Canonical Announces Mir: A New Display Server Not On X11 Or Wayland 354
An anonymous reader writes "On the Ubuntu Wiki is now the Mir specification, which is a next-generation display server not based on X11/X.Org or Wayland. Canonical is rolling their own display server for future releases of Ubuntu for form factors from mobile phones to the desktop. Mir is still in development but is said to support Android graphics drivers, open-source Linux graphics drivers, and they're pressuring hardware vendors with commercial closed-source drivers to support it too. They also said X11 apps will be compatible along with GTK3 and Qt/QML programs. Canonical isn't using X11 or Wayland with their future Unity desktop as they see many shortcomings from these existing and commonly used components."
wayland's flopping, lets try again! (Score:2, Interesting)
Seriously? Can't leave it well enough alone? Can't even focus your energy on one replacement, you want to work on another too?
Canonical swirling down to irrelevance. (Score:4, Interesting)
Unless they can convince the wider Linux community to adopt some of their technologies, Canonical is basically going to end up forking the platform. If that happens, it will be a fairly major step backwards for Linux on the desktop since developers will be on the hook to adjust to supporting not just multiple packaging systems and multiple library versions, but also multiple incompatible core system API's. Essentially Ubuntu will no longer be "Linux" in any way that matters to developers and all the support for Linux out there now will either die or just switch over to being Ubuntu specific and I don't see how that benefits anyone in the community.
Terrible news from the Soviets at Canonical (Score:1, Interesting)
This is a terrible development. The splintering of the Linux desktop into a bunch of incompatable window systems is the last thing we need and which has been prevented for years by the X standard. While the Mir says it will support X applications, the threat comes from the fact that there may also be Mir applications which will not be able to run on other distributions, will not be able to run on X server root displays of other distributions. Another danger is not only Canonical trying to create a fleet of Mir only applications that cannot run on X server, but as well, end up creating a driver mess with drivers that can only run on Mir, or where driver vendors will now be faced with supporting many incompatable driver APIs for all of these windown systems, which will deter hardware vendors from supporting the platform.
The presents of so many incompatable window systems will simply make Linux appear to be a splintered, fractured platform that will appear impossible for hardware vendors to support.
Another problem with this Mir idea is that it takes away the ability of Linux users to continue to use their fine tuned, customized X desktops which so many have invested time in tailoring to their liking, and with their own choice of window manager.
I also find the name to be odd. Do they name it after a soviet space station as an indication that they are planning to take away our rights in a soviet style dictatorship? Canonical has been acting like a soviet style dictatorship that has forced its own obscene agendas on its users for years, including the Unity atrocity which is utterly antagonistic towards users. The Unity environment was designed to force on users some convuluted model of how things should work, as if Canonical has felt it needs to take away users freedom and force on users things that will cause them great pain and discomfort because the pain is good for them, as a price to pay for Canonicals fetish for bizarre and unueable user interfaces. It is sort of like how an interior decorator designs extremely uncomfortable furniture that inflicts misery on the users of that furniture, because to the interior decorator, such misery is a beautiful thing, they think that deprivations and discomfort have an aesthetic value to them.
Re:No, not again (Score:5, Interesting)
I think Shuttleworth has just decided (probably correctly) that he can't make any money on the desktop, but mobile is still a possibility. The Unity interface and now this are an attempt to compete with Android.
I abandoned Ubuntu for my desktop when Unity came, but I think I might actually like it on a tablet or phone. Anyway, I'll try to keep an open mind when the devices actually come out. I hope one of non-Android Linux phone efforts finds a niche, whether it's Ubuntu, Jolla, Tizen, or Firefox OS. If Shuttleworth can pull it off, then more power to him.
Buntu's Track Record. (Score:2, Interesting)
This will go nowhere. Cananonical has "completion" issues. Look at their past track record on linux. The focus on a feature for a release or two and then either declare it done or stop talking about it. They were going to make everything easy, printing, wifi, audio. Pulse Auido is still far from perfect and network manaeger still has issues. Then we have 10 second boot times, better looking that Mac, Desktop notifications, Wayland and 200 million users by October 2013.
Back in October of 2011 I predict the death of Wayland on my blog which I almost never post to. http://elder-geek.blogspot.com/2011/10/ubuntu-is-failure.html [blogspot.com]
Unity is still here, but instead of fixing it for the desktop, more work will go into making it run on other platforms. I love Linux with all of my heart. But Ubuntu is so preditible on how they are going to fail. They never complete anything that they start. Linux will be safe in the long run from the Distro that strives to remove the word "Linux" from their users minds.
Terrible clever idea or just terrible? (Score:5, Interesting)
I really don't have the technical knowledge to praise or damn the idea, but as I understand it, there are some clever moves in this;
It appears that they rip out enough of Android that they can use the Android graphic drivers for Mir, so that every device with android drivers delivers "free" drivers for Mir too. That would give them a huge advantage in the Smartphone and Tablet arena.
