Canonical Continues Working On XMir Performance

[DDF420]

Mir is the worst thing to happen to Linux, ever.

How miserably inadequate your life must be. In your countless wasted hours, boring us to death everyday playing "i hate everything canonical" in every post, like the attention whore you are then feel the need to throw in another opportunist obtuse one liner announcing your apathetic roll in the whole saga. !!

[scionicspectre]

A lot of people love Linux but don't care much for what Canonical's doing with it, much like how many people don't really like what Android has built above the very fundamental bits. They identify with a different kind of Linux that respects some of the values they hold dear. Maybe Dee is just upset because Mir seems like a large political snafu, which has the potential to undermine a lot of what's good about Linux by pressuring developers out of the traditional ecosystem.

I think it's fine either way, but that's only because I've seen that openness and collaboration always find a way to assert themselves. So the Linux we love will survive, no doubt, although it may not be as widespread as Ubuntu. Much like how MeeGo is just barely scraping by on mobile while Android drowns out its influence.

Time will tell, but I don't think people are pathetic or live inadequate lives just because they're concerned this will make things difficult outside of the Ubuntu community.

[dee.]

How miserably inadequate your life must be. In your countless wasted hours, boring us to death everyday playing "i hate everything canonical" in every post, like the attention whore you are then feel the need to throw in another opportunist obtuse one liner announcing your apathetic roll in the whole saga. !!

Oh honey, are you proposing to me? This is so soon!

[dee.]

A lot of people love Linux but don't care much for what Canonical's doing with it, much like how many people don't really like what Android has built above the very fundamental bits. They identify with a different kind of Linux that respects some of the values they hold dear. Maybe Dee is just upset because Mir seems like a large political snafu, which has the potential to undermine a lot of what's good about Linux by pressuring developers out of the traditional ecosystem.

No, I don't really care about "political snafus". Politics has really nothing to do with it. Mir is horrible because it's horrible to the desktop Linux ecosystem. To understand, you need to know a bit of history, you need to see the whole ugly story.

See, we now have a situation where we have one display server - Xorg. Everything right now runs and is compatible with Xorg, no matter what distro, what DE, you can run the same software in Gentoo/KDE as in Ubuntu/Unity as in Fedora/GNOME. The so-called "fragmentation" isn't even really fragmentation, since all the desktops play nicely enough together that any software written for one DE can still be ran in another (ie. it doesn't really matter if your DE uses GTK or Qt, you can still run both GTK or Qt based apps).

Now, Xorg has many flaws, many of which are inherent to its design and cannot thus feasibly be fixed in any meaningful way. So many Xorg developers figured, we need a new display server. That's where Wayland came to the picture. These people, with years of experience working with the Linux graphics stack, have been spending a huge amount of time and effort writing code, and thanks to them, we now have a display server that is modern, fast, efficient, fixes all the flaws in Xorg, provides tear-free desktop compositing with perfect frames, and even provides backwards compatibility via XWayland - something they're well-equipped to do, being mostly Xorg-developers themselves.

So, everyone pretty much saw the advantages of Wayland, and that we should be moving towards it - even Canonical. Shuttleworth himself said "Canonical fully supports Wayland, Ubuntu will use Wayland in the future". Every other major distro, every desktop environment and toolkit, Mesa, the rest of the graphics stack, threw their support towards Wayland - so we had a clear, easy migration path away from X: Distros could provide Wayland with XWayland for legacy X apps, while most apps would be getting Wayland support via modern toolkits such as GTK3, Qt5 or EFL. Legacy apps that still use older toolkits would use Xwayland as rootless application-specific X servers and everything would be fine. To top that, we were getting mobile Linux systems (Sailfish, Tizen) using Wayland, Wayland was being adopted in embedded and IVI systems - we were finally getting a display server that would unify the entire Linux ecosystem, not just on the desktop but everywhere. You'd have to be a fool not to see the benefits of this.

