The Story of Ubuntu's Mir Abstraction Layer (MirAL)

[TheBlackCat]

All of these applies to the Wayland too. It causing additional work, and toolkits can't hide all differences (you mentioned firefox), it doesn't bring any visible advantages, it will harm users, there will be tons of problems with outdated apps, drivers etc. Why not stick with X11?

The big difference is that Wayland does bring visible advantages over X11. Better security, less tearing, less lockups, better performance, fewer layout glitches, etc. Mir, on the other hand, doesn't have any advantages (at least none that the Mir developers have bothered to express and that weren't immediately shown to be false), and lots of disadvantages.

[TheBlackCat]

look at the people commenting on Mir articles against Mir and look for them adding something to an article about Wayland.

Yes, adding falsehoods and FUD about Wayland. There are a lot of misconceptions about Wayland in the Mir camp, even from the Mir developers.

It seems to me that the people who do write interesting articles or comments about Wayland, are not the same people who are constantly attacking Mir, because they're focusing on Wayland.

A lot of people associated with Wayland development have written articles criticizing Mir from a number of standpoints, right from the very beginning when the justification for Mir from Canonical was based on outright falsehoods about Wayland.

On the other hand, the people who are writing interesting articles and comments about Mir, doesn't attack Wayland, because they're focusing on Mir.

Except the very first article about Mir, the one announcing its development, attacked Wayland based on false reasons. bregma in this very thread has repeatedly attacked Wayland based on false reasons.

The Free Software community is never going to evolve into a Single Hierarchy, which to my mind is a good thing.

That would make sense in the Wayland/Mir debate if Mir actually had anything new, interesting, or even fundamentally different from Wayland to offer. But not only does it not, the small differences it did have are almost all gone as Mir has gotten closer and closer to Wayland and has adopted more and more of the Wayland stack.

[Khrundel]

The big difference is that Wayland does bring visible advantages over X11. Better security, less tearing, less lockups, better performance, fewer layout glitches, etc. Mir, on the other hand, doesn't have any advantages (at least none that the Mir developers have bothered to express and that weren't immediately shown to be false), and lots of disadvantages.

You must be kidding.
First, Wayland, being a protocol, can't fix any glitches/tearing. Wayland compositors can. But all modern X11 WMs are using composition, so, maybe there are some problems with glitches on X11, but I haven't noticed one for a long time. Second, all layouting currently done by toolkits, they use same layouting on X11. Third, better security always mean less possibilities. In case of Wayland, for example, for some security reasons, they don't allow top level windows to determine their screen positions. They say this decision is final. This means apps must use some non-portable workaround simply to save window position! Wine developers can't implement Wayland backend because Windows apps heavily rely on ability to open a window at any specified location.
So, these advantages over X11 are controversial. And your confidence about Mir lacking same advantages based on nothing.
The funny thing about all Mir critics is that you think about Wayland as some kind of rightful heir to X11 while Mir is usurper. That is not true.

[Gusar]

The funny thing about all Mir critics is that you think about Wayland as some kind of rightful heir to X11 while Mir is usurper. That is not true.

Everyone agreed on a X11 -> Wayland transition. Where "everyone" means windowing system devs, GPU driver devs, toolkit devs, application devs... They all agreed that a move to something more modern than X11 should be made, and because such a move is very disruptive, it was important to agree on the destination. All those listed agreed on Wayland being the destination, and worked towards making it happen. Only much later, when a lot of that work was already done, Canonical came along with their brilliant idea and suddenly there was a second destination, increasing the disruption.

Wayland *is* the heir, based on consensus. And Mir *is* the usurper, going against the consensus and causing unnecessary disruption.

[TheBlackCat]

You must be kidding.
First, Wayland, being a protocol, can't fix any glitches/tearing. Wayland compositors can. But all modern X11 WMs are using composition, so, maybe there are some problems with glitches on X11, but I haven't noticed one for a long time.

It isn't a matter of having a compositor or not. That is a necessary, but not sufficient, condition. The design of X11 makes these sorts of problems impossible to solve. Wayland, on the other hand, is specifically designed to provide the right APIs to allow applications to avoid them. For example there is no reliable way for applications to properly align videos with the contents of a window, menus with the contents of windows, or even the contents of a window with the window decoration, under X11, while there is under Wayland. There is no way under X11 for applications to get the timing of individual rendered frames while there is under Wayland. So although yes, compositors and toolkits need to be designed to work with the Wayland API, there is simply no way for them to fix these problems under X11 while there is a way under Wayland.

Second, all layouting currently done by toolkits, they use same layouting on X11.

That is only true for widgets. For things like overall window contents, "special" areas like videos or openGL sections, and even menus, you need to talk to the windowing system. Wayland provides good ways to do all of those things, while X11 doesn't. Even if you do that through a toolkit, there is no way for the toolkits to reliably implement such things under X11, while there is under Wayland.

Third, better security always mean less possibilities. In case of Wayland, for example, for some security reasons, they don't allow top level windows to determine their screen positions. They say this decision is final. This means apps must use some non-portable workaround simply to save window position!

Of course, the whole point of a window manager is that it manages windows. Window managers can, should, and do remember window positions under X11 and Wayland. Why does the application have to duplicate that functionality?

And your confidence about Mir lacking same advantages based on nothing.

I meant Mir advantages over Wayland, not over X11.

The funny thing about all Mir critics is that you think about Wayland as some kind of rightful heir to X11 while Mir is usurper. That is not true.

It isn't a matter of being an usurper, it is a matter of it being a waste of time and resources without providing any benefit to anyone over something that already existed. And not just Mir developers' time, it also wastes the time of people who had no say in Mir's development.

[shmerl]

Amazing. Canonical just announced the end of Unity. So Mir will go away as well I suppose.

[cynical]

Amazing. Canonical just announced the end of Unity. So Mir will go away as well I suppose.

Yeah I wonder what you people will have to whine about now.