Linked by David Adams on Tue 31st Jan 2012 23:50 UTC, submitted by Michael
Graphics, User Interfaces This weekend at FOSDEM 2012 what Kristian Hogsberg is expected to say in Brussels will surprise many of you: Wayland 1.0 is gearing up for release as their first stable release. Wayland [a new X server for Linux] is supposed to be ready to take on the Linux desktop world.
Order by: Score:
it's not a X server.
by hussam on Wed 1st Feb 2012 04:18 UTC
hussam
Member since:
2006-08-17

it's not a X server.

Reply Score: 17

RE: it's not a X server.
by Laurence on Wed 1st Feb 2012 09:13 UTC in reply to "it's not a X server."
Laurence Member since:
2007-03-26

it's not a X server.

True, but I was reading last night that Wayland can run Xorg as non-root for compatibility.

Different thing entirely I know, but I found that interesting at the time so thought I'd share

Edited 2012-02-01 09:13 UTC

Reply Score: 4

RE[2]: it's not a X server.
by fithisux on Wed 1st Feb 2012 10:08 UTC in reply to "RE: it's not a X server."
fithisux Member since:
2006-01-22

It seems that xorg is pushed completely to the user space like Xming/Xphoto/Xquartz.

Reply Score: 3

RE[3]: it's not a X server.
by pgeorgi on Wed 1st Feb 2012 11:24 UTC in reply to "RE[2]: it's not a X server."
pgeorgi Member since:
2010-02-18

Or XGGI, in 1998 (http://marc.info/?l=ggi-develop&m=90986709728710&w=2)

The cabal finally caught up \o/

Reply Score: 2

krh
by diegoviola on Wed 1st Feb 2012 05:15 UTC
diegoviola
Member since:
2006-08-15

if you're reading this:

Thank you so much for Wayland, you rock. ;)

Reply Score: 7

Comment by lucas_maximus
by lucas_maximus on Wed 1st Feb 2012 08:30 UTC
lucas_maximus
Member since:
2009-08-18

Being Released under the MIT license.

Reply Score: 0

Comment by orestes
by orestes on Wed 1st Feb 2012 12:20 UTC
orestes
Member since:
2005-07-06

Cool. Now, we just need to have desktop environments and WMs to finish being built for it so normal users can take it out for a spin.

I know I'm looking forward to seeing how fast the Fedora team can get it included as a workable experimental option

Reply Score: 3

RE: Comment by orestes
by shmerl on Wed 1st Feb 2012 16:38 UTC in reply to "Comment by orestes"
shmerl Member since:
2010-06-08

It's also a question of drivers. Nvidia didn't publish any so far for example.

Reply Score: 2

RE[2]: Comment by orestes
by orestes on Wed 1st Feb 2012 16:46 UTC in reply to "RE: Comment by orestes"
orestes Member since:
2005-07-06

It'd almost be better off if they didn't. Binary blobs are a pain in the ass for anyone remotely interested in the bleeding edge. This would be the best shot at a clean start with only open drivers we've had in forever.

Reply Score: 3

RE[3]: Comment by orestes
by shmerl on Wed 1st Feb 2012 20:17 UTC in reply to "RE[2]: Comment by orestes"
shmerl Member since:
2010-06-08

Ideally yes. But so far we still don't have open accelerated Nvidia drivers even for X11. Reverse engineering didn't get further than Nouveau so far. Why do you think Wayland drivers will succeed in that better without Nvidia's input? And they so far didn't show any desire to open specs.

Edited 2012-02-01 20:18 UTC

Reply Score: 5

RE[3]: Comment by orestes
by lemur2 on Wed 1st Feb 2012 22:13 UTC in reply to "RE[2]: Comment by orestes"
lemur2 Member since:
2007-02-17

Bill Shooter of Bul:

I've found the best results come with the Open source Gallium drivers with video cards 2-3 years old. I have two separate screens, and maximising works as expected on each in all desktop environments and window managers I've tried.

According to Phoronix, the Open Source Nvidia drivers are improving at a pretty decent rate as well. Soon I imagine they'll be good enough for normal desktop use for a majority of Nvidia cards.


orestes
It'd almost be better off if they didn't. Binary blobs are a pain in the ass for anyone remotely interested in the bleeding edge. This would be the best shot at a clean start with only open drivers we've had in forever.


tuma324
The nvidia blob is a half-assed port of the Windows drivers to Linux, that's why it sucks so badly for things like 2D, etc.

