Mir: Stop this train before it derails Ubuntu.

[tartalo]

And then you (and other DEs too) would be more than happy to use Unity + Mir if Wayland compatibility would be maintained. Are you really sure because i doubt that!

But you and other DEs would probably still not be using it? It's a bit pointless then to debate strongly over the decisions made isn't it?

You insist, so this seems important for you, but you are mixing different questions so it's a bit difficult to understand what you ask, let me try:

* Would other DEs besides Unity use Unity if Mir was Wayland compatible?

Err, no? At most some parts, like someone using Unity's HUD or something like that, but I've heard that reusing parts of Unity is difficult because of it's intricate design.

* Would other DEs besides Unity use Mir as the default compositor if Mir was Wayland compatible?

Perhaps.

There are DEs that rely on X11 Window Managers that didn't develop themselves, like LXDE using Openbox. They do have plans for future Wayland compatibility so unless Openbox decides to move to Wayland they'll have to use something else, If Mir was Wayland compatible they could have considered it as an option, now I don't see it happening.

The DEs that develop their own X11 Window Managers already had plans to make them Wayland compatible, and that probably wouldn't change in the short term. But in the long term who knows, for example KDE is now recommending Gstreamer, which was a Gnome thing, these kind of reevaluations happen all the time, specially if they clearly benefit the users.

* Would users of DEs other than Unity use Unity if Mir was Wayland compatible?

More likely than with Mir being Wayland incompatible. Most that don't use Unity probably simply prefer other alternatives, but some can't use it because it only works properly in Ubuntu, and porting it will be more difficult after Mir. I don't really know if the people that have been trying to hack Unity in to Arch are doing it because of the challenge or because they actually like Unity as a DE, I also don't know if any people that like Unity have sacrificed it because what they don't like are Mir future plans, but people that liked Ubuntu as a whole did jump ship because of Mir not being Wayland compatible.

* Would users of DEs other than Unity use Mir if Mir was Wayland compatible?

Definitively yes. People make every kind of heterodox combinations: KDE on Openbox, XFCE on Compiz... So, why not?

What's important to understand is that current Mir plans close a lot of doors to Ubuntu users and would complicate developers' life if they chose to acknowledge it's existence. I have used the Word Wide Web to exemplify the importance of commonly agreed standards before and I'll do it again. You probably have heard about XAML and Silverlight, when everyone was discussing what we know today as HTML5 for rich interfaces in the web, Microsoft attempted to push their own thing instead, it was somehow "open" and there was a half working implementation for Linux too, Netflix used it, but it was an attempt to gain control of the web (again) and such things tend to fail in the end. Would I use Internet Explorer if they had backed HTML5 since the beginning, probably not, but that's not what matters, when there's a commonly agreed standard, which implementation I choose to use becomes irrelevant.

[EgoGratis]

You insist, so this seems important for you, but you are mixing different questions so it's a bit difficult to understand what you ask...

OP said he does not use Unity that is why i asked him why does he have strong opinion on something he does not use in the first place. He insist Mir should be Wayland compatible right from the start and i asked what purpose would that have and if he expects other DEs would jump on the ship and use Mir. As long as common drivers are used for all solutions i don't see a problem given the fact there will not be only one display server or only one implementation to support Wayland protocol anymore. Once Wayland matures i don't see why Mir could not gain support for it and while maturing i don't see why Wayland protocol couldn't implement support for something tested and used in Mir.

I think it's not difficult to understand the question and it doesn't need further explanation.

Perhaps.

They can still use Mir (add support) if it suits their purpose. About Canonical being upstream for others and not only Ubuntu perhaps in the future when things mature. Canonical started by positioning itself to serve what upstreams have to offer to end users now this changed a bit and they produce code for Ubuntu purposes and who knows in the future they might become important upstream to serve others with the code too and yes as strong upstream they will/would have stronger control over it. What will/would they do with that control? I guess we will just have to wait and see how will/would they act as strong upstream because we are not there yet.

There are DEs that rely on X11 Window Managers that didn't develop themselves, like LXDE using Openbox. They do have plans for future Wayland compatibility so unless Openbox decides to move to Wayland they'll have to use something else, If Mir was Wayland compatible they could have considered it as an option, now I don't see it happening.