Of course, as so often happens, someone dropped the ball - and this was Canonical. What they did was abominable. Despite for years publicly backing Wayland, even offering it in repos, promising people that Ubuntu would run on Wayland, they suddenly turned coat and informed they had been working on their own display server in secret. Not only that, but they released a bunch of FUD and lies about Wayland, some of which is still floating around thanks to people who are so fanatical in their following of their favorite corporation that they don't take the time (or just don't care) to look up the facts.

What they did was, they backstabbed the entire Linux ecosystem. Instead of backing the modern next-gen display server they had promised, they made their own in-house solution, which is basically just a rip-off of Wayland and other people's code. XMir is a plain xerox copy of XWayland. Libhybris was first developed for Wayland, by Jolla employees, to allow Wayland to use Android drivers, then Canonical ripped it off and even tried to present it as their own work. Because of this, many people are STILL touting Mir's ability to run on Android drivers as if it's an advantage against Wayland - when the whole library was developed for Wayland in the first place! And Shuttleworth keeps pissing everyone off, telling everyone how he's certain Mir will be better than Wayland, more popular, that everyone will eventually support Mir instead of Wayland... and other Canonical employees have to pretty much just shuffle their feet, try to do damage control and wish their boss would just shut up.

So now, instead of that clean, easy migration path to Wayland, we have this competitor - Mir, and no one really likes it - the toolkits, other DE's, etc. - none of them really want anything to do with it. Which is understandable, as they already have a much better solution which they have been investing time on - Wayland, and Wayland is developed for the needs of the entire Linux ecosystem in mind, whereas Mir is only developed for the needs of Unity with no care for other distros - Canonical has explicitly stated so. So Canonical is probably going to have to implement toolkit support etc. upstream, since none of the toolkits want to do it, and even Mesa patches will have to be maintained upstream. All this is making Ubuntu more and more separate from the rest of the Linux ecosystem - requiring more and more in-house patches to function. This is a bad thing. People have been saying "sure, Ubuntu does stupid shit, but at least it brings people to Linux" - this won't be the case, it will do no good if Ubuntu becomes so different and separate from other Linux systems that it's basically a different OS.

The worst impact will be on software that doesn't for some reason use toolkits, but needs to talk to the display server directly - think: games, proprietary software ported from other OS's - those will have to decide if they support Mir or Wayland. They need to choose between the distro with most users, or all the other distros (which still combined have more users) - a lose-lose scenario. So the lowest common denominator will be X, since both will support X via XWayland or XMir. Without Mir, they'd have a no-brainer decision, to support Wayland. So think how this hurts the chances of Linux becoming a real, serious gaming platform.

So any way you slice it, Mir is a bad idea and it's harmful to Linux. Forget politics, forget hurt feelings and shit-slinging, it's bad because it causes REAL fragmentation and it's technically inferior, a copycat solution that only works for one distro and hurts everyone else. Jolla's first phone runs Wayland, so there's no reason Canonical couldn't have done the same. They needed control? That's fine, they could have written their own Wayland compositor, and keep the control of the code 100% to themselves, choose whatever license they wanted - there was simply NO good reason to create Mir. It was a stupid decision, and now Canonical is too stubborn to back out of it, because they think PR and marketing-speak will fix everything.

[dee.]

It's good Shuttleworth finally learned some business 101. Praising Wayland while developing an alternative secretly was smart from him (assuming that is what happened and not just some coincidence). Everybody thought he was cool, they turned their back and he stabbed them 20 times. Dude, it's business, this is how things run my friend. It's a competition and Canonical wants to win. They're finally learning the art of getting into power. If winning means crushing a few people and companies so be it. Nobody cares. Canonical might be on the road to something big. Finally. Think about it, the ones that get into power like Microsoft, Apple, Oracle, aren't nice. They don't mind killing and eating you. Finally Shuttleworth learns from the best. I'm starting to think Ubuntu might have a chance to play with the real big boys after so many years of obscurity.

Of course all this is assuming that what .dee says is true and not just some conspiracy theory.