QtMir, QtUbuntu, Qt/QML; it looks like Ubuntu dumps Gnome/GTK in favour of Qt5 for core OS (GUI) development. As I see it they will clone KDE/Qt, substituting the KDE parts with QtUbuntu.
Their time line seems very optimistic though.
Re:No, not again (Score:4, Interesting)
I think Shuttleworth has just decided (probably correctly) that he can't make any money on the desktop, but mobile is still a possibility.
It is highly doubtful that he can make any money in the mobile sphere, that is pretty well decided now, too. He probably stood a better chance with the desktop, particularly after Windows 8.
The Unity interface and now this are an attempt to compete with Android.
If the goal was to compete with android, they should have gone KDE. KDE active is a much more attractive development environment and much further along than Ubuntu's mobile offerings, which don't even use the standard Unity interface.
I abandoned Ubuntu for my desktop when Unity came, but I think I might actually like it on a tablet or phone. Anyway, I'll try to keep an open mind when the devices actually come out. I hope one of non-Android Linux phone efforts finds a niche, whether it's Ubuntu, Jolla, Tizen, or Firefox OS. If Shuttleworth can pull it off, then more power to him.
Study after study shows that Unity does not work well on a tablet/touch device. It only looks like it should work, but all of the apps are mouse centric. The problem for Canonical going mobile is that most of the apps in their repositories, which is a large selling point (even if free), won't work on mobile. So from the very start, they will be competing with Apple and Android who have a huge head start and even Microsoft who while a very distant third is lightyears ahead of Canonical.
As I said earlier, they should have gone Plasma Active. If all of the resources that they dumped into Unity and now their mobile offerings had been used to further that project, they would have been to market earlier and had apps ready to deploy. Instead they chose to go their own way, which is their right, but not necessarily the wisest business decision as even Microsoft is late to the game.
Re:wayland's flopping, lets try again! (Score:5, Interesting)
I think we're seeing a natural cycle in the software world. During the 80s there were dozens of architectures, operating systems, languages, etc. and the best (for some definition of best) became dominate and during the 90s consolidated. Now we're in the midst of another explosion in new technology (languages, display servers, processor architectures, perhaps even operating systems) that will eventually lead to reconciliation and consolidation in another five to ten years.
Things like Wayland have to appear, and even fail: their existence allows new ideas to be tested giving us a better idea of where to go from here.
Re:why (Score:4, Interesting)
I understand the desire to replace X.
It's the desire to trash everything and start again, but this time doint it *right*.
Big chunks of X either aren't needed any more or have moved into other locations (mostly the kernel).
Yes and no. Mostly no.
For better or worse, quite a bit of the hardware side has moved into the kernel.
The other bits (old-style graphics and font rendering) is no longer big. It was big in 1987, but by 2013 standards it's a few k, perhaps even a few M of memory. Utterly irrelevant.
The other parts of X work really pretty well.
Sure there are warts. But the better solution is not to nuke it from orbit, it's to come up with protocol fixes to give thigs like persistence and fewer round trips (e.g. like NX). The trouble with nuking things is that all the edge and corner and even marginally non mainsream cases just get thrown away too.
X does a lot of things well, and large parts of the protocol have aged very gracefully. Did you know that copy/paste with advanced (non text) types and drag and drop is all implemented using mechanisms compatible with the original 1987 X protocol?
Oh, and you can pry my server side decorations from my cold, dead hads :)
Also what moron on the X team got rid of the keycombo to nuke server grabs for misbehaving applications? I think the reasoning was that it shouldn't be necessary because that's an application bug and should never happen. No shit it's a bug, sherlock! Now these monkeys are trying to give us the next great compositor.
Basically they have no respect for the user.
Re:Canonical swirling down to irrelevance. (Score:4, Interesting)
On the desktop I've switched from Ubuntu to Mint, but on the server I've changed from Debian to Ubuntu LTS. For my use case having more up-to-date software is more important than utter stability and the outside chance of major-vendor support for programs that I'm not going to run anyway.
In fact, my roll-my-own-distro choice is now Ubuntu Server, which is far less likely to spontaneously break than the current favorite riceboy distro, Arch.
Re:Good luck with that! (Score:2, Interesting)
Actually the client-server thing isn't that stupid. The really nice thing about it is that it now also fits the difference we have between GPU and CPU memory, something on the display server (in X11 the client) is likely to be in GPU memory, and something on the application (in X11 the server) is definately in CPU memory. This means if you optimize for networked X11 by limiting what you copy between client and server, you end up with something also optimized for hardward accelerated rendering (keeping pixmaps in textures and working on them remotely rather than doing local computation). Of course X11 really really needs an update, but it could be better handled by throwing out shit, mandating new shit and calling it X12.