Sign in to follow this  
Followers 0
f0rk_b0mb

ubuntu Mir: Ubuntu's New Display Server

119 posts in this topic

Well at the end of the day, no-one is developing gnome2 anymore, it's like using windows 2000, if it works, it works, if it doesn't, it isn't getting patched up.

There's MATE but progress on that is very slow.

nope. gnome2 is nothing like win 2000. comparing a gui/shell with a whole OS is not serious at all. also why should be gnome 2 developped still? its highly customizable anyway, much more than gnome 3, gnome shell or unity.

Share this post


Link to post
Share on other sites

I'm talking about in terms of updates, if there's some huge security flaw discovered in windows 2000, it won't be patched.

If there's some huge security flaw in gnome2, it won't be patched (that's not to say there won't be A patch but nothing will be officially released by the gnome team).

I agree it is very customisable, and think even to this day it's an utter disgrace the gnome team removed loads of features from gnome2 in gnome3 such as the ability to specify your own volume step. Why should it be developed still? because it uses GTK2 and not GTK3, it's lacking a lot.

Share this post


Link to post
Share on other sites

nope. gnome2 is nothing like win 2000. comparing a gui/shell with a whole OS is not serious at all. also why should be gnome 2 developped still? its highly customizable anyway, much more than gnome 3, gnome shell or unity.

I'd actually argue that gnome 3 is more customizable than gnome 2, since its written in javascript and easily extensible.

Anyway, regarding MIR, after reading a lot about this situation I've become rather angry at Canonical's behaviour here.

Basically here's what happened:

Canonical "thoroughly researches" wayland to see if it will fit their needs. Only they failed to get even a basic understanding of how wayland works, and never once reached out or spoke to a single wayland developer. A canonical developer even admitted in the wayland IRC after mir was announced, that they didn't really understand how the wayland input model worked.

Then they start working on MIR in secret. They announce it, with a list of "technical reasons" why wayland is inferior and doesn't fit their needs. This list is very quickly debunked by wayland developers as basically being completely untrue FUD, but Canonical keeps truckin with their pointless new display server, which is a great example of this: http://en.m.wikipedi...n_of_commitment

If they had just come out and said "we made this because we want complete control over our display server", I still wouldn't be super happy about it, but I would understand their justification. Instead they clearly didn't even understand how wayland works, and tried to use a list of totally untrue reasons as to why wayland doesn't work for them as justification.

2 people like this

Share this post


Link to post
Share on other sites

Anyway, regarding MIR, after reading a lot about this situation I've become rather angry at Canonical's behaviour here.

Basically here's what happened:

Canonical "thoroughly researches" wayland to see if it will fit their needs. Only they failed to get even a basic understanding of how wayland works, and never once reached out or spoke to a single wayland developer. A canonical developer even admitted in the wayland IRC after mir was announced, that they didn't really understand how the wayland input model worked.

Then they start working on MIR in secret. They announce it, with a list of "technical reasons" why wayland is inferior and doesn't fit their needs. This list is very quickly debunked by wayland developers as basically being completely untrue FUD, but Canonical keeps truckin with their pointless new display server, which is a great example of this: http://en.m.wikipedi...n_of_commitment

If they had just come out and said "we made this because we want complete control over our display server", I still wouldn't be super happy about it, but I would understand their justification. Instead they clearly didn't even understand how wayland works, and tried to use a list of totally untrue reasons as to why wayland doesn't work for them as justification.

I agree, it sounds like Canonical disliked the idea of simply being a "contributor" so said "Screw it, we'll make our own display server, with Blackjack, and hookers".

1 person likes this

Share this post


Link to post
Share on other sites

I withdraw my previous comments, it's a shame that Ubuntu won't work with others considering how popular the distro is

Share this post


Link to post
Share on other sites

Haha brilliant;

'The Mir specification removed the Wayland criticism and added that Wayland's input event handling has indeed been improved compared to X, plus added some more jargon, and ultimately concluded with: "However, we still think that Wayland's attempt at standardizing the communication between clients and the display server component is very sensible and useful, but due to our different requirements we decided to go for the following architecture [with regard] to protocol-integration."'

I.E. canonical's staff know **** all. How idiotic they look now.

