Advanced

Canonical's Mir for Ubuntu Linux: New Open Source Innovations?

Posted by thevarguy2 on · 56 comments
http://www.thevarguy.com; By Christopher Tozzi
Mail this story
Print this story

If Canonical succeeds in making Ubuntu Linux widespread on phones and tablets, it will revolutionize the role of open source software in mobile computing. But in a sense, the company has already achieved some remarkably revolutionary feats with Mir, the display server it started developing several months ago. This may sound like something only geeks can appreciate, but it could actually become a huge deal for mainstream IT.

Full Story

  Nav
» Read more about: Story Type: News Story; Groups: Community, Ubuntu

» Get new stories by e-mail: nightly or as they are published

« Return to the newswire homepage

  Related stories
» More like this

Comments (56)

36 earlier comments in the full thread.

It doesn't matter to me if composting and GLX is the modular part, or network invisibility is the modular part, or both in a structure which loads modules as they're needed (where have I heard that before?).

What matters to me is the seeming relentless drive to dump the UNIX environment that has worked so well.

I have no doubt that X11 is ancient and crusty and deserves to be replaced. Neither do I run the 2.4 Linux kernel any more.

Network invisibility may seem irrelevant to someone who just wants a drop-in Microsoft replacement.

That is not how I use my systems. Is it, really, how you use yours?
> What matters to me is the seeming relentless drive to dump the UNIX environment that has worked so well.

1) X11 is not a key part of the UNIX system and it never was.

2) The philosophy behind Wayland is not like systemd in any respect. It is effectively a reversal of the principal architecture of X11. X11 puts the network first. Wayland puts the compositor requirement first.

In praise of the developers of Wayland, they have decided on a very narrow remit for Wayland. They will not encroach beyond the core remit of Wayland which is to achieve a specific narrow goal, to provide a compositing and rendering core for Linux systems. Network and X11 support will be available but they will be separate agents and not part of Wayland.

In some ways, X11 as implemented on most distributions is a convoluted mixture of various components squeezed together in an unholy alliance. In a sense the scope of X11 with its oft accompanying modules has grown beyond the core remit of a network protocol.

The simple reason that it has got so crufty is exactly the reason that it has not stayed strong to the "do something simple, do it really well" idea. It has become uncomfortable within the environment that it is commonly used for.

I don't doubt that a large number of people use the X11 protocol for remote desktop-ing. The problem is that the "putting the network first and foremost" is not necessary the best fit for all the common use cases for desktop Linux now.

And as for running X11 and Wayland together as separate co-operating processes: what's so wrong with that? One is a compositor and one is a network component. Isn't that how we do things in Linux/Unix? It's certainly sounds a more pure architecture than the monolithic beast with "patches" that the typical X11 installation has become.
Quoting:What matters to me is the seeming relentless drive to dump the UNIX environment that has worked so well.


They way I see it, part of the idea behind Wayland is to prevent Unix/Linux from being dumped all together just because of its "ancient and crusty" windowing rendering system. :-)

> What matters to me is the seeming relentless drive to dump the UNIX environment that has worked so well.

Exactly my point. For once I 100% agree with Bob_Robertson.

I have nothing against evolution, or even redesigning something from scratch, as long as the new implementation doesn't ditch crucial features that are still well used.

You are right, it seems these days a new generation of young programmers that grew up on Windows have taken over who don't have any clue about all the nice features that Linux inherited from Unix and how much these features are key to why Linux is superior to Windows.

All they want is to make Linux more WIndows like... :-(
> the idea behind Wayland is to prevent Unix/Linux from being dumped all together just because of its "ancient and crusty"

That's just nonsense, Linux is currently running on more devices than ever before, nobody is dumping it and there is no real reason why Wayland would change this either way. If anything Wayland will contribute to divide the Linux desktop even more and hence make it less attractive.
> 1) X11 is not a key part of the UNIX system and it never was.