Nouveau already does many things better than the blob, like 2D acceleration, KMS, and other things.

I have no doubt Nouveau will surpass the blob in other areas in the future.


Just on the continuously-improving capabilities of the open source drivers, it has been recently announced that R600 Gallium3D Can Now Do OpenGL 3.0, GLSL 1.30.

http://www.phoronix.com/scan.php?page=news_item&px=MTA1MDM

"The R600g driver now joins Intel Sandy Bridge, Gallium3D Softpipe, and Nouveau NVC0-Fermi as being the open-source Mesa / Gallium3D drivers that are currently capable of doing OpenGL 3.0 / GLSL 1.30."

Since the closed-source binary blob drivers won't be able to support Wayland, it is good to finally see the open source drivers start to catch up in capabilities.

Reply Score: 2

RE[4]: Comment by orestes
by twitterfire on Fri 3rd Feb 2012 14:25 UTC in reply to "RE[3]: Comment by orestes"
twitterfire Member since:
2008-09-11


Just on the continuously-improving capabilities of the open source drivers, it has been recently announced that R600 Gallium3D Can Now Do OpenGL 3.0, GLSL 1.30.


Yes, three years and a half after Nvidia's drivers did that.

Beside that, open source drivers have a puny 3D acceleration.

Reply Score: 3

RE[3]: Comment by orestes
by twitterfire on Fri 3rd Feb 2012 14:06 UTC in reply to "RE[2]: Comment by orestes"
twitterfire Member since:
2008-09-11

It'd almost be better off if they didn't. Binary blobs are a pain in the ass for anyone remotely interested in the bleeding edge. This would be the best shot at a clean start with only open drivers we've had in forever.


Why would binary drivers be a "pain in the ass"? At least they are usable and you get decent performance from them.

Open source drivers for AMD and Nvidia cards suck badly.

Reply Score: 2

RE[4]: Comment by orestes
by orestes on Fri 3rd Feb 2012 16:01 UTC in reply to "RE[3]: Comment by orestes"
orestes Member since:
2005-07-06

Haven't had a single issue with the open drivers for my 4800 series Radeon card since... right around Fedora 13 if memory serves.

As for the binary blobs, where do I start? They lag behind on supporting new X releases to a degree that's utterly inexcusable at times, need to be rebuilt every time you upgrade a kernel, still fail to support KMS, and they complicate system troubleshooting.

Oh yeah, and there's also the little matter of not being portable to other platforms and being handcuffed to the vendor as far as when support will end for the card.

Edited 2012-02-03 16:01 UTC

Reply Score: 2

Hope they can deliver
by Invincible Cow on Wed 1st Feb 2012 12:25 UTC
Invincible Cow
Member since:
2006-06-24

I love this. I just hope it can deliver on its goals, which is a tear-free graphics system.

It wouldn't be fun to go through a rough phase of changes across the graphics stack, only to find nothing has changed for the better.

Also, I hope wayland will solve the problem of smooth transitions between bootsplash, login screen and desktop.

Reply Score: 3

RE: Hope they can deliver
by Gullible Jones on Wed 1st Feb 2012 13:26 UTC in reply to "Hope they can deliver"
Gullible Jones Member since:
2006-05-23

The hell with tear-free... I'll settle for a graphics stack that runs decently fast on a typical Pentium II box, and doesn't have a fit when I try to enable 2D acceleration.

When I was an acne-pocked teenager, I thought Linux graphics ought to be perfect. Now I think they ought to be stable and snappy, in that order of importance. If looking more like Windows 2000 than Windows 7 is the price to pay... So be it.

P.S. Windows 2000 was awesome. I have seen it crash once from an incompatible third-party driver, and maybe twice from hardware problems, and that's it. It can run fast on a Pentium II with 128 MB of RAM, and much faster on a higher-end machine. Really a shame it's no longer supported.

Reply Score: 5

RE[2]: Hope they can deliver
by ggeldenhuys on Wed 1st Feb 2012 13:46 UTC in reply to "RE: Hope they can deliver"
ggeldenhuys Member since:
2006-11-13

Win2000.....faster on a higher-end machine. Really a shame it's no longer supported.