Share this post


Link to post
Share on other sites

There's a bright future for Mir, considering they've shown they don't even understand how Wayland works.

Share this post


Link to post
Share on other sites

Well, maybe there's something they're keeping from us? They said key parts will be kept secret, so maybe they have a killer feature up their sleve that Wayland doesn't do, or can do but, it's very slow and unstable? I'm not defending them, I'm just throwing out options.

Edit: I found a demo of Mir. http://www.omgubuntu...ts-demoed-video

Share this post


Link to post
Share on other sites

Well, maybe there's something they're keeping from us? They said key parts will be kept secret, so maybe they have a killer feature up their sleve that Wayland doesn't do, or can do but, it's very slow and unstable? I'm not defending them, I'm just throwing out options.

Edit: I found a demo of Mir. http://www.omgubuntu...ts-demoed-video

Or they got caught talking crap about stuff they didn't understand, and now they're trying to save their face with some more made up crap as they scramble to try and create some valid reason in a coding field they have no experience in and that's a very specialized coding field requiring a life time of experience.

Basically they got caught with their pants down as incompetent in this field and they don't want to admit it.

The whole "we have super secret reason" is so obvious you could spot it from mars without a telescope.

Share this post


Link to post
Share on other sites

This will turn into the modern day sound server fiasco.

In my opinion ubuntu would benefit more from contributing to Wayland and other projects rather than creating their own and trying to push it as mainstream.

Share this post


Link to post
Share on other sites

give it a few months and Ubuntu will say they need to create their own new sound server as well :)

Share this post


Link to post
Share on other sites

This is actually good news and I look for forward to the future. Honestly, I am really glad that Canonical is taking this step although a bit late, better than never.

Correct me if I am wrong for the following:

  • Microsoft created their own display server called "Windows" that ran on top of DOS and it was highly successful back in the days
  • Apple has their own display server for OS X running on top of Unix and is highly successful
  • Android has its own display server and is highly successful

Linux distributions (in general) have had X for ages, an ancient display server and is gone nowhere (for desktop computing).

It is about time someone took the matter in their own hands and did something. My respect for Canonical has skyrocketed. I want to see Canonical do what Apple did with their own display server. Keep up the good work!

Share this post


Link to post
Share on other sites
give it a few months and Ubuntu will say they need to create their own new sound server as well :)

Don't give them any ideas, too many as it is.

Share this post


Link to post
Share on other sites

This is actually good news and I look for forward to the future. Honestly, I am really glad that Canonical is taking this step although a bit late, better than never.

Correct me if I am wrong for the following:

  • Microsoft created their own display server called "Windows" that ran on top of DOS and it was highly successful back in the days
  • Apple has their own display server for OS X running on top of Unix and is highly successful
  • Android has its own display server and is highly successful

Linux distributions (in general) have had X for ages, an ancient display server and is gone nowhere (for desktop computing).

It is about time someone took the matter in their own hands and did something. My respect for Canonical has skyrocketed. I want to see Canonical do what Apple did with their own display server. Keep up the good work!

You are forgetting about wayland. You are correct, X is crap and badly needs to be replaced with a more modern display server. This display server has already been in development for years, its called wayland. Canonical first said they would adopt wayland, and then all the sudden announced that they were developing mir (which looks to very similarly architecturally to what wayland is already) behind closed doors. The technical reasons canonical listed as the reasoning that wayland wouldn't meet their needs have all been thoroughly debunked already, Mir is pointless duplication of effort and can potentially cause harm to the linux desktop due to how fragmentation will effect driver development from AMD/Nvidia.

If wayland didn't exist, and we were all still stuck with x for the foreseeable future, I would fully agree with you and be happy with canonical's decision, but as it stands canonical's decision is damaging and doesn't make any sense.

Even one of the mir developers can't even articulate why mir needs to exist: http://pastebin.com/KjRm3be1

The fact that canonical's developers seemed to totally lack any technical understanding of how wayland works does not inspire confidence in their own display server... The wayland developers are mostly former and current xorg developers and are much more experienced in this area.

Share this post


Link to post
Share on other sites

This is actually good news and I look for forward to the future. Honestly, I am really glad that Canonical is taking this step although a bit late, better than never.