The DEs that develop their own X11 Window Managers already had plans to make them Wayland compatible, and that probably wouldn't change in the short term.

If Mir matures and proofs itself i don't see why support could not be added and the same applies for Wayland that needs maturing too. First somebody has to beat X11/Xorg in meaningful way, adding 10% overhead and not much benefit will not beat X11/Xorg anytime soon for example. Who knows what will be the pace of Wayland/Mir based solution in achieving this goal but i doubt anybody wants to wait for this to happen for more than let's say 2 years in the future.

[buzzingrobot]

... current Mir plans close a lot of doors to Ubuntu users and would complicate developers' life if they chose to acknowledge it's existence.

As a user, I have to admit I am not especially concerned if the lives of developers are complicated or not. Developers make software. If I like what they make, I use it.

As a user, I see three competing pieces of software here: X, Mir, Wayland. Two of them aren't ready for use. When they are, I'll make my judgement based on what I see. If Distribution X with Wayland is best, I'll use it. ("Best" defined broadly, including access to applications that run on it.) If Ubuntu with Mir is best, I'll use it. If something running X is best, I'll use that.

Competition is good. I'm not feeling any of the angst about Mir that some folks seem to feel.

(Standards are good, too, usually. But, they can be counterproductive when the need to remain faithful to an aging piece of software that enforces de facto standards stifles innovation and inmprovement.)

[tartalo]

OP said

Please stop depersonalizing quequotion, he doesn't like it and I'd like you to respect it.

As long as common drivers are used for all solutions i don't see a problem

Drivers are just one part of the equation, an important one indeed, but there are others, like toolkits, DEs that previously worked on Ubuntu, other distributions... Mir departing from the agreed standard is very disruptive, and it will either give developers headaches or damage Ubuntu users' choices. There's still much to see, but the second option seems to be the one taking more weight (and it doesn't surprise me a bit).

Once Wayland matures i don't see why Mir could not gain support for it

In my opinion that would be a wise decision that shouldn't be delayed.

I don't see why Wayland protocol couldn't implement support for something tested and used in Mir.

Wayland is the agreed standard, if Mir developers come up with an idea that would benefit everybody I don't see either why that couldn't be added to the standard, bending an agreed standard to a "one distro only" solution is however quite a different thing.

They [other DEs] ]can still use Mir (add support) if it suits their purpose.

Not quite, Mir is not developed with anything but Unity in mind, there are contradictory messages about API / ABI stability ... They, Canonical, are not putting it easy for others to adopt Mir, and I don't really understand why they try to make users think otherwise. If there was no agreed standard adopting Mir as a de-facto standard could be a price someone might want to pay, but with Wayland already properly defined Mir is not very attractive.

About Canonical being upstream for others

Canonical has already been upstream for others, for example with upstart, or with lightdm. But with Mir it seems difficult, honestly.

[EgoGratis]

Please stop depersonalizing quequotion, he doesn't like it and I'd like you to respect it.

It's not depersonalizing it's standard procedure.

like toolkits

More work for Canonical.

DEs that previously worked on Ubuntu, other distributions...

Will remain working in my opinion for foreseeable future just like they do now and there will be some new possibilities on how exactly to use them.

Mir departing from the agreed standard is very disruptive, and it will either give developers headaches or damage Ubuntu users' choices.

I don't think users of Ubuntu will have less choices in foreseeable future probably there will be more choices but let's wait and see.

In my opinion that would be a wise decision that shouldn't be delayed.

Wayland still needs time to mature and delay is something that will happen anyway.

Wayland is the agreed standard, if Mir developers come up with an idea that would benefit everybody I don't see either why that couldn't be added to the standard, bending an agreed standard to a "one distro only" solution is however quite a different thing.

But imagine Ubuntu Touch using SurfaceFlinger at the beginning on mobile devices and it worked? And imagine Mir would be Wayland compatible ATM and "qq" and other DEs not using it anyway... And imagine this that "agreed standard" and Mir are not yet suitable replacement for X11/Xorg ATM. First we have to get there.