Yeah, and next, Shuttleworth needs to assert his dominance of the marketplace by raping a couple of babies and kicking some puppies. Because, you know, that's just smart business. I know so because I've watched two seasons of The Apprentice and Donald Trump says so.

Bo$$, I'm pretty sure I'm speaking for all the other Ubuntu fans when I say: Please, stop helping. You're making even the Ubuntu fanboys look bad.

[Maxjen]

No, I don't really care about "political snafus". Politics has really nothing to do with it. Mir is horrible because it's horrible to the desktop Linux ecosystem. To understand, you need to know a bit of history, you need to see the whole ugly story.

See, we now have a situation where we have one display server - Xorg. Everything right now runs and is compatible with Xorg, no matter what distro, what DE, you can run the same software in Gentoo/KDE as in Ubuntu/Unity as in Fedora/GNOME. The so-called "fragmentation" isn't even really fragmentation, since all the desktops play nicely enough together that any software written for one DE can still be ran in another (ie. it doesn't really matter if your DE uses GTK or Qt, you can still run both GTK or Qt based apps).

Now, Xorg has many flaws, many of which are inherent to its design and cannot thus feasibly be fixed in any meaningful way. So many Xorg developers figured, we need a new display server. That's where Wayland came to the picture. These people, with years of experience working with the Linux graphics stack, have been spending a huge amount of time and effort writing code, and thanks to them, we now have a display server that is modern, fast, efficient, fixes all the flaws in Xorg, provides tear-free desktop compositing with perfect frames, and even provides backwards compatibility via XWayland - something they're well-equipped to do, being mostly Xorg-developers themselves.

So, everyone pretty much saw the advantages of Wayland, and that we should be moving towards it - even Canonical. Shuttleworth himself said "Canonical fully supports Wayland, Ubuntu will use Wayland in the future". Every other major distro, every desktop environment and toolkit, Mesa, the rest of the graphics stack, threw their support towards Wayland - so we had a clear, easy migration path away from X: Distros could provide Wayland with XWayland for legacy X apps, while most apps would be getting Wayland support via modern toolkits such as GTK3, Qt5 or EFL. Legacy apps that still use older toolkits would use Xwayland as rootless application-specific X servers and everything would be fine. To top that, we were getting mobile Linux systems (Sailfish, Tizen) using Wayland, Wayland was being adopted in embedded and IVI systems - we were finally getting a display server that would unify the entire Linux ecosystem, not just on the desktop but everywhere. You'd have to be a fool not to see the benefits of this.

Of course, as so often happens, someone dropped the ball - and this was Canonical. What they did was abominable. Despite for years publicly backing Wayland, even offering it in repos, promising people that Ubuntu would run on Wayland, they suddenly turned coat and informed they had been working on their own display server in secret. Not only that, but they released a bunch of FUD and lies about Wayland, some of which is still floating around thanks to people who are so fanatical in their following of their favorite corporation that they don't take the time (or just don't care) to look up the facts.

What they did was, they backstabbed the entire Linux ecosystem. Instead of backing the modern next-gen display server they had promised, they made their own in-house solution, which is basically just a rip-off of Wayland and other people's code. XMir is a plain xerox copy of XWayland. Libhybris was first developed for Wayland, by Jolla employees, to allow Wayland to use Android drivers, then Canonical ripped it off and even tried to present it as their own work. Because of this, many people are STILL touting Mir's ability to run on Android drivers as if it's an advantage against Wayland - when the whole library was developed for Wayland in the first place! And Shuttleworth keeps pissing everyone off, telling everyone how he's certain Mir will be better than Wayland, more popular, that everyone will eventually support Mir instead of Wayland... and other Canonical employees have to pretty much just shuffle their feet, try to do damage control and wish their boss would just shut up.

