Forgot your password?
typodupeerror
Ubuntu

Ubuntu's Mir Gets Delayed Again 241

Posted by samzenpus
from the just-a-little-longer dept.
jones_supa writes "Delays keep piling up for the Mir display server on the Ubuntu desktop. After already being postponed multiple times, Mir might not be enabled by default on the Ubuntu Linux desktop until the 16.04 LTS release — in two years time! This was the estimate by Mark Shuttleworth in a virtual Ubuntu Developer Summit. Using Mir, Mark says, will lead to supporting more hardware, obtaining better performance, and 'do some great things' with the technology. He expects some users will start using Mir on the desktop over the next year. Mir is already packaged as an experimental option, along with an experimental Unity 8 desktop session."
This discussion has been archived. No new comments can be posted.

Ubuntu's Mir Gets Delayed Again

Comments Filter:
  • tablet and phone (Score:2, Informative)

    by Anonymous Coward on Friday March 14, 2014 @12:33AM (#46480185)

    It's already default on the tablet and phone, which is what Shuttleworth is excited about these days. So in that sense, it is already here.

    So wayland is going to have to do a lot more than make decent ground if Ubuntu is to drop Mir. Wayland will need to do everything that Mir and X11 can do, and exceed them, and also be on a mature and well tested code base. Merely being an adequate competitor won't cut it.

  • by jones_supa (887896) on Friday March 14, 2014 @03:13AM (#46480637)
    Much better performance, no tearing problems, smooth compositing and desktop effects, old legacy X11 crap thrown away.
  • by Anonymous Coward on Friday March 14, 2014 @04:20AM (#46480845)

    network transparency

    Which is of course why X gets used where it is used in the first place. Wrap your head around X being used on MS Windows guys - once you work out why it is used there you'll stop laughing at network transparency and see why people are using it.

    X is only network transparent if all your apps are from 1995 and are written against Motif. Everything newer than that is not network transparent, it's just shoving uncompressed bitmaps across the network in a highly inefficient wrapper protocol that makes large numbers of inefficient, lag inducing round-trips.

    A rootless VNC-esque protocol (Xpra) is a superior solution in every way.

  • by thegarbz (1787294) on Friday March 14, 2014 @06:37AM (#46481277)

    If you think that he's wrong maybe you should look at how any modern X system works. Both X developers and Wayland developers have discussed in detail that there is nothing network transparent about any modern release of X which does any kind of direct rendering or hardware acceleration, something that was introduced around the mid 90s, so the parent's comment is actually right on the mark.

  • by Bengie (1121981) on Friday March 14, 2014 @07:13AM (#46481403)
    Wayland is a simple abstract layer. You can make your own remote protocol and have receive your draw commands and render it locally. One can fully implement X in Wayland, but one cannot implement Wayland in X. Therefore, Wayland is better. If you need X like functionality, someone could easily create a wrapper, which they already have for X.
  • by squiggleslash (241428) on Friday March 14, 2014 @07:40AM (#46481569) Homepage Journal

    That's complete claptrap. Yes, very recent (last 5-10 years) widget toolkits have started to force use of features that result in bitmaps being sent across the network, but that's hardly a reason to throw out X. And it's essentially a lie to pretend it means, somehow, that X11 doesn't have network transparency.

    I find it pretty bad, to be honest, that the same devs who are protesting that X11's network transparency isn't what it could be are:

    1. The devs that did this in the first place, refusing to advance the protocol to include the features that GTK3 et al required.
    2. Apparently think the solution is getting rid of network transparency altogether.

    I'm staggered, to be honest, by the whole thing. I appreciate old code is sometimes awkward to support, but the solution isn't a wholesale replacement of the project. Mozilla's developers may have made many mistakes in their decision to throw out the Netscape code that delayed the release of a great browser for many years, but they were NEVER, NEVER so stupid as to say "Well, we looked at the Netscape source code, and we think the solution is to get rid of HTML. It's too kludgy! I mean, just look at it, it's impossible to add features to it cleanly!"

    If we were talking about a rewrite of X.org, nee XFree86.org, nee X86, that'd be one thing. It's probably necessary by now although I'd still say they need to seriously think in terms of refactoring the project. But throwing out the entire protocol because they refused to add the features necessary to make the protocol efficient? Fuck that.

  • by Anonymous Coward on Friday March 14, 2014 @09:12AM (#46482305)

    Have you tried using Xpra [xpra.org]?. It does exactly that with X. It works great over the internet, and lets you attach and detach X applications across X servers. Basically, it's screen for X.

The most exciting phrase to hear in science, the one that heralds new discoveries, is not "Eureka!" (I found it!) but "That's funny ..." -- Isaac Asimov

Working...