...
"Tongue tied and twisted, just an Earth bound misfit
I really doubt that network transparency is what is slowing down your $200 video card. X applications have had channels for communicating with local X servers without going over the network or even TCP/IP for a very long time. Your $200 video card is slowed down because the company that made it only implemented core features in the X driver. They are probably even less interested in developing full drivers for yet another system.
Also, I am not an 'admin guy'. I do use network transparency though.. a lot. I don't really use 3d acceleration, at least not for more than eye candy that isn't really very important to me. I don't want your favorite feature to go away though! Why do you want to take away mine?
Why would anyone want to talk to FreeBSD users?
“Common sense is not so common.” — Voltaire
The only part of X11 that people need is network transparent remote display, but that's the one part that the Wayland developers are absolutely hell bent on removing.
They aren't removing anything... they are building something new from the ground up. In doing so, they are choosing not to build the whole castle in one go, but to focus their technical expertise on building a low-level protocol/library, on top of which other developers will inevitably keep building and add support for remote rendering.
Wayland isn't removing your network transparency, it's delegating that responsibility to other yet-to-be-written software. This would be a concern if all distros were moving to Wayland next month, but they aren't. Wayland is a long-term strategy.
-1, Too Many Layers Of Abstraction
And how exactly does remote X with a modern UI toolkit differ from VNC?
Holy crap; you really don't get "client server".
I want "my program" to be running on the honking biggest and fastest box to which I have access. I want that process to display its results on the dinky, pathetic little thing in front of me that I have to use to tell it to do that. I don't want that dinky, pathetic, under-powered and poorly implemented piece of crap in my hand to have to do any more than it absolutely has to, because it sucks at even that!
Play around with "ssh -X" for a few days and get back to us. VNC, Jeebus.
"Tongue tied and twisted, just an Earth bound misfit
The Wayland team is working with KDE and Gnome who are going to implement something like RDB for their respective GUIs. That's going to be how networking is going to be handled at the GUI level.
I would be quite happy with a means to re-connect X applications, unfortunately rather than doing something useful like that (and protocol improvements for a faster display), Wayland wants to throw the baby out with the bathwater.
Couldn't you just use Outlook's web interface instead? Otherwise you could use RDP or Citrix or any number of other methods for remote access to a Win server that has Outlook on it.
~Petaris "The world is open. Are you?"
The important thing about X is very precisely the fact that it's a client-server protocol with network transparency, so the client app doesn't need to be on the same machine as the display server. When I was a sysadmin supporting a bunch of machines, the ability to have lots of windows on my screen that were running clients on lots of different hosts was really valuable, and now that I'm running VMware, I'd be a lot happier if I could do the same thing instead of running a lot of ssh logins and the VMware console widget. It's a bit less critical than it used to be, because lots of the "remote desktop" functionality is built into browsers now, but the tradeoff is big ugly browsers that crash a lot, burn gigabytes of RAM, and aren't very fast.
There are other windowing protocols that also do this (for instance, NeWS, which renders things in Postscript, so what you see is not only what you get, but it's also what you want, and it supports downloading functions to the server so you can avoid round-trips during mouse motion, etc.), but the network transparency is the big win. Running NeWS on an 8MB Sun-3/50 was quite nice, even though these days I've got wristwatches with more horsepower.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
So just this morning I had to reboot my Win7 machine, because I had an RDP session that wouldn't die. Was it because I hibernated the machine to bring it in to the office, disconnecting the VPN in the process? I don't know, but the "Are you sure you really want to close the window, we'll save the session for you, is that ok?" dialog box wouldn't go away and wanted to be on top.