Correct me if I am wrong for the following:

  • Microsoft created their own display server called "Windows" that ran on top of DOS and it was highly successful back in the days
  • Apple has their own display server for OS X running on top of Unix and is highly successful
  • Android has its own display server and is highly successful

Linux distributions (in general) have had X for ages, an ancient display server and is gone nowhere (for desktop computing).

It is about time someone took the matter in their own hands and did something. My respect for Canonical has skyrocketed. I want to see Canonical do what Apple did with their own display server. Keep up the good work!

I think you're confusing the OS with a display server, the display server is just an app that's responsible for showing window contents on the screen (So that apps don't have to do it themselves)

Canonical could write their own new display server from scratch, and you'd only notice if it wasn't working well (Which will probably happen), it's just like how Windows didn't radically change with the addition of the DWM.

Share this post


Link to post
Share on other sites

Hi :)

I do not believe I confuse anything and I made it very clear but here is a detailed version of my post:

  • Microsoft created their own display server called "Windows" that ran on top of DOS and it was highly successful back in the days

DOS == the OS, Windows == the GUI (display server)

  • Apple has their own display server for OS X running on top of Unix and is highly successful

OS X == Command line, GUI + Darwin where the GUI is running on top of the Unix/Darwin

  • Android has its own display server and is highly successful

Android == Linux + the Display Server

In relation to Android, I have no idea what display server is running and/or if it's related to X.

I think you're confusing the OS with a display server, the display server is just an app that's responsible for showing window contents on the screen (So that apps don't have to do it themselves)

Canonical could write their own new display server from scratch, and you'd only notice if it wasn't working well (Which will probably happen), it's just like how Windows didn't radically change with the addition of the DWM.

Share this post


Link to post
Share on other sites

Ubuntu, why no NO use Wayland? Building their own stuff will only break compatibility when wayland becomes mainstream.

what is this wayland you speak of and who said this will be mainstream ?

seriously asking I have no idea

thanks

Share this post


Link to post
Share on other sites

what is this wayland you speak of and who said this will be mainstream ?

seriously asking I have no idea

thanks

Wayland is the successor to X that is currently in development. It is being developed by many of the same developers that currently develop X. Its the upstream backed X replacement, its expected that most non-ubuntu distros will move to wayland in the future (and before ubuntu started this pointless project it was expected that ubuntu would move to it too). Here's a post I saw earlier that explains quickly how X works and why it needs to be replaced:

So the current heirarchy looks something like this (corrections welcome):

Display < GPU < Framebuffer < X < GL < Compositor < Toolkit < Application

The problem here is the X part is based upon IPC and windowing from the 70's where the Application/toolkit talks to X, X draws something (incorrectly), X talks to the Compositor, the Compositor redraws the something (correctly) and then informs X and then the Application something has been drawn. This is the way things have been done since Compiz and Metacity+Compositing came out. We're taking a huge IPC hit for no reason other than X is ubiquitous. That's like running a LAMP server as a VM inside Windows, and allowing Windows' networking stack to run the routing tables, just because Windows is ubiquitous.

The new heirarchy will look like this:

Modern:

Display < GPU < Framebuffer < GL < Compositor < Toolkit < Application

Legacy:

Display < GPU < Framebuffer < GL < Compositor < Toolkit < X < Application

Here, both Mir and Wayland are essentially the Compositor. Notice how we've actually removed an entire IPC step to get stuff onto the screen for everything except programs like xeyes? Notice how we still keep network transparency for X-based servers? That is why we would choose to use the new display managers.

Hi :)

I do not believe I confuse anything and I made it very clear but here is a detailed version of my post:

  • Microsoft created their own display server called "Windows" that ran on top of DOS and it was highly successful back in the days

DOS == the OS, Windows == the GUI (display server)

  • Apple has their own display server for OS X running on top of Unix and is highly successful

OS X == Command line, GUI + Darwin where the GUI is running on top of the Unix/Darwin

  • Android has its own display server and is highly successful

Android == Linux + the Display Server

In relation to Android, I have no idea what display server is running and/or if it's related to X.