Huh? Of course it is, not for servers but it's THE Unix standard for any desktop system and GUI application

> X11 puts the network first

network transparency is KEY to every aspect of Unix, Unix was conceived from the very beginning as networked operating system, that's why Unix and Linux had such a headstart on Windows when the Internet became popular

> And as for running X11 and Wayland together as separate co-operating processes: what's so wrong with that? One is a compositor and one is a network component. Isn't that how we do things in Linux/Unix?

But as far as I understand that doesn't help at all, the presence of X11 next to Wayland is only meant to support 'legacy' X11 applications, all new Wayland applications (including desktop enviroments for Wayland) will not work over the network.
X11 being Network Transparent has nothing to do with network capabilities of Linux/Unix, which no one here is questioning.

X11 was used because there was nothing else at the time. When Wayland is ready, you will see a mass migration from X11 to Wayland. Heck, most popular DE already supporting it.

Quoting:Linux is currently running on more devices than ever before, nobody is dumping it and there is no real reason why Wayland would change this either way.


All the devices you speak of don't run X11, they run a tailored GUI like Android using WallpaperService Engine or something else the vendor of certain device developed. Ubuntu is starting its own Mir. I wonder why, have you wondered why?

It is obvious that the very informative material I supplied links to either didn't help or you didn't read. Either way, it seems useless to continue debating this issue with you while you keep confusing things.

> while you keep confusing things.

I'm afraid that's exactly what YOU are doing especially if you say nonsense like this:

> X11 being Network Transparent has nothing to do with network capabilities of Linux/Unix, which no one here is questioning.

Everything in Unix is build around the network, so the fact that X11 is and has been network-centric since the beginning is only a logical consequence of that.

> All the devices you speak of don't run X11, they run a tailored GUI like Android using WallpaperService Engine or something else the vendor of certain device developed.

And there is nothing wrong with that, "horses for courses" as the saying goes. What's wrong is trying to impose a technology suited for handheld devices onto the desktop.

I did read several of the links mentioned in this discussion, but most of the info presented there told me nothing I didn't already know. It only confirmed my concerns about the direction this is going.

> X11 was used because there was nothing else at the time. When Wayland is ready, you will see a mass migration from X11 to Wayland. Heck, most popular DE already supporting it.

That's exactly what I'm concerned about as that will only further divide and cripple the Linux desktop.

> Everything in Unix is build around the network, so the fact that X11 is and has been network-centric since the beginning is only a logical consequence of that.

Everything in Unix is *socket* oriented. That networks also use sockets in Unix is a logical consequence of that.

I have to ask though: if Wayland were to be widely adopted with an X protocol client, what do you fear that you would not be able to do, which you do now as a consequence?
So Wayland is modular and extendable, and merely puts the composter in the center instead of the 'side'.

Ok.

So take the network transparency function, make a module. Just as modern systems are so resource-heavy that the present X is _able_ to work, then the network transparency will work just fine as a module.

Done well, the switch from X11 to Wayland will be invisible, because the same functionality will exist in the new as existed in the old.

Try to swap them without that functionality, and I'll be miffed. And you don't want me miffed. Grumble grumble.

P.S. 4567, Indeed we do agree on the principles involved. I'm very glad to see it regardless of other disagreements. I have no "party loyalty" which means I can't be glad to work with someone where we agree just because we have some points of disagreement elsewhere. I've met such people with their "party loyalty", and they are pathetic.
>So take the network transparency function, make a module. Just as modern systems are so resource-heavy that the present X is _able_ to work, then the network transparency will work just fine as a module.

As far as I understand Wayland, that doesn't seem to be possible because Wayland lets the clients do the rendering directly. Correct me if I'm wrong as I don't yet fully understand this part of Wayland.

@Bob_Robertson: I have no "party loyalty" either, I think for myself and agree or disagree with people on each subject individually based on my thoughts, research and conclusions about the specific subject, not because of affiliations.

> what do you fear that you would not be able to do, which you do now as a consequence?

I'm not quite sure I understand your question, I already said network transparency is very important to me, as I make use of it all the time both privately and professionally.
I see we are not making progress and I can't come up with a better convincing arguments than This Video

So I suggest you watch it fully again and this time make sure till the end since that is where the important points are.