So now, instead of that clean, easy migration path to Wayland, we have this competitor - Mir, and no one really likes it - the toolkits, other DE's, etc. - none of them really want anything to do with it. Which is understandable, as they already have a much better solution which they have been investing time on - Wayland, and Wayland is developed for the needs of the entire Linux ecosystem in mind, whereas Mir is only developed for the needs of Unity with no care for other distros - Canonical has explicitly stated so. So Canonical is probably going to have to implement toolkit support etc. upstream, since none of the toolkits want to do it, and even Mesa patches will have to be maintained upstream. All this is making Ubuntu more and more separate from the rest of the Linux ecosystem - requiring more and more in-house patches to function. This is a bad thing. People have been saying "sure, Ubuntu does stupid shit, but at least it brings people to Linux" - this won't be the case, it will do no good if Ubuntu becomes so different and separate from other Linux systems that it's basically a different OS.

The worst impact will be on software that doesn't for some reason use toolkits, but needs to talk to the display server directly - think: games, proprietary software ported from other OS's - those will have to decide if they support Mir or Wayland. They need to choose between the distro with most users, or all the other distros (which still combined have more users) - a lose-lose scenario. So the lowest common denominator will be X, since both will support X via XWayland or XMir. Without Mir, they'd have a no-brainer decision, to support Wayland. So think how this hurts the chances of Linux becoming a real, serious gaming platform.

So any way you slice it, Mir is a bad idea and it's harmful to Linux. Forget politics, forget hurt feelings and shit-slinging, it's bad because it causes REAL fragmentation and it's technically inferior, a copycat solution that only works for one distro and hurts everyone else. Jolla's first phone runs Wayland, so there's no reason Canonical couldn't have done the same. They needed control? That's fine, they could have written their own Wayland compositor, and keep the control of the code 100% to themselves, choose whatever license they wanted - there was simply NO good reason to create Mir. It was a stupid decision, and now Canonical is too stubborn to back out of it, because they think PR and marketing-speak will fix everything.

Could'nt agree more. I wish everybody would read this.

No need my friend. We are telepathically linked. He sees what I see.

Maybe BO$$ is Mark Shuttleworths secret phoronix account. I feel like if he were to express his real feelings instead of acting like Steve Jobs all the time he would sound just like BO$$.

[dee.]

It's also not just a bad move for the rest of the ecosystem - it's also a bad move for Canonical itself. The Ubuntu fans who keep on defending Mir and blaming everyone who criticizes it of "hating Canonical" should take a step back and look at things rationally. People who love Ubuntu have the most reason to be against Mir.

I speak for myself only, and I'm sure there are some who really, really dislike Canonical and Ubuntu for whatever reasons, but I don't. I like Ubuntu and wish only good things for it. I really would like to see Canonical succeed in creating a mainstream Linux OS, bringing desktop Linux to the mainstream. That'd be great. Only, what they're doing with Mir is not helping. They're actually shooting themselves in the foot. Even the dash adware isn't that big of a blunder, they could just turn it off by default and it wouldn't matter. Even if they keep it on by default (as they seem to be doing) it still wouldn't hurt them as much as this Mir idiocy.

It'd be different if they developed their entire OS in-house, then it might in some way make sense to say "to hell with the competition" and cheer them on to crush every other OS (an unrealistic goal, but it would at least make sense). However, when you consider the fact that Ubuntu is 90% dependent on the Linux ecosystem - 90% dependent on other people's code, it doesn't make sense at all. A Linux distro cannot survive by crushing their competition, because the whole open-source ecosystem works in a way that the competition is necessary - the collaboration and sharing of effort is needed to utilize scarce resources. Say Ubuntu becomes the only Linux distro - what then? What do they base their distro on if Debian fails? Where do they get the bulk of their code? Since they're unable to even create their own display server from scratch - how would they be capable of developing and maintaining their entire software stack?

In the Linux world, alienating the entire community and ecosystem with poor choices and bad behaviour isn't a viable strategy. Microsoft could pull that shit off back in the 90s when they had the entire market by the balls. Even they can't get away with it anymore, so what reason is there to believe Canonical - a small startup with a tiny marketshare - could? There's simply no viable path for Canonical to even get to that position, because the more they hurt the ecosystem that supports them, the more they're shooting themselves in the foot.

