Ubuntu Allegedly To Have Its Own X, Wayland Alternative

[dee.]

This is so typical of Canonical of late. They don't care about the ecosystem, they don't care about upstream nor downstream, they just want to be the new apple.

Are Ubuntu devs so antisocial that they are unable to work with or collaborate with any outside devs? Why do they need to reinvent everything in-house?

[F i L]

So why not wait till there's some actual information before having an opinion.

It's not bad for Canonical to hear a bit of outcry from those who understand what Wayland is, even at only their consideration of going their own way. The display server is the next most important thing to the kernel for making a consistent, cross-distro environment for developers. Proprietary drivers (those needed for games, especially upcoming Valve games) need to specifically support each display server, so adding another layer of fragmentation here would be suicide to Linux Gaming (right when it's actually picking up momentum). The majority of the Linux community and Xorg developers are now supporting Wayland (which is a major improvement to X), it's niave of Canonical to think they could match this movement. Much better for them to contribute and be more aggressive with influencing it's design direction.

Keep in mind one thing however: Canonical employ intelligent and capable engineers. One expects that whatever decision they take, they'll have some fairly good reasons for it.
( anticipating moans about them dropping Gnome shell for Unity, answer this question honestly - knowing their goal was to build the Unity experience, do you honestly think they could have done it while also aligning development with Gnome shell? (whether you like Unity or not is irrelevant) )

Given Canonical's history, it's not hard to imagine them jumping from "we're thinking about it" to "we're doing it" prematurely (like they did with Unity, regardless of weather you like it or not). Even brilliant engineers are only capable of so-much work per-person. If Ubuntu had chosen Gnome (or forked it like Mint/Elementary), then that would be receiving a lot more attention now, and be a lot more developed today. I like both Gnome and Unity, and I have complaints about both. But it's funny to me now, that with only a few extensions Gnome can behave very similarly to Unity.

[MartinN]

Jono Bacon is just the community manager, so he is basically the messenger.
Don't shoot the messenger.
He isn't the guy who decided that they should develop this software.

Be that as it may, despite my over-reacting, this just wreaks of disaster. Linux -must-/-must-/-must grow up, that is, its developers must MATURE to the point where they need to stand on each other's shoulders rather than stand on each others' toes. The single, sole reason, IMO, why Linux hasn't conquered the desktop is this Linux NIH syndrome, for a lack of a better term....

Please people - wake up - there hasn't been a radical idea in CS since the 'invention' of primitive recursive function - stop stroking your own egos and get with the program. We've got Apples to eat and Microsofts to bury.... And it won't get done without surrendering your egos!

[MartinN]

The display server is the next most important thing to the kernel for making a consistent, cross-distro environment for developers. Proprietary drivers (those needed for games, especially upcoming Valve games) need to specifically support each display server, so adding another layer of fragmentation here would be suicide to Linux Gaming (right when it's actually picking up momentum). .

Thank you for saying it like it is!

[MartinN]

Keep in mind one thing however: Canonical employ intelligent and capable engineers. One expects that whatever decision they take, they'll have some fairly good reasons for it.
( anticipating moans about them dropping Gnome shell for Unity, answer this question honestly - knowing their goal was to build the Unity experience, do you honestly think they could have done it while also aligning development with Gnome shell? (whether you like Unity or not is irrelevant) )

Yes, they may employ the most intelligent and capable engineers - but do they employ wise engineers and/or management+leaders? If they re-invent Wayland, in any shape or form, however miniscule, to cause division then I would say they lack wisdom, despite their technical brilliance.

[Ericg]

Wayland = good
Anything not wayland = bad (for the rest of human existence)

Is this correct?

Its a little too generalized but the jist of it.... yes. Or atleast:

Wayland = good
X and Android DS = Bad.

