Xfce, LXDE, & GNOME Are Running On Ubuntu XMir

[chrisb]

Ehhr?!? Why would anyone do that again? Your distribution is not going to use it, why spend time on it?

Developers tend to be inquisitive and interested in software; I had assumed that anyone commenting on Wayland/Mir would have at least tried them both out. It is hardly possible to comment on technical solutions without having spent some time investigating the options.

At the moment I haven't seen anyone outside of Ubuntu looking into Mir.

That's a shame. If no one has spent any time looking into it, then no one is going to be able to form a rational opinion about it. And without that, all we are left with is emotions, personal bias and politics.

A lot of things *could* be done, but I just don't get why anyone would.

Like I said, cross-pollination of ideas is a good thing. Ubuntu is the top Linux desktop distribution right now, and they got there by solving some problems that the others didn't. It's worth a small amount of effort to keep an eye on what they are doing. In general, it is a good idea to expose yourself to as many ideas and concepts as possible to gain an understanding of what is possible, and what is desirable. Isn't learning from others one of the main benefits of open source?

[phoen1x]

Ubuntu is the top Linux desktop distribution right now, and they got there by solving some problems that the others didn't.

Care to name atleast one problem that Ubuntu solved? Thanks.

[mrugiero]

a large part of Mir is copied straight from Wayland

Aside from the idea and the algorithms involved, is there any Wayland code in Mir?

We have 3 or more different Kernel, We have 3 or more sound server, we have X different file system, We have X different package manager....
OSx or Windows have more than 1 DE ? bad comparison. Or you suggest we need to do the same with Linux ? 1 Kernel, 1 File system, 1 DE, 1XServer, 1 Company ?

I love when people from open source community trying to discredit Canonical explaining that the better way to do is the way of proprietary solutions... Big Facepalm.

When you suggest that a company/people can't do it's own project you fail totaly. Open Source encourage fork, new projects, creativity and all people to do new stuffs... If you rant against this point for some idiot reason you are against the basis of the Open Source rules.

None of your examples are valid, for one basic thing: POSIX. As long as they respect the POSIX standard, you can write your software for one platform and it will be outright (at the source level) compatible with all of them. That's why they are the same as different DEs to this subject, because none of them breaks source compatibility.

I read perfectly and is more absurd think I have ever see.
If red-hat do the same, all the people say ... nothing because lot of people don't do the same mental storm when Red-Hat take decisions. In Fact Google do, Mozilla do and no one cry about the end of the open source world...

You don't respond to my first question :
- So you consider BSD, PulseAudio, SystemD and all File system like absurd projects and horrible business decision ?
Or from some magic trick this rules apply only on Canonical Display Server ?

Your example of proprietary system is only here to demonstrate that you are totally wrong... But the worst, you don't realize how...

Well, for a start, both Google and Mozilla have completely different targets than Ubuntu. None of them targets the desktop. When Android came out, the only other Linux based OS for smartphones was Meego. There wasn't a developed ecosystem to screw up to begin with. If Mir was only targeted to phones, I wouldn't mind. I think it's more open than Android anyway, so it might actually get a real free software community on that side.

About Red Hat, the proof charge is on you, come up with an example of a disruptive Red Hat decision and I'll tell you what I think about it.

3 - Yes big communication mistake here. But we can read a lot of "incorrect statement" from the community against Ubuntu project from long time ago...
4 - No they speaking about other possibility than wayland from one year I think, staying officially with Wayland. And Canonical do what they want. As red-Hat can do it the same. Is business decision.

In fact I can understand some people don't like the method but the Mir project stay a free software project and we have to judge it from a technical point of view, not from some obscur and absurd ideology.

Point 3: name someone that MATTERS that spreads FUD about Mir. I'm not talking about Phoronix readers, I'm talking about real devs, Red Hat, something like that. There's a big difference on spreading FUD while being a John Doe (only the morons really buy that) or while being Canonical (where a lot of maybe not too techie but relatively smart people will buy anything you say).

About the technical points, the thing is, when you are the one opposing consensus, you are the one who should give technical arguments for that. Prior to Mir, there was agreement between most of the graphical developers on the Linux world that Wayland was the way to go. For the time being, the only technical argument was 'server allocated buffers', which had no detailed explanation of why they consider it better. The ones who weren't on the Wayland's wagon at that time was mostly interested in network transparency, and that's not solved by Mir either. On all the other things, a Mir dev admitted Mir and Wayland are supposed to behave the same, or Wayland is flexible enough to give it a shot.

Making technical decision is not a "toxic" decision. Ranting again project, people and company just for pleasure is it.
And in your link you have another example that Canonical in 2010 was looking for different solution than Wayland.

Wrong again. A technical decision must be, before, an informed decision. If you say it was a 'technical' decision, then they should have:
1. Asked if what they considered missing was going to be part of Wayland. This is standard procedure when deciding to implement something on the open community, you inform yourself about existing projects with common goals, and check if there is any real difference with what you want to do. As a Mir's dev said on a G+ note (I'll look for it if you don't believe me, but I'd have to check a big mail list to do so), most differences were actually things Wayland does have now. Only, they didn't ask if they were planned. They just wasn't there a year ago, when they decided to do their own.
2. If there is any difference between what you want and what is done or being done, you either justify (just to know what you are doing) why you think the way to go is other, or you accept the other people's point of view. Both imply in practice the same procedure, which is a careful analysis of the reasons that lead to that difference. You might even convince the other and avoid a fork if the current plan can't meet your needs.
3. If the above lead to negative answers, you do your own.