Win7 really is a lot better than XP, and XP was a lot better than 95/SE/ME were, but I still have to reboot it a couple of times a month. And stuff doesn't die for no good reason anywhere near as often as it used to (except Firefox, but our work IT department insists on using older versions with long-term support :-) (And at least Firefox saves its sessions relatively cleanly when it dies, unlike IE8.) And since I'm sometimes using an external monitor as well as the laptop screen, when I'm just on the laptop, sometimes there will be windows I can't open because they're on the Other Screen. If I were using X, these things wouldn't be problems.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
My normal work environment is a Win7 laptop, either on a 10 Mbps LAN at work or VPN from home, with Windows Remote Desktop Protocol connection through a firewall to a Win2008 server (currently running on VMware ESXi), which runs browsers and putty SSH connections to a variety of other machines in my lab. For the things it does, it mostly works ok. Cut&paste works pretty well, remote disk mounting is possible though I don't use it much, web pages with no sound and not too much movement render ok. Watching Youtube with sound on fails badly - way too jerky and painfully out of sync. Youtube with the sound off redraws the pictures better, but especially full-screen you can see the redraws happening on the LAN, and worse remotely. Traffic webcams do fine, network performance graphers do fine. On a LAN, it's better than running Motif on a 386/33 was (ok, it's better than that, because I've got a 1920x1080 32-bit screen instead of a 640x480 8-bit screen... but considering that I've got a few thousand times more horsepower now, it ought to be enough better to be extremely transparent.)
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
So needed some good mod points for this.
The number of times I've watched my own projects go from "beautiful and clean" to "crufty hack jobs" all because of handling corner-cases is probably well... the same as the number of projects I have :(
So if I want to have my desktop connected to a few dozen target machines using RDP, how do you suggest I do it? Keep a few dozen RDP login windows active? Clunky! At that point I might as well just run a few dozen ssh sessions, or stick to browser-based management applications. RDP is fine if I want to talk to one machine at a time (and in fact my Windows desktop usually has an RDP session to a Windows server on my lab network, which has a few dozen ssh and browser sessions open running management applications, because they're mostly on target machines that don't run X.)
(And if all your foes are spelling or grammar Nazis, that means you only make enemies of intelligent people, which _is_ a step up from the kinds of enemies lots of people have, I suppose....)
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
You've got it backwards - if I wanted Linux to be purely a server OS, I'd be fine with administering it over ssh with an occasional https light-GUI tool. With Linux on a desktop running an X window server, I can also run graphical applications on other not-purely-server Linux machines. Also, as VMware and its competitors in the dedicated and cloud businesses are proliferating, more and more things are becoming servers (even my laptop, which often has multiple VMware clients running.)
It would be nice if X had evolved a bit more than it has, but simply running it on machines that are thousands of times more powerful than the Sun 3/50 and 386/33 I first used for X windows in the 1980s really is a good start.
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
You may look at:
http://code.google.com/p/partiwm/wiki/xpra
Everything I write is lies, read between the lines.
I'm not against change, sound drivers suck and I'm use pulseaudio now, and I'm mostly happy with the change. CUPS was a great change. I switched my workstation to systemd this year. I'd been waiting a decade for something that was a step forwards from SysV on that. I use dbus for IPC when I can.
X11 isn't perfect, but network transparency is just not a feature I would give up. It is something I use and I won't be without it. Presumably all the new mac-clone interfaces will switch to whatever this new thing is, and the traditional GUIs will insist on staying with X11.
Try doing that same workload with X11 and see how it compares.
I run: Windows, OS X, Linux, FreeBSD. Just because you have a hammer, doesn't mean everything is a nail.
BSD currently has no plans to support Wayland, and usually takes a while before catching up (unless it's a standard like IPv6). That said, the guys who do PC-BSD could strive to make Wayland at least an option on PC-BSD 10. Other BSD server based platforms, like FreeBSD, OpenBSD may continue to rely on either X or the plain CLI, but PC-BSD is one BSD platform that could make good use of Wayland.
RDP is good, I agree - I hate to admit it - but there are also limitations (cost and setup..)
VNC is just plain outdated (1980s technology)
NX is now closed source and the last open source version is unmaintained
Xpra is the new kid on the block (the xpra.org version) - from your comments, it sounds like you should try it
On *nix, only the last 2 support seamless mode. (RDP could do it but is too much work to setup)
Disclaimer: xpra.org maintainer
TODO: 753) write sig.
True. I was using KDE3, and then KDE4 came out ... with all the nicest usability features removed. "Uh, I don't think so Tim." and became a Gnome2/Compiz user. Then they "modernised" Gnome to Gnome 3. "Aaaaagghhhh". :D
Didn't update my computer for a year, which is ages for me.
Thankfully KDE has listened enough to users to re-implement some of my favourite features (desktop folder view specifically) and I'm now back on KDE. Just wish panel icons were dragable by default, rather than having to enter panel settings first, and that I could drop and drag from menu favourites directly into the panel.