Useful info about where X was started and it wasn't on Unix/Linux.

Wikipedia: X Window System

Quoting: Origin and early development [edit] The original idea of X emerged at MIT in 1984 as a collaboration between Jim Gettys (of Project Athena) and Bob Scheifler (of the MIT Laboratory for Computer Science). Scheifler needed a usable display environment for debugging the Argus system. Project Athena (a joint project between Digital Equipment Corporation (DEC), MIT and IBM to provide easy access to computing resources for all students) needed a platform-independent graphics system to link together its heterogeneous multiple-vendor systems; the window system then under development in Carnegie Mellon University's Andrew Project did not make licenses available, and no alternatives existed.

The project solved this by creating a protocol that could both run local applications and call on remote resources. In mid-1983 an initial port of W to Unix ran at one-fifth of its speed under V; in May 1984, Scheifler replaced the synchronous protocol of W with an asynchronous protocol and the display lists with immediate mode graphics to make X version 1. X became the first windowing system environment to offer true hardware independence and vendor independence.
That's a 45min video, I wish I had that much spare time...
You can skip the beginning.

> As far as I understand Wayland, that doesn't seem to be possible because Wayland lets the clients do the rendering directly. Correct me if I'm wrong as I don't yet fully understand this part of Wayland.

One of the biggest bugbears about X is that the simple render is hardly used by anyone at all these because it's so useless. Everyone uses XRender instead which, if I understand correctly, means that you draw your own stuff anyway.

Even if this wasn't the case, most apps these days use toolkits which should insulate the apps from the backend implementation.
Some interesting comment on the state of X:

http://people.freedesktop.org/~daniels/lca2013-wayland-x11.p...

The pdf is a bit annoying: I just scroll rapidly through it.
> I'm not quite sure I understand your question, I already said network transparency is very important to me, as I make use of it all the time both privately and professionally.

Just curious: in what way do you use network transparency?

Are you running a thin client to an X-"hosted" app (e.g. ssh) or a 'thick' app connected to an X-"hosted" renderer?
> Just curious: in what way do you use network transparency?

I ssh into remote machines and then lauch GUI apps that run on those remote machines but display their GUI on my desktop. This is quite common practice (ok maybe not for those people who only have a single PC at home, but for lots of people with multiple PC setups and in work environments).
> I ssh into remote machines and then lauch GUI apps that run on those remote machines but display their GUI on my desktop.

Exactly.

The network agnosticism allows anything that can be reached through an SSH session to open a GUI application to display on my local machine. This traverses NAT, firewalls, etc, and is an excellent example of stacking relatively simple applications to achieve very complex results.

I'm not wedded to any code-base or application. I neither code for nor compile the GUI, I don't care if it's called Wayland, X, or Cheese. What I have seen is functionality that has been established through 30 years of people USING it, and any replacement SHOULD begin by implementing that same basic functionality, first.

For example, the change from Xfree86 to Xorg. I wasn't involved in how it was done, by fork or re-write or whatever. What I saw was that Xorg "picked up where Xfree86 left off". I consider that a good example.

If Wayland is better, and I have no reason to doubt it can be, then great. Wayland should start where X is now.

Maybe it will be Wayland with an X "plug in", this "coexist side by side" thing. Sure. Whatever it takes.
@Bob_R,

Rome wasn't built in a day

I believe that Wayland developers are trying to get all of X11 functionality in as quickly as possible by making Wayland & X11 coexist. I am pretty sure they will rip X11 apart eventually and either merge functions in Wayland/Weston or just make a miniaturized (Throw away useless code) X11 to coexist better. I believe the 2nd option will win because the developers are being very keen and careful about keeping Wayland as small as possible for better performance, maintainability and extensibility.

Post a comment

A free account is needed to post. Write your comment here first: your text is kept in this browser while you sign up or log in, and is waiting for you when you come back.

Your message:
|||||
10
Markdown: **bold**, *italic*, [text](https://link), - lists, > quotes, $x^2$ for math. Formatting help

Sign up and post (it is free)  or  Log in and post