+1
The best Windows OS to date! I still run Win2000 on most of my Windows development machines - rock solid babies with light resource usage (unlike todays OSes).

Reply Score: 2

RE[3]: Hope they can deliver
by ebasconp on Wed 1st Feb 2012 22:29 UTC in reply to "RE[2]: Hope they can deliver"
ebasconp Member since:
2006-05-09

Can it still run, say, VisualStudio 2010 or 2008?

Reply Score: 2

RE: Hope they can deliver
by ggeldenhuys on Wed 1st Feb 2012 13:48 UTC in reply to "Hope they can deliver"
ggeldenhuys Member since:
2006-11-13

I just hope it can deliver on its goals, which is a tear-free graphics system.

I hope so too! The current Linux graphics system and windowing environment is a total joke. I really hope with the fresh start, that Wayland can deliver.

Reply Score: 1

RE[2]: Hope they can deliver
by Gullible Jones on Wed 1st Feb 2012 14:46 UTC in reply to "RE: Hope they can deliver"
Gullible Jones Member since:
2006-05-23

Unfortunately it seems to me that the graphics situation has gotten even worse lately. Before KMS, you could disable hardware acceleration if you had problems, and things would generally work (albeit with somewhat poorer performance). Now you're stuck with the fbdev driver as a fallback, and IMO it's a pretty trashy fallback - you can't easily change the screen resolution with it, it won't work at all with some configurations (e.g. PPC Macs that require an unaccelerated framebuffer), and IIRC it has serious problems with multiple monitors.

Granted there is a "Shadow" option for Intel now, that's supposed to work like the old "NoAccel" and "ShadowFB"... Last time I tried it, it made Xorg segfault on start.

Compare FreeBSD 9, which still uses XAA/EXA without KMS. Start the X server... Oops, there's a hardware acceleration bug, and everything is slow with a tendency to freeze. Disable hardware acceleration with "NoAccel" and everything works! Damn shame that the Xfce developers decided to ditch HAL support, and automount/power management doesn't work on the BSDs any more.

Reply Score: 3

RE[3]: Hope they can deliver
by No it isnt on Wed 1st Feb 2012 15:24 UTC in reply to "RE[2]: Hope they can deliver"
No it isnt Member since:
2005-11-14

The graphics situation might be getting worse for ancient hardware, but at some point you can't expect anything else. And although the situation might have become worse for the two of you who run Linux on PPC Macs, it's much better now for those who use current, supported hardware than it was, say, eight years ago when you Mac still wasn't obsolete.

Making use of actual, working hardware acceleration is more important than the ability to turn it off. Which is actually something Linux does better than FreeBSD 9.0 and ATI.

Reply Score: 4

RE[4]: Hope they can deliver
by Gullible Jones on Wed 1st Feb 2012 16:39 UTC in reply to "RE[3]: Hope they can deliver"
Gullible Jones Member since:
2006-05-23


Making use of actual, working hardware acceleration is more important than the ability to turn it off. Which is actually something Linux does better than FreeBSD 9.0 and ATI.


Absolutely not IMO. It's important for the OS to make proper use of the hardware. It's more important for the OS to remain usable when it can't make proper use of the hardware, which is going to happen sooner or later, given Linux's lack of market share.

In short, better to work with somewhat reduced functionality than to not work at all.

(On a related note, the not-very-functional fbdev driver is far faster on my Intel chipset laptop than the current intel driver, which is fully hardware accelerated. How's that for making full use of the hardware?)

Reply Score: 3

RE[5]: Hope they can deliver
by No it isnt on Wed 1st Feb 2012 17:27 UTC in reply to "RE[4]: Hope they can deliver"
No it isnt Member since:
2005-11-14

You're wrong. The FBdev driver seems faster because it doesn't do nearly as much as the Intel driver. Try playing a film in fullscreen, for instance. If you want functionality, buy supported hardware. Don't complain that the software won't willingly cripple itself to work poorly with unsupported hardware.