The display server and the GUI aren't the same thing, the GUI stuff runs on top of the display server. The display server is what talks to the video hardware and in some cases it also handles the 3d compositing (which is the case for wayland and mir for example. X couldn't handle compositing on its own so it requires that you run a compositor on top of it which is rather inefficient, one of the many reasons it needs to be replaced)

2 people like this

Share this post


Link to post
Share on other sites

Edit: As confusing as it is, Wayland isn't the name of a display server, it's the name of the protocol for talking to one. Weston is the name of the reference display server that implements Wayland (And X11). Wayland is meant to be small enough that each DE would provide it's own display server, e.g. Kwin for KDE, Mutter for Gnome, etc. instead of relying on a single X11 server

Hi :)

I do not believe I confuse anything and I made it very clear but here is a detailed version of my post:

  • Microsoft created their own display server called "Windows" that ran on top of DOS and it was highly successful back in the days

DOS == the OS, Windows == the GUI (display server)

  • Apple has their own display server for OS X running on top of Unix and is highly successful

OS X == Command line, GUI + Darwin where the GUI is running on top of the Unix/Darwin

  • Android has its own display server and is highly successful

Android == Linux + the Display Server

In relation to Android, I have no idea what display server is running and/or if it's related to X.

Well, no. "Windows" isn't a display server, it's an OS that didn't even have a display server until Vista (The DWM), before that you had a mix of GDI and Direct3D/OpenGL that drew directly to the screen. Mac OS gained a display server with OS X called Quartz Compositor.

Share this post


Link to post
Share on other sites

No matter what they do with thier GUI it's still problematic distro and a cluster****. Do we really need another display manager? Nope.

Share this post


Link to post
Share on other sites

Why Your Desktop Won't Be Running Mir/Wayland Anytime Soon -- skip the blog entry, read A. Seigo's comment.

Yeah, Aaron's comment is a very good understanding of the situation. Thats right on par with what I expected, X definitely isn't completely going away anytime soon, we will still need xwayland (and mir's equivalent) for a long time. The good thing is that even wayland/mir + the nested rootless X server legacy solution is still much better than what we have now with just X. Wayland/mir would be the 'system compositor" and all the compositing stuff would still be going through wayland, so we'd still get wayland's advantages such as no more video tearing/more direct access to the hardware when it comes to compositing/reduced IPC, much reduced flickering and other graphics glitches :)

It would be: Display < GPU < Framebuffer < GL < Compositor < Toolkit < X < Application vs the old: Display < GPU < Framebuffer < X < GL < Compositor < Toolkit < Application. Both more efficient and keeping excellent backwards compatibility with X.

I'm hoping within the next 1-2 years we start seeing some distros adopt wayland in this manner (or at least have it so its easy for early-adopters to test). It depends heavily on driver support though, since ATM wayland needs a driver that supports EGL, and afiak right now thats only the open source drivers. Since both wayland and mir require EGL, and X/GLX will eventually be considered legacy, I expect Nvidia/AMD to provide an EGL driver eventually.

Share this post


Link to post
Share on other sites

Because it's linux, and fragmenting is better than cooperating for a common goal.

this s seriously the main downfall of linux, and what keeps tripping it whenever it starts to get a good run.

LinuxHater has a nice post regarding this situation: http://linuxhaters.b...gmentation.html (NSFW)

Share this post


Link to post
Share on other sites

LinuxHater has a nice post regarding this situation: http://linuxhaters.b...gmentation.html (NSFW)

After reading the post you linked I think the author came to the right conclusion for the wrong reasons. He essentially claims that fragmentation of the GNU/Linux ecosystem isn't a problem because it is already heavily fragmented and the marketshare is too small for anybody to care. Although the marketshare of desktop Linux is small, there are many people who care about it. In fact, the open-source nature of the platform encourages active participation of users in their favorite software projects. Unlike fragmentation of proprietary platforms, there are not permanent inherent compatibility problems when open-source software fragments. There will simply be intense debate for several years over which platform is best for accomplishing the task. All competing platforms will take ideas (and probably code) from the others until one emerges as clearly superior. The winner will become the platform upon which everything will be standardized. Fragmentation encourages choice, competition, and innovation; it is one of the major advantages of GNU/Linux over other platforms.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0

  • Recently Browsing   0 members

    No registered users viewing this page.