If that's not done, then, by basic logic, your decision wasn't really technical, because you didn't even know what technical differences were between what you did and what the others offered you.
If it's not clear enough, I'll put it in other words. If you are diverging because of technical reasons, there must be technical differences between what you want and what there is. Right now, the only one Canonical did named was the allocation on the server side. Which I'm not sure if Wayland is even unable to do, so there might still be ZERO differences in the tech side. And nobody likes, and less the ones who does it for free, having to port something or risking the project itself (Ubuntu is still the majority of the consumers desktop) with no technical reasons to do so other than 'we didn't feel like talking'.

no! it was the only alternative while many (not all!!!) agreed we need something new over X. regarding wayland itself many if not most people were skeptical at the beginning. in fact there were many voices say "why waste resources into wayland... we do not need another display server".

forgot that?

So? There's the ones that thought Wayland was the way to go, and then the ones who thought there was no need for another display server. Tell me which ones are asking 'well, maybe YET ANOTHER one is needed, let's go Mir' in your statement.
In fact, most of Wayland detractors WANT NETWORK TANSPARENCY, the same you do. Does Mir solve this? I bet it doesn't, or it would have been one of the technical arguments.

[mrugiero]

Developers tend to be inquisitive and interested in software; I had assumed that anyone commenting on Wayland/Mir would have at least tried them both out. It is hardly possible to comment on technical solutions without having spent some time investigating the options.

Investigating doesn't mean using. You read documents, check what technical differences there are, and if they seem compelling to you, maybe you develop for it.
Pro-tip: there aren't real tech differences aside from the buffers being allocated server-side and a different API. Not really a programmer's wet dream, comparing with Wayland at least.

That's a shame. If no one has spent any time looking into it, then no one is going to be able to form a rational opinion about it. And without that, all we are left with is emotions, personal bias and politics.

Same answer. If it were about implementation, you'd be probably right, but is not, it's about specification. Nobody is saying 'Mir is full of bugs' or anything like that, which would require using it to know.

Like I said, cross-pollination of ideas is a good thing. Ubuntu is the top Linux desktop distribution right now, and they got there by solving some problems that the others didn't. It's worth a small amount of effort to keep an eye on what they are doing. In general, it is a good idea to expose yourself to as many ideas and concepts as possible to gain an understanding of what is possible, and what is desirable. Isn't learning from others one of the main benefits of open source?

And most programmers are open for this. They are exposed to the idea. The idea doesn't offer them anything new compared to Wayland, and that's why they don't care supporting it; because it's just extra work, with no extra benefit.

[k1l_]

Care to name atleast one problem that Ubuntu solved? Thanks.

Way back noone thought you will need a GUI Installer that could be used by your grandma. Ubuntu solved that.

[bkor]

Developers tend to be inquisitive and interested in software; I had assumed that anyone commenting on Wayland/Mir would have at least tried them both out. It is hardly possible to comment on technical solutions without having spent some time investigating the options.

Mir is specific to one distribution. If you're developing something for multiple distributions and using Ubuntu I can understand that you look into Mir. But at the moment one distribution goes its own way. Their target is Unity, not the desktop environment you're developing for. Futher Mir requires switching distributions, etc. That is a significant time investment for unknown benefit.

That's a shame. If no one has spent any time looking into it, then no one is going to be able to form a rational opinion about it. And without that, all we are left with is emotions, personal bias and politics.

I've repeatedly given rational arguments why Mir is not at option for GNOME. Yet again you suggest that "all we are left with is emotions, personal bias and politics". A good technical argument is that the focus is Unity and that API is not stable. That is what was mentioned to GNOME. It changed recently. Going by history, it probably will change soon again.

Note that 60% of the GNOME modules work under Wayland. None of them work under Mir. Probably it works on some Mir developers machine or something, but kind of hard to collaborate.

Like I said, cross-pollination of ideas is a good thing. Ubuntu is the top Linux desktop distribution right now, and they got there by solving some problems that the others didn't. It's worth a small amount of effort to keep an eye on what they are doing. In general, it is a good idea to expose yourself to as many ideas and concepts as possible to gain an understanding of what is possible, and what is desirable. Isn't learning from others one of the main benefits of open source?

You assert things without backing things up. E.g. why they got popular (wtf does that have to do with things), small amount of effort (already told you that it is not a small amount of effort and things have been hidden and things have changed drastically).

Anyway, suggest to quit playing the emotional card. Mir is not an option and that is not due to Canonical hate.

[phoen1x]

Way back noone thought you will need a GUI Installer that could be used by your grandma. Ubuntu solved that.

False. When i first tried mandrake linux (version 9 or smth) it had GUI installer and was pretty easy to install for a complete linux newbie.

[erendorn]

Aside from the idea and the algorithms involved, is there any Wayland code in Mir?

I think that at least XMir and the libs used to make Mir work with android graphic drivers reuse heavily code written for their wayland counterparts.

[PsynoKhi0]

Care to name atleast one problem that Ubuntu solved? Thanks.

Increased public awareness regarding alternatives to windows and mac on the desktop.

Back to topic: had anyone else scratched their heads seeing the memory usage?

[Ikem]

You can do that already, as there are native X-Servers for Android.

I only know of android-x-server. And that's more a proof of concept.