Reply Score: 1

RE[6]: Hope they can deliver
by cmchittom on Wed 1st Feb 2012 17:56 UTC in reply to "RE[5]: Hope they can deliver"
cmchittom Member since:
2011-03-18

Don't complain that the software won't willingly cripple itself to work poorly with unsupported hardware.


One person's "willingly crippled" is another's "failing gracefully."

Reply Score: 3

RE[7]: Hope they can deliver
by No it isnt on Wed 1st Feb 2012 19:29 UTC in reply to "RE[6]: Hope they can deliver"
No it isnt Member since:
2005-11-14

No, they are two entirely different things.

Reply Score: 2

RE[6]: Hope they can deliver
by Gullible Jones on Wed 1st Feb 2012 20:29 UTC in reply to "RE[5]: Hope they can deliver"
Gullible Jones Member since:
2006-05-23

BTW I just love it when people say "LOLOL buy better hardware LOL." There are good reasons why some people won't be willing to do this. Leaving those people behind won't make Linux any more popular.

Reply Score: 4

RE[3]: Hope they can deliver
by ggeldenhuys on Wed 1st Feb 2012 20:08 UTC in reply to "RE[2]: Hope they can deliver"
ggeldenhuys Member since:
2006-11-13

Now you're stuck with the fbdev driver as a fallback, and IMO it's a pretty trashy fallback

Correct me if I am wrong, but there is also VESA drivers you can use. I know I used that the other day as a fallback, but can't remember what Linux distro I used (Slackware, Ubuntu or OpenSUSE - I think it was Slackware).

At least the open source AMD/ATI and nVidia drivers are getting better and better performance wise. But I do understand the issue with bugs. The open source ATI driver has had a painting bug since after Ubuntu 8.04 (I'm a GUI toolkit developer), and it still exists in Ubuntu 11.10 - no movement on the bug report either. :-(

Reply Score: 2

RE[4]: Hope they can deliver
by Gullible Jones on Wed 1st Feb 2012 20:13 UTC in reply to "RE[3]: Hope they can deliver"
Gullible Jones Member since:
2006-05-23

Nope, you can't use the vesa driver. vesa + KMS = crash city. fbdev does work with KMS, but it has the usual shortfalls of fbdev.

(You could use the vesa driver and the nomodeset kernel option, but then you can't get the correct screen resolution on laptops. Well, not without some ugly hacks. Anyway the point of a fallback option is to be there in case something goes wrong, not to be a geeky thing that takes lots of effort to get working.)

Reply Score: 2

RE[2]: Hope they can deliver
by bouhko on Wed 1st Feb 2012 15:18 UTC in reply to "RE: Hope they can deliver"
bouhko Member since:
2010-06-24

The drivers are at fault too. Take dual screen for example. For some reason, both Nvidia and ATI choosed to implement their own dual screen stack in their drivers.
As a result, the ATI implementation sucks badly, because what it basically does is simulate a big desktop, so when you maximize a window, it gets maximized across screens.
Enabling xinerama on the ATI drivers solve this, but then I get tearing everywhere and moving a window put my quad-core and my radeon 6950 on its knees...

When I owned a Nvidia card, I had almost the same problem.

So I hope Wayland can offer a better API so driver vendors just implement what's needed and we get a uniform API that apps can use for multiple screens and other things.

Reply Score: 2

Bill Shooter of Bul Member since:
2006-07-14

I've found the best results come with the Open source Gallium drivers with video cards 2-3 years old. I have two separate screens, and maximizing works as expected on each in all desktop environments and window managers I've tried.

According to Phoronix, the Open Source Nvidia drivers are improving at a pretty decent rate as well. Soon I imagine they'll be good enough for normal desktop use for a majority of Nvidia cards.

Reply Score: 5

RE[3]: Hope they can deliver
by phoenix on Wed 1st Feb 2012 17:03 UTC in reply to "RE[2]: Hope they can deliver"
phoenix Member since:
2005-07-11

The radeon driver has supported Xrandr for multi-monitor support for several years now.

Prior to that, though, I found the MergedFB support in the radeon driver to be far superior to nVidia's TwinView.

I was a staunch ATI fan for many many years, going from radeon to firegl to radeon again, always with dual monitors, never with any issues with "maximise covers both monitors".

It was only when the district standardised on nVidia for student desktops that I switched away from ATI. And it's taken a couple of years to get a working xrandr setup using the nouveau driver. Something I had from the get-go with radeon/mergedfb.

