Linked by Thom Holwerda on Fri 8th Mar 2013 16:13 UTC
Ubuntu, Kubuntu, Xubuntu Mark Shuttleworth: "I simply have zero interest in the crowd who wants to be different. Leet. 'Linux is supposed to be hard so it's exclusive' is just the dumbest thing that a smart person could say." He's right. Lots of interesting insights in this blog post - I may not agree with everything Ubuntu does, but at least it's doing something.
Thread beginning with comment 554730
To view parent comment, click here.
To read all comments associated with this story, please click here.
Wayland and Mir
by ndrw on Fri 8th Mar 2013 17:37 UTC in reply to "Comment by Laurence"
ndrw
Member since:
2009-06-30

Looks like Mark has stabbed Wayland project in the back. If they play it out well (i.e. don't bundle it with a ton of Unity crap) it may be _the_ successor of X in Linux.

The reasons are mostly political (technical "issues" pointed by Canonical are a stretch, to say the least). The biggest issue with Wayland is that they just move too slowly and have engaged in ideology (by depending on half-baked in-kernel GPL graphics stack). Anyone who gets rid of this dependency and deploys the stuff first is the winner.

Mir has its problems too - CLA, being late to the market. But it will have the drivers (both Android and proprietary) and will be shipped in millions by a major Linux distribution.

Wayland still has a chance but they:
- have to be more flexible about the graphics stack.
- need a deployment vehicle (like Wayland-on-X) to get Wayland to distributions. That's not unlike how OpenGL compositors were deployed (Xgl).
- have to hurry up.

Reply Parent Score: 3

RE: Wayland and Mir
by ndrw on Fri 8th Mar 2013 17:47 in reply to "Wayland and Mir"
ndrw Member since:
2009-06-30

> technical "issues"

There are actually two technical issues that bother _me_, not sure if Mir is addresses them:
- Client-side window decorations. Window management is a compositor's job.
- Yet another IPC bus. There was no reason not to use DBUS even in 2008.

Reply Parent Score: 2

RE[2]: Wayland and Mir
by ssokolow on Fri 8th Mar 2013 19:46 in reply to "RE: Wayland and Mir"
ssokolow Member since:
2010-01-21

- Client-side window decorations. Window management is a compositor's job.


Could go either way. On the one hand, they were pushing for them but, on the other hand, they did move away from that kind of design with appindicators.

(Though I always switch back to the traditional tray icon or find a new application when faced with appindicators because I insist on left-click being "hide/show main window" and right-click being "show context menu". Last I checked, the libappindicator API simply does not support letting an application register a handler beyond "show context menu".)

Reply Parent Score: 3

RE[2]: Wayland and Mir
by Lunitik on Fri 8th Mar 2013 21:09 in reply to "RE: Wayland and Mir"
Lunitik Member since:
2005-08-07

Window decorations is not a valid complaint, as an example, KWin is implementing Wayland and thus still be doing decorations.

Again, Wayland is a protocol which Weston implements. Weston is only a reference implementation though, Mutter of Gnome is also implementing it as far as I can tell. At the end of the day it is highly unlikely anyone will actually use Weston at all.

This is really not a good situation though, everyone moving away from Xorg will have to implement the Wayland protocol in their code, multiple duplication of effort will be happening. I would be surprised if there is not another project that turns up too since everyone will be writing display managers anyway. As I said elsewhere, I just hope closed drivers start targeting KMS and GEM et al rather than trying to track all of the mess that is going to happen.

For me, the real reason is that Wayland is another IPC, there is no need when DBUS is already widely used. This is the key takeaway from all the docs about Mir, which I think they understood they'd have to write anyway - just calling it Wayland support in Compiz, it would still be a reimplementation of Weston.

Edited 2013-03-08 21:11 UTC

Reply Parent Score: 3

RE: Wayland and Mir
by _txf_ on Fri 8th Mar 2013 19:12 in reply to "Wayland and Mir"
_txf_ Member since:
2008-03-17

Mir has its problems too - CLA, being late to the market. But it will have the drivers (both Android and proprietary) and will be shipped in millions by a major Linux distribution.


You're also forgetting the lack to technical ability at canonical. The fact that they couldn't understand wayland shows that lack of skill. They might be able to get something half working, but I'll bet (like Unity) it will be absolutely plagued by bugs and incompatibilities for some time.

One doesn't simply just produce a quality display manager easily and quickly, see the fact that Androids SurfaceFlinger still isn't anywhere near as good and flexible as other display managers (even after all these years)..

Note that the most experienced people in linux graphics support Wayland. And some of those didn't have very nice things to say about the code quality of Mir.

but we'll just have to wait and see if Canonical can avoid the most likely outcome.

Edited 2013-03-08 19:16 UTC

Reply Parent Score: 4

RE: Wayland and Mir
by Lunitik on Fri 8th Mar 2013 20:02 in reply to "Wayland and Mir"
Lunitik Member since:
2005-08-07

The graphics companies don't need to do anything about Mir or Wayland (actually Weston, Wayland is just a protocol.) What they need to do is write drivers for the current in-kernel stuff. The display system doesn't need to be implementing drivers, it needs to be displaying things on the screen. It is the job of the kernel to interact with hardware.

The problem is Nvidia and ATI just can't be asked to adopt the new stuff, but if they did it wouldn't even matter what display server is used, everyone would benefit. That shouldn't be a concern of anyone but Nvidia and AMD though, they are the ones refusing to adopt the new stack.

Reply Parent Score: 2