Have tried some lighter desktops but I do admit I like my wibbly windows too much.
xpra.org may be worth a try. Although it is not a "remote desktop" solution but a seamless remote application solution.
TODO: 753) write sig.
It is certainly possible, but easy it is not.
The latest instructions on how to get seamless RDP to work with Linux are well out of date and never worked reliably or properly for me...
TODO: 753) write sig.
Is this sarcasm?
Certainly.
My point was that the Sun 3/60 debuted in (what?) 1985. Back then people accused X11 of being slow and bloated. Well al those so-called "legacy" parts have stayed the same or even shrunk since then. If it was slow and bloated in 1985 and hasn't changed much, then it's tiny and ligntning fast by today's standards.
The point about xterm was that many people claimed that the old parts of X11 are "legacy" and slow and bloated (still, after nearly 30 years) and noone uses them anyway so they should just go, in defiance of all the happy and productive xterm users out there like me.
Emacs (eight megs and constantly swapping) used to be a hog and now it's considered a lightweight editor. X11 somehow is still considered slow and bloated.
X11: Unchanged yet still mysteriously slow and bloated since 1985.
SJW n. One who posts facts.
Guten morgen. Buenos dias.
They mean it doesn't look enough like Win* or Macs.
True. I was using KDE3, and then KDE4 came out ...
Every morning it seems, I stumble out of bed hoping to read something like, "You ASSHOLE!!! !@#$%^^&* ..."
That tends to wake me up and get me thinking again. Instead, what do I get? "I love KDE!" I hate KDE. Would somebody please shoot me, or at least beat me to death with a stick?
Not your fault. I'm a difficult case. There's six !@#$%^& inches of snow on the gawd damned ground, damnit.
I really ought to take another look at KDE, considering it's been about ten years since I last did, but !@#$, it looked awful then. I think the prettiest X ever looked was DECWindow(s?), and I even liked Motif. Then I'm forced to endure KDE? $PUKE.
[I'll be ok. Go hug your wife/mom/kids.]
"Tongue tied and twisted, just an Earth bound misfit
Is this sarcasm?
Certainly.
Whew. :-|
Emacs (eight megs and constantly swapping) ...
It's just a full featured (okay, bloated!) OS. And you can edit text files (and read News, ... :-) with it.
-rw-r--r-- 1 root root 2301536 2012-09-23 07:40 vmlinuz-2.6.32-5-686
(0) kiak /home/keeling_ ls -AlF /usr/bin/vim.basic /usr/bin/vim.basic*
-rwxr-xr-x 1 root root 1510796 Jul 11 2010
The Linux kernel is 2.3Mb. Modern day vim (basic) is 1.5 Mb.
X11: Unchanged yet still mysteriously slow and bloated since 1985.
Morons. :-P I blew close to C$100.00 for a forty Mb disk in the early '90s.
[Ob: GTFOML.]
"Tongue tied and twisted, just an Earth bound misfit
Not really. It's a much more painful to try to remote a browser vi X11 on a LAN than it is to VNC/RDP into windows. The main issue is that while bandwidth has improved greatly from the early days of X11 there's only so much you can do about latency, and nearly no apps are written to handle that network latency well. It's actually a much better idea to let them do all their tiny operations on a local framebuffer and stream the whole image as a big blob of pixels. Back when X11 was designed, it was basically unthinkable to throw pixel streams across the network, because bandwidth was such a scarse resource (even on a LAN, 10 MB shared collision domain? Ouch.). When your design constraints change so drastically it's only natural that the optimal solution is no longer the same.
The thing is, with a modern design around the capabilites of video cards you have a much cleaner way of grabbing the individual window frabebuffers (textures? Not sure what they're called when compositing) and streaming them as an image remotely. You can couple that with some hints ("Scroll down 50 pixels") and get a very efficient network protocol - that the applications can safely ignore. As far as they're concerned, everything is happining locally, and they don't care that instead of getting scaled and wobbly they're being encoded and streamed across a lan. As long as you forward the same input events, what's the difference?
The thing about X11 is that you can take a butt-ugly-looking Xt-based toolkit and sling C code from waaaaay back in the day and there's a good chance that it still runs.
Main difference between the BSD license and the GPL license: one is from California and the other is from Massachusetts
At least for me, xterm with server-side fonts is fast; ctrl-middle-click and ticking "Enable Jump Scroll" helps too.
I speak England very best