Reply Score: 2

RE[3]: Hope they can deliver
by cyrilleberger on Wed 1st Feb 2012 19:48 UTC in reply to "RE[2]: Hope they can deliver"
cyrilleberger Member since:
2006-02-01

As a result, the ATI implementation sucks badly, because what it basically does is simulate a big desktop, so when you maximize a window, it gets maximized across screens.


Curiously, the ATI driver does not maximize the window across screens for me. Maybe because you can select how you want your multidisplay to be handled by the driver.

Reply Score: 2

RE[3]: Hope they can deliver
by Yagami on Wed 1st Feb 2012 19:51 UTC in reply to "RE[2]: Hope they can deliver"
Yagami Member since:
2006-07-15


Enabling xinerama on the ATI drivers solve this, but then I get tearing everywhere and moving a window put my quad-core and my radeon 6950 on its knees...

When I owned a Nvidia card, I had almost the same problem.



No it doesnt. I have dual screens ( 1920x1200 ) and it runs as smooth as single.

I have an ati 4890.

Granted, see my previous posts ( on kde 4.8 ) , its basicly :

Funtoo -> flys

openSuse -> crawls

Reply Score: 1

RE[3]: Hope they can deliver
by ggeldenhuys on Wed 1st Feb 2012 20:20 UTC in reply to "RE[2]: Hope they can deliver"
ggeldenhuys Member since:
2006-11-13

As a result, the ATI implementation sucks badly, because what it basically does is simulate a big desktop, so when you maximize a window, it gets maximized across screens.
Enabling xinerama on the ATI drivers solve this,

That's weird. I use an ATI 45xx (I think) card at work. Ubuntu 8.04 and Ubunutu 10.04 using the open source ATI drivers works beautifully in a dual monitor setup. And no, when I maximise a window, it maximise only on the current monitor, not over both displays. I also don't use Xinerama - that just sucks big time.

I also tried the proprietary ATI driver under Ubuntu 10.04 - yes I have better performance, but the dual monitor setup was more of a pain to setup than with the open source driver (monitor 1 & 2 was swapped around). In both case, I never use Xinerama.

The only odd thing I did find (with some applications), is that popup dialogs might appear centred between the two monitors - but again, only with certain software.

My issue with X11 is the disconnected window frame from the application (application can't reliably query the window frame size or width of borders etc), huge amount of lag while resizing (mouse cursor is faster than the window frame can keep up), tearing while dragging a window around etc.

Reply Score: 2

RE[3]: Hope they can deliver
by tuma324 on Wed 1st Feb 2012 20:57 UTC in reply to "RE[2]: Hope they can deliver"
tuma324 Member since:
2010-04-09

The drivers are at fault too. Take dual screen for example. For some reason, both Nvidia and ATI choosed to implement their own dual screen stack in their drivers.
As a result, the ATI implementation sucks badly, because what it basically does is simulate a big desktop, so when you maximize a window, it gets maximized across screens.
Enabling xinerama on the ATI drivers solve this, but then I get tearing everywhere and moving a window put my quad-core and my radeon 6950 on its knees...

When I owned a Nvidia card, I had almost the same problem.

So I hope Wayland can offer a better API so driver vendors just implement what's needed and we get a uniform API that apps can use for multiple screens and other things.


The nvidia blob is a half-assed port of the Windows drivers to Linux, that's why it sucks so badly for things like 2D, etc.

Nouveau already does many things better than the blob, like 2D acceleration, KMS, and other things.

I have no doubt Nouveau will surpass the blob in other areas in the future.

Edited 2012-02-01 21:03 UTC

Reply Score: 2

RE: Hope they can deliver
by renox on Wed 1st Feb 2012 17:29 UTC in reply to "Hope they can deliver"
renox Member since:
2005-07-06

I love this. I just hope it can deliver on its goals, which is a tear-free graphics system.
I'm not sure that it is the goal..
But to answer your question: it depends!
If wayland is used with its default compositor then a client draw all the window including the decoration so it should be tear-free, this has the cost (a high cost IMHO) that the window may become unresponsive if the client is busy..
KDE plans to use server side decoration, so it will provide responsive windows but not necessarily tear-free.

