Waste of resources and development: likely to ostracize Ubuntu from the linux community.

[quequotion]

Developing Mir is a waste of resources that could be invested in improving other projects, particularly Wayland.

There is a great risk of Ubuntu being ostracized from the linux community. Most, if not all, Ubuntu derivatives have no plans or means to support Mir and will not be able to keep up-to-date with Ubuntu development. Very few valid reasons have been given for the creation of Mir over using that time and those developers to improve Wayland; and most of them have been redacted.

I personally doubt that Mir will be a viable replacement for Xorg at any time in the foreseeable future and I feel that much of the work put in to Mir was less than honourably "borrowed" from Wayland, which Canonical had promised to support and which was finally seeing the possibility of adoption in several other distributions. Although claims have been made that corporate sponsors are interested in Mir, I doubt their dedication and I don't see that as reason enough to invalidate another, more complete, project.

What is the real motivation behind Mir's development? Is it worth the cost of time and resources if it fails? Is it worth being all alone and incompatible with everything? Is it worth losing lots and lots of users?

[Daniel van Vugt (vanvugt)]

I respect your opinion. And your opinion is valid, even if some facts like "borrowing" work from Wayland are inaccurate. If you look at the Mir source, you will find it has been written entirely from the ground up by Canonical, only using some external libraries from other projects to handle input etc.

But opinion is not a bug. And bugs.launchpad.net is not the right place to have this argument.

[quequotion (quequotion)]

I understand. "Borrow" is indeed an inaccurate term.

I was referring to this statement:

"Mir is a Wayland-alike; we're piggybacking on a lot of good work done for Wayland."

https://plus.google.com/113883146362955330174/posts/PXc93m8nKwk

Perhaps I've filed the bug on the wrong target.
I see Mir as a flaw in the Ubuntu desktop development process.

[Daniel van Vugt (vanvugt)]

Oh, right. "we're piggybacking on a lot of good work done for Wayland" is referring to changes already made to the Mesa code. The comment is not related to the Mir code.

[quequotion (quequotion)]

Ah, that isn't at all clear the way it's presented.

Actually, this statement opens up a few cans of worms: it gives the impression that Mir will be somewhat Wayland compatible, which may be entirely false. It also implies that Canonical took advantage of work done on/for Wayland when abandoning it for Mir which should be completely legal since the code is not derivative and open-source licensing may even allow it anyway, but it sounds a bit cutthroat.

[Daniel van Vugt (vanvugt)]

It's not really cut-throat.

There's a certain amount of work that Mesa needed to have to support new platforms like Wayland and Mir. And both must use Mesa. So if both are dependent on the same Mesa libraries then it makes sense there will be some overlap in effort.

If Wayland had not done the work then Mir would have. But the Mir team has done further work in Mesa on top of that since then.

There's nothing particularly controversial about Mir and Wayland using some of the same open source libraries.

[quequotion (quequotion)]

Indeed, I meant that the wording of the statement could be misinterpreted as such.

[quequotion (quequotion)]

This is not invalid. It is my opinion. I believe Mir represents a substantial flaw in Ubuntu's design process.

[quequotion (quequotion)]

Fix Released:
https://insights.ubuntu.com/2017/04/05/growing-ubuntu-for-cloud-and-iot-rather-than-phone-and-convergence/

[Daniel van Vugt (vanvugt)]

Actually Mir is still being developed by Canonical right now. Just not being used on Ubuntu Desktop.

It's important to note that:
  Mir is a display server AND a protocol.
  Wayland is NOT a display server and is ONLY a protocol.

I didn't understand the difference fully before, but I do now.

Since Mir and Wayland only conflict in terms of whose protocol you use, that's what this complaint is really about if anything. However good news: In Mir 0.28 they have added support for the Wayland protocol. Now Mir coexists with Wayland and everybody wins. There is now zero overlap and zero conflict.

[quequotion (quequotion)]

That's good to know, but I was never concerned about the technological similarities and differences of the two. I have always considered it a waste of resources to pursue the Mir project. Given the unlikelihood of Canonical securing any contracts with GPU vendors to supply drivers that support Mir, the ongoing and gaining momentum of Wayland, the ostracization of Ubuntu from other GNU/Linux distros, and the unpopularity of Unity, it has never been worth doing, in my opinion.

I hope I haven't jumped the gun, I marked the bug "Fix Released" as I interpreted Shuttleworth's intention to give up on such projects as Mir.

[quequotion (quequotion)]

That said, if it is now possible to integrate the two I see even less reason for Mir to exist.

[Daniel van Vugt (vanvugt)]

If Mir is speaking the Wayland protocol now like other shells, that means it makes no sense to compare Mir to Wayland. They are not conflicting things and they are not trying to do anything similar to each other.

Instead any shell built on Mir should be directly compared to Gnome Shell, KDE, MATE, Xfce etc. And if Mir has done a good job then it should be easier to build shells on it than it was to build those other existing shells.