Not quite, Mir is not developed with anything but Unity in mind, there are contradictory messages about API / ABI stability ... They, Canonical, are not putting it easy for others to adopt Mir, and I don't really understand why they try to make users think otherwise. If there was no agreed standard adopting Mir as a de-facto standard could be a price someone might want to pay, but with Wayland already properly defined Mir is not very attractive.

You see being Wayland compatible right from the start would not made that much difference would it. That is why i do believe it's quite pointless to discus this before we have something that we can actually use in real world and to work better compared to X11/Xorg.

Canonical has already been upstream for others, for example with upstart, or with lightdm. But with Mir it seems difficult, honestly.

Yes upstart is good example how first there was upstart then somebody else made something else and now we have 2 competing solutions and that is about it. The sky isn't falling.

[neu5eeCh]

If you're interested, this thread is essentially identical to the thread I started here called Mir vs. Wayland & effect on Linux "Ecosystem"? Maybe the two threads should be combined?

Anyway, my impression is that Ubuntu is the 986 pound gorilla and that, like it or not, the other distros can either cling to Wayland -- and lose even more market share -- or adopt MIR. My 2 Cents (Canadian): Companies like AMD and INTEL are not going to be developing new drivers for Wayland and MIR. They're going to pick one and it's going to be MIR because that's where the market share is (or promises to be).

Regardless of my preferences (and I really only want what works) a better title to this thread might be addressed to distros committing themselves to Wayland: Stop this train before you're derailed by Wayland. I see a future where every distro that is not MIR-based is forced to rely on open source drivers or simply don't work on newer hardware. To make a very loose analogy, Ubuntu will be the Windows of the Linux-verse and the rest of the distros will be the marginal 1% (and that's probably how a nice little niche of linux users prefer it).

[llanitedave]

As a user, I have to admit I am not especially concerned if the lives of developers are complicated or not. Developers make software. If I like what they make, I use it.

If the developers are put off, they'll be more likely to make the software you like for some other system rather than Unity.

That said, I'm not feeling any sense of forboding. It's not like Mir is going to suddenly make all previously developed software nonfunctional. It's not the application developers that are going to be inconvenienced anyway. They use standard GUI tools that are already compatible with X. Mir will still have to allow all existing applications that run on the QT or Wx toolkits, for example. It will have to be compatible with the graphic outputs of existing game and imaging applications. It's the toolkit developers that will have to be kept happy. If future versions of QT or Wx won't work on Mir, then neither will the applications that use them. Application developers will either have to use toolkits developed specifically for Mir, or older versions that retained compatibility. Otherwise, it's the maintainers of Mir that are going to be kept busy constantly backporting.

[buzzingrobot]

If the developers are put off, they'll be more likely to make the software you like for some other system rather than Unity.

I have no deep-seated preference to Unity or any other platform. It's up to Canonical to make sure apps work on Mir. They have the incentive to do that, other developers don't. It's not like the OS X independent developer arena, where incomes depend on adapting to whatever new thing Apple rolls out.

If Canonical doesn't make that happen, c'est la vie.

Application developers will either have to use toolkits developed specifically for Mir, or older versions that retained compatibility. Otherwise, it's the maintainers of Mir that are going to be kept busy constantly backporting.

It would make sense for Canonical to provide those toolkits. Ideally, they'd engineer some magic into Mir to handle the compatibility.

On the device front, I'd expect Ubuntu to be the gatekeeper for those apps, giving them the opportunity to weed out any that don't deliver on Mir.

[EgoGratis]

Anyway, my impression is that Ubuntu is the 986 pound gorilla and that, like it or not, the other distros can either cling to Wayland -- and lose even more market share -- or adopt MIR.

I think we will be using Mir, X11/Xorg, Wayland based solutions for quite some time and there is no need for ultimatums like that.

My 2 Cents (Canadian): Companies like AMD and INTEL are not going to be developing new drivers for Wayland and MIR.

Sure they will as they already do in a way and this will only improve in the future.

[buzzingrobot]

I think we will be using Mir, X11/Xorg, Wayland based solutions for quite some time and there is no need for ultimatums like that.

Seems more of a prediction than an ultimatum.

In any case, it's pointless to discuss market share when we're discussing desktop Linux, There is no desktop Linux market. A market is a place to buy and sell. No one is buying and selling desktop Linux.