Mir Now Allows Multi-Threaded Compositing

[TheBlackCat]

"De facto standard" IS the standard. It is one only when everyone agrees to follow it; who cares what is written on paper if it does not coincide with reality?

No, sorry, you don't get to arbitrarily redefine words to suit the argument you are making. "Standard" in computing has a very specific definition. "de facto standard" also has a very specific, completely different definition. The fact that these definitions contradict your position is a problem with your position, not with the definition.

I can't remember Debian abandoning dpkg because of LSB-compliance issues.

I also don't recall IE quickly embracing svg when it became a w3c standard (or are you saying the w3c is not a standards-setting organization?). Just because something is a standard does not mean everybody automatically uses it.

[ворот93]

Server-oriented means nothing in your context of energy consumption. Do you have _any_ numbers to back it up? Canonical has a big PR mouthpiece that keeps talking about efficiency while writing a lot of stuff in Python.

What's your problem with Python? Last I checked, portage in gentoo works a lot faster than its C++ rewrite.

[mark45]

What's your problem with Python? Last I checked, portage in gentoo works a lot faster than its C++ rewrite.

That it's interpreted? Also, Ubuntu isn't gentoo and the stuff that runs on Ubuntu in Python (software updater, software center, whatever) are not portage.
And on a different note, I used to bitch about how Java is faster than C/C++ because there were some benchmarks for some situations under some conditions that prove that, yes, I used to be an asshole just like you are now, except that I've grown up, and except that you're even more full of shit by saying it works "a lot faster", not just faster.

Server-oriented means nothing in your context of energy consumption. Do you have _any_ numbers to back it up?

[dee.]

Interpreted languages have their place, they're fine for non-speed-critical applications, and they're generally great for portability. Python is a great language, it has a very nice syntax, but there are things you don't want to write in it or any other interpreted language. When speed is an issue, compiled languages still generally sweep the floor with interpreted ones.

Mir is not going to become the desktop standard, for several reasons:
1 - Wayland has/will have the first-to-market advantage, it's more finished and will be usable sooner.
2 - People (rightly) don't trust Canonical, with their track record, to maintain a display server that would provide the needed functionality for everyone, not just for Unity.
3 - Canonical doesn't collaborate with other distros or desktop environments enough. To get your work used by others, you need to play nice in the sandbox.

Every Ubuntroll who's here just religiously defending Canonical no matter what they do should remember that critical thinking is allowed even if you like something. I used to like Ubuntu, in fact I still do in some ways - I think 12.04 was a great release, and still is. Since 12.10 however they've been moving in a very bad direction, and so people get divided - some see it and say "Ok, this looks very problematic, maybe we should re-evaluate our stance about Ubuntu"... and others double down, refuse to see or hear anything that contradicts their fanboyish love of Ubuntu, because they care more about being a part of a group, about defining their identity by what OS they use (they see themselves as Ubuntu-users, for better or worse) than looking at things honestly and critically.

[Sonadow]

Mir is not going to become the desktop standard, for several reasons:
1 - Wayland has/will have the first-to-market advantage, it's more finished and will be usable sooner.
2 - People (rightly) don't trust Canonical, with their track record, to maintain a display server that would provide the needed functionality for everyone, not just for Unity.
3 - Canonical doesn't collaborate with other distros or desktop environments enough. To get your work used by others, you need to play nice in the sandbox..

How can Wayland have first-to-market advantage when it's still not even considered stable enough for mainstream deployment. Having packages for it and Weston in the repositories does not mean anything the display still defaults to using X. And Wayland support for the respective graphical toolkits are still moving along in their own sweet time.

As much as I don't like Mir, the fact that Ubuntu (and its derivatives) is the most widely used distribution, coupled with the recent announcement that Kylin's new base will be Ubuntu, only makes it appear that adoption of Mir is almost guaranteed instead of Wayland simply because of sheer numbers. Don't forget that Ubuntu has been working to seduce Nvidia and AMD to release Mir-compatible versions of their blob drivers too, and that is enough to sway a lot of people over. Especially those who always play the upgrade game with new hardware on launch.

The only way Wayland can beat Mir in the adoption game is to make it fully usable, have all graphical toolkits support it and have all distributions default to using Wayland in the default installation by Jan 2014.

[SilverH]

Hi guys

I'm a Ubuntu user, but I'd really like to understand all this mess.