If you want to know what makes Wayland good or what makes X bad, check the Wayland page up on freedesktop.org. What it comes down to is: X was designed for a VERY different era of computing with a VERY different focus and architecture. "X" is a protocol with a lot of overhead and a lot of legacy. Its also an affirmed standard so its not like X.org can break the "X" standard and no one will care. They'd have to write their own protocol anyway, name it something different. Which basically happened anyway cuz most of the Wayland developers are all X.org developers. So they KNOW what X's big limits are and can design around them.

People get pissed about Wayland breaking things cuz its not "X" but the reality is the X server is bypassed as much as humanly possible nowadays anyway. Even the entire "X" library got replaced awhile back (Xlib has been formally depreciated and replaced by XCB) This is mainly because Xlib is synchronous, ergo every command that is sent has to have a return command that acknowledges its acceptance and success before the next command can be sent. It sounds like a great idea but when you are talking about something that needs to be firing off hundreds of commands a second and it really doesnt matter if ONE frame is dropped, all that waiting and all that round-tripping is actually bad.

[MartinN]

Its a little too generalized but the jist of it.... yes. Or atleast:

Wayland = good
X and Android DS = Bad.

If you want to know what makes Wayland good or what makes X bad, check the Wayland page up on freedesktop.org. What it comes down to is: X was designed for a VERY different era of computing with a VERY different focus and architecture. "X" is a protocol with a lot of overhead and a lot of legacy. Its also an affirmed standard so its not like X.org can break the "X" standard and no one will care. They'd have to write their own protocol anyway, name it something different. Which basically happened anyway cuz most of the Wayland developers are all X.org developers. So they KNOW what X's big limits are and can design around them.

People get pissed about Wayland breaking things cuz its not "X" but the reality is the X server is bypassed as much as humanly possible nowadays anyway. Even the entire "X" library got replaced awhile back (Xlib has been formally depreciated and replaced by XCB) This is mainly because Xlib is synchronous, ergo every command that is sent has to have a return command that acknowledges its acceptance and success before the next command can be sent. It sounds like a great idea but when you are talking about something that needs to be firing off hundreds of commands a second and it really doesnt matter if ONE frame is dropped, all that waiting and all that round-tripping is actually bad.

Thank you Eric for this comment - just want to add, Google's management actually IS wise, and I am going to place a $100 USD bet right now if any takers exist, that Google will eventually (e.g. 2 years from now) make use of Wayland once it is stable enough, and use it as Android's display server.

Who wants to prove me wrong?

[johnc]

Yes, they may employ the most intelligent and capable engineers - but do they employ wise engineers and/or management+leaders? If they re-invent Wayland, in any shape or form, however miniscule, to cause division then I would say they lack wisdom, despite their technical brilliance.

Considering how often my Ubuntu machine boots to a blinking cursor on a black screen, I really question the "technical brilliance" part.

That's why I feel confident in saying that they have no way of pulling this off... at least not on the desktop. No way on the desktop.

[johnc]

Thank you Eric for this comment - just want to add, Google's management actually IS wise, and I am going to place a $100 USD bet right now if any takers exist, that Google will eventually (e.g. 2 years from now) make use of Wayland once it is stable enough, and use it as Android's display server.

Who wants to prove me wrong?

It's possible, but I don't see the advantage. What's worse is that it would provide another dust-up in the Android OS version fiasco... because I suspect vendors won't be porting their existing GPU drivers to Wayland. Maybe okay for future SoCs, but existing devices won't be updated.

[MartinN]

It's possible, but I don't see the advantage. What's worse is that it would provide another dust-up in the Android OS version fiasco... because I suspect vendors won't be porting their existing GPU drivers to Wayland. Maybe okay for future SoCs, but existing devices won't be updated.

As long as "vendors" in your statement excludes Intel - who has OSS-ed the Intel HD drivers, quite frankly I don't care if the narcissists at NVidia or the day-late-and-a-dollar-short AMD/ATI folk don't port their GPU drivers to Wayland whatsoever. Just give me a laptop with Intel HD graphics, a solid Linux Intel HD GPU driver/OpenGL implementation that works with Wayland smoothly, and I will forego NVidia and AMD/ATI any day ... If I want to play games, I'll play them on my XBox....