[synaptix]

Giant wall of text, but a great read.

I do remember Shuttleworth saying Ubuntu was fully supporting Wayland and would eventually be moving to Wayland, which was great considering that means my current distro (Xubuntu) would most likely follow suite.

I thought Wayland? Cool, something new that would be better than Xorg (Xorg is still great, but does have it's problems). I was happy to see the most populated distro in Linux supporting this project.

But then Canonical vomited out Mir. Ubuntu will be using Mir from now on soon enough, but I'm not so sure about it's derivatives such as Kubuntu, Xubuntu and Lubuntu and of course Linix Mint as well. Think I read somewhere that Xubuntu would be staying on Xorg and wouldn't go Mir.

[Vim_User]

It's good Shuttleworth finally learned some business 101. Praising Wayland while developing an alternative secretly was smart from him (assuming that is what happened and not just some coincidence). Everybody thought he was cool, they turned their back and he stabbed them 20 times. Dude, it's business, this is how things run my friend. It's a competition and Canonical wants to win. They're finally learning the art of getting into power. If winning means crushing a few people and companies so be it. Nobody cares. Canonical might be on the road to something big. Finally. Think about it, the ones that get into power like Microsoft, Apple, Oracle, aren't nice. They don't mind killing and eating you. Finally Shuttleworth learns from the best. I'm starting to think Ubuntu might have a chance to play with the real big boys after so many years of obscurity.

Actually, it is a really dumb move to stab those in the back that you are relying on. Without the community Canonical has nothing they can base their OS on, their only product that they can sell. Now they stabbed the community in the back. Really smart move, kill those whose work you need for your product. Shuttleworth will learn the hard way that something that works in the proprietary OS market will not necessarily work in the FOSS market.

[dee.]

I do remember Shuttleworth saying Ubuntu was fully supporting Wayland and would eventually be moving to Wayland, which was great considering that means my current distro (Xubuntu) would most likely follow suite.

I thought Wayland? Cool, something new that would be better than Xorg (Xorg is still great, but does have it's problems). I was happy to see the most populated distro in Linux supporting this project.

But then Canonical vomited out Mir. Ubuntu will be using Mir from now on soon enough, but I'm not so sure about it's derivatives such as Kubuntu, Xubuntu and Lubuntu and of course Linix Mint as well. Think I read somewhere that Xubuntu would be staying on Xorg and wouldn't go Mir.

None of the derivatives will use Mir. Everyone is either going to stay on X for now, or go for Wayland. Xfce has Wayland support on the roadmap I think, and Xubuntu will likely migrate once Xfce finishes porting the DE to GTK3. LXDE is migrating to Qt5, and probably also going to use Wayland in the future. KDE has stated very clearly that they want nothing to do with Mir and will be supporting Wayland.

Canonical suggested to the derivatives, "just run over XMir", and to top that off, they're now running the main distro that way - a stupid, really stupid kludge: XWayland and XMir are meant to be used for applications, not desktops or window managers - now, there will basically be an extra layer that provides no benefit whatsoever. Heck, the current Ubuntu 13.10 runs on XMir and it actually shows up two mouse cursors on the screen - and, get this: the Canonical devs are claiming "it's not a bug, it's a feature" because "it lets people know they're running on top of two display servers"! No, really. I wish I was making this up.

So Lubuntu and Kubuntu at least have both - sensibly - said "no thanks", they'll have nothing to do with Canonical's kludge solution, they won't be running entire desktops over XMir, and are instead staying on X and later moving on to Wayland. I haven't heard explicit announcements from Xubuntu, but I don't think they'll be running over XMir, because it makes no sense for them to do so. Mint hasn't made an explicit announcement, but Clem (the main dev) has stated that they don't care about Mir and it doesn't affect Mint's plans. So it's probably Wayland for Mint.