It wouldn't be fun to go through a rough phase of changes across the graphics stack, only to find nothing has changed for the better.

I think that many issues are caused currently by drivers so wayland won't help here, it could even be worse: new code and new way to use existing code --> new bugs (remember PulseAudio?).

Also, I hope wayland will solve the problem of smooth transitions between bootsplash, login screen and desktop.


Uh? With KMS it should already be a solved issue..
But Wayland will indeed "solve" the issue as it needs KMS to work ;-(

Reply Score: 3

RE: Hope they can deliver
by tuma324 on Wed 1st Feb 2012 20:50 UTC in reply to "Hope they can deliver"
tuma324 Member since:
2010-04-09

I love this. I just hope it can deliver on its goals, which is a tear-free graphics system.

It wouldn't be fun to go through a rough phase of changes across the graphics stack, only to find nothing has changed for the better.

Also, I hope wayland will solve the problem of smooth transitions between bootsplash, login screen and desktop.


KMS already solves this, but it will be better with Wayland since Wayland actually requires KMS to work.

You also want to avoid using the crappy blobs like "nvidia" and use the open source drivers that support KMS, like Nouveau, Intel or radeon.

The open source drivers have KMS, the blobs don't.

Edited 2012-02-01 21:06 UTC

Reply Score: 3

Comment by Drunkula
by Drunkula on Wed 1st Feb 2012 14:27 UTC
Drunkula
Member since:
2009-09-03

+1 on the Win2K comments. Shame its no longer "supported."

Reply Score: 1

RE: Comment by Drunkula
by Gullible Jones on Wed 1st Feb 2012 16:49 UTC in reply to "Comment by Drunkula"
Gullible Jones Member since:
2006-05-23

To tell the truth, it's probably still usable as a desktop OS, if you're willing to be cautious about how you use it (which is probably a good rule of thumb for any Windows version). Firefox (and thus Noscript) works, Flash works, most desktop applications work... There are even some security programs (Sandboxie, Outpost Firewall Free, SuRun) that work with it. It's gotten to be a bit of a patchwork OS, but in practice I think a knowledgeable desktop user can avoid most of the security pitfalls. The more major problem would be lack of support for encrypted wifi and such.

OTOH, I would not trust Win2K at all in an office environment, or for anything requiring file or print sharing. YMMV.

Reply Score: 3

RE[2]: Comment by Drunkula
by Lennie on Wed 1st Feb 2012 17:19 UTC in reply to "RE: Comment by Drunkula"
Lennie Member since:
2007-09-22

Well, actually I think Adobe stopped creating (even security) updates for Flash on Windows 2000.

Reply Score: 2

RE[3]: Comment by Drunkula
by Gullible Jones on Wed 1st Feb 2012 17:23 UTC in reply to "RE[2]: Comment by Drunkula"
Gullible Jones Member since:
2006-05-23

Link please? It still says Flash is compatible with Win2k here: http://www.adobe.com/shockwave/download/alternates/

Reply Score: 2

RE[4]: Comment by Drunkula
by Lennie on Wed 1st Feb 2012 17:34 UTC in reply to "RE[3]: Comment by Drunkula"
Lennie Member since:
2007-09-22

I just remember seeing an updater complain that it can't install.

Maybe the machine had a service pack missing ?

It has been a while since I last looked at the problem.

Reply Score: 2

RE[2]: Comment by Drunkula
by zima on Tue 7th Feb 2012 23:59 UTC in reply to "RE: Comment by Drunkula"
zima Member since:
2005-07-06

I still have win2k on one machine (dual Pentium II ...a bit short on RAM; plus IIRC it uses an older multiprocessing hw standard, which can cause problems under XP or non-ancient Linux)

And yeah, it's quite fine when I boot it from time to time. On such hw, the relative snappiness of Opera can be really appreciated (especially older versions, 9.27 most notably ...curiously, with websites dropping IE6 support, such old Opera tends to work better now - focus on standards compliance at work, I guess; except for js, better to disable it completely if only due to slowness), Flash is better to be forgotten (a bit weird / sad - it was "the thing" back then but a different thing: vector animations), and security is not much of an issue behind router.

Edited 2012-02-08 00:09 UTC

Reply Score: 2