Ubuntu alone is not a significant user base, if it alienates the rest of the Linux users (who will be using Wayland). With the current move they'll probably start losing users. In general, in order for Linux gaming efforts to gain traction, there should be unified driver base. If that won't happen - Canonical would achieve a real sabotage for the whole industry.
Originally Posted by e8hffff
Agreed - I was very happy with their earlier versions, particularly with the netbook editions that tweaked the UI for smaller screens. Unfortunately, it's been a steady decline for the past couple of years, resulting in me moving back to Fedora in 2011...
Originally Posted by bobwya
Why can't they compete by adopting Wayland and creating features for it, implement their own Ubuntu compositor instead of Weston along with new features? If every corporation that pumped propane invented their own gasket for the propane tanks we use instead of agreeing on a -standard- for gaskets, then I'd have to go to one place only instead of a dozen in a radius of 5 miles to refill my propane tank... Your point, while valid, has its trade-offs which you omitted to mention.
Originally Posted by BO$$
Do you work for Canonical by any chance?
I doubt mir will win, since IMO it is pretty much impossible for 5 or so developers with no prior experience in that field to write something that actually works...
Originally Posted by BO$$
Also Wayland is backed by Intel which is about 1000 times bigger than Canonical and actually makes a profit
No fragmentation? Wayland developers already have everything being ported to Wayland. The toolkit development has gotten further.
Originally Posted by BO$$
No other distro is going to use Mir, and they are going to be using Wayland, but now thanks to Canonical, it looks like we're also going to have all these Mir apps floating around, that won't run on any other distro at all, unless Weston or the like adds another plugin that does what xwayland does, only for Mir. This would probably require another process, the rootless MIR server.
Most of Wayland's development cycle is design. If Canonical goes as fast as they go, we're going to be stuck with a far inferior solution. They have so many complex things that they still have marked as "TODO", and they still aren't sure what they are going to do from what I can tell from RAOF on the Wayland IRC. Remember, they've been working on this thing in secret for nine months.
The false assumption here I think from some people is that the majority of Linux users run Ubuntu. While I have no firm data to back this up (and neither do the people claiming otherwise), I would wager that while Ubuntu probably has a plurality of the user share the total number of users of other distributions probably out numbers it by a fair margin. And as other people have posted, when it comes to driver support Nvidia and AMD are primarily targeting 3D workstations, which probably use a solution such as RHEL or SUSE Enterprise. So I hardly see Ubuntu forcing their solution on users killing off Wayland.
Also, keep in mind that not all Ubuntu users use Unity, and since they basically said Unity and Mir are going to work in lock-step...
Last edited by Hamish Wilson; 03-05-2013 at 11:13 PM.
I don't know about the future of linux display server, but for sure ubuntu has lost one user: me.
Generally speaking, I agree about the positive effect of the competition, such as web browser competition, which can stimulate each players to do their best.
But for the display server, there is a big problem in the middle: the GPU drivers.
It is no more "I do my stuff, you do your stuff and the user will decrete the winner", I worry that it will be more similar to "I do my stuff, you do your stuff and GPU maker will ignore both".
I feel like the ubuntu devs started to work on wayland, looking to the philosophy behind it, looking how wayland arrange all the puzzle's pieces and, at some point, they say "hey, why not continue working in house? I understand the general idea, I understand how the story will end, we don't need to follow them no more!".
Looking to the Mir spec, many things sound similar to wayland's design. I think that for wayland, the big effort it is been analysis and develop a very smart archetecture, a modern project. Mir did not spend that time and energy because the just had copy the other work, cannibalizing the others work, but I say this: ubuntu devs don't have the know how for walking on their legs.
Look to the TODO list:
- Modality state into surface model
- Window decoration
- Window resizing
- Semi-maximised states
- Proximity regions
- Overlay scrollbar
- Side stage
- Input transformation
- Workspaces in surface model
- Non-homogeneous DPI displays
- Window deformation
- Inter-session effects
- Ready notification
- Boot / shutdown
- Mir on Mir
- Greeter on Mir
- Client transition animations
- Form-factor information status update
- Remote access
- Shell input control - pointer barriers, acceleration etc
- GTK+ support
- XUL support
- Libreoffice support
- Update control center
- Input methods, accessibility
Ready for October 2012? Yeah, dreaming is free.
For that date, we will have wayland in 2.0 version, xwayland in xorg stable release and MIR in the middle of nothing.
The opinions of Martin Gräßlin (developer of Kwin) summarize perfectly my actual feeling.
Starting from now, ubuntu users = ubuntu users - 1