Correct me if a got anything wrong (besides my poor english ):
Wayland is a library which will be a standard for ppl to develop X like screen servers.

If it's just a library, what is the tecnical reasons to Ubuntu avoid using wayland?
Is there any solid tecnical reason to avoid it? (from the Ubuntu dev's point of view)

Another thing I don't understand is why ppl get so concerned about fragmentation. I thought linux users in gerenal were used to it.
Why fragmentation in this matter should be different from the others?


Peace

[erendorn]

How can Wayland have first-to-market advantage when it's still not even considered stable enough for mainstream deployment. Having packages for it and Weston in the repositories does not mean anything the display still defaults to using X. And Wayland support for the respective graphical toolkits are still moving along in their own sweet time.

Can you please make a comparison of all your points above between Mir and Wayland?
For example, how would your consider the speed of the Mir support in the various toolkits, compared to Wayland one?
Do you have any data on where Nvidia support stands for both display servers? Maybe you even have data on how fast this support is progressing? That would be awesome.

[Ericg]

Only one? Hey, I've been telling anyone who listens that Wayland is a standard, over and over... but the Ubuntrolls don't listen. They just go on and on about "lolz wayland still isn't finished we want a new display server now!!!!111" which kind of makes me wonder why they think it's a good idea to start an entirely new one, instead of using all the effort already put in to wayland...

Sorry Dee :P just seems like me and daniel stone are the only ones posting correct info on wayland/X sometimes

[ворот93]

1 - Wayland has/will have the first-to-market advantage, it's more finished and will be usable sooner.

It doesn't function yet so no, it does not necessarily have

2 - People (rightly) don't trust Canonical, with their track record, to maintain a display server that would provide the needed functionality for everyone, not just for Unity.

Mom basement geeks care, "trust" etc etc etc. Simple users pick what works and Ubuntu works best.

3 - Canonical doesn't collaborate with other distros or desktop environments enough. To get your work used by others, you need to play nice in the sandbox.

So what? Unity still beats GNOME Shell by a VERY long shot.

The "community" are losers and disappointment who were impotent when it came to solving bug #1. Canonical has what it takes to actually get mass adoption.

[erendorn]

Hi guys

I'm a Ubuntu user, but I'd really like to understand all this mess.

Correct me if a got anything wrong (besides my poor english ):
Wayland is a library which will be a standard for ppl to develop X like screen servers.

If it's just a library, what is the tecnical reasons to Ubuntu avoid using wayland?
Is there any solid tecnical reason to avoid it? (from the Ubuntu dev's point of view)

Another thing I don't understand is why ppl get so concerned about fragmentation. I thought linux users in gerenal were used to it.
Why fragmentation in this matter should be different from the others?
Peace

Wayland is both a protocol and a library (but you could implement your own library and still follow the wayland protocol).
The protocol and the lib are quite similar to what Mir intends to do. There are no technical reasons at the moment to avoid it (none. even about the "android graphics", it's the same amount of effort to do it in Wayland an Mir). Maybe in the future, some new need will arise, and Wayland protocol will only be extended (because standard), while Canonical will be able to change Mir (because it won't be standard: they can decide anything).
Also, Canonical can do things quicker, because they do not have to follow specs, or implement things they don't think they need. On the other side, because it will not be standard, they need to support the rest of the stack (kernel side, should be easy, and toolkit side, which will be more difficult).
Canonical have decided that for them, control is better than shared development, moving to an open source, closed governance model. It will probably work for them, but much less for people that work with them.

People got concerned about the move, because:
- Canonical said false negative statements about wayland in their first blog post about mir. They have been removed since, but people still resent this.
- Governance model is becoming an important subject in the FOSS community (at least as much as licensing model). Canonical is growing more like Google (android) or Samsung (Tizen): you get the code, but you don't get to say what gets in it. It still better than closed source, but it still annoys the community.
- Open source projects can deal fragmentation, unless there is closed source code involved: the binary drivers (until now there was only X, so no fragmentation here). Both Mir and Wayland use an EGL base (standard, easily supported by any driver), and a few (one or two) incompatible extensions. That means that an EGL driver can support none, one, the other, or both display server. As each server needs additional work from the GPU vendor (who are many on the phone/tablet space, with limited support already), the fear is that only one of the two will get supported (at least for some hardware), and nothing could be done about that.

Hope it helps.