personally I have hope for this project, we'll see what happens though, canonical does however have the resources to do this...as they proved with Unity
personally I have hope for this project, we'll see what happens though, canonical does however have the resources to do this...as they proved with Unity
This was my own concern, more or less; though I don't know to what degree a differing display server affects programming. If Ubuntu is the Linux distro of choice for third party software developers, and if their software is developed specifically for MIR, then this would seem to complicate matters for other distros still using X or using Wayland. I could be wrong though, which is why I'm asking the question.
I freely admit that I'm no expert on these things, but can Unity (at the moment a layer built upon Compiz/Gnome 3) be compared to building a display server from scratch? I'd think building a display server is vastly more complicated than making a DE.
Correct me if I'm wrong...
The Roadmap they have is for it to be functional for 14.04.
;p
Last edited by mikodo; March 21st, 2013 at 09:49 PM. Reason: grammar
Unless I'm much mistaken, this isn't the case. Steam hasn't chosen Ubuntu. It's chosen Linux. It chose Ubuntu as its first approach to Linux, as the perception is it's the most popular Linux distro, but what Ubuntu does next will have zero effect on Steam I imagine. The Steam box isn't going to be Ubuntu-based (I don't know where this idea comes from, so please enlighten me if there's solid info I missed), but Linux based, and the client is Linux based. Of course what comes after X will affect how distros work with the Steam client on the desktop, but Steam in and of itself is not solely intended for Ubuntu.
My personal take is Ubuntu and Mir will pull further away from Linux as we know it. Whether this will result in the success they want is debatable. I suspect Wayland will be what the Linux eco-system uses, and Mir will be what Ubuntu uses, much like how things are right now with Unity. I don't think this will affect drivers from Nvidia etc, as you'll see the same driver across both approaches. From what I gather, it's not like how X was and shouldn't cause a divide from a driver point of view.
Intel Core i3-8100, 16GB Corsair DDR4 2400 RAM, 240GB Crucial SSD+1TB HDD, Nvidia GTX 1050 Ti
As I said in an earlier post, third party software developers are likely to use gtk, Qt or some other toolkit library which would have X/Wayland/MIR support. True, there are certain applications that do in fact draw directly to the display server such as Blender, and these applications will have to deal with the complications - but in most cases I don't think it would affect application developers directly.
The main issue is that the fragmentation is likely to result in us taking even longer to move away from X11 because the toolkit developers need to put in work to support multiple display servers. The one thing that might make everything work out is if Canonical pulls out amazing engineering chops and their Qt/Gtk/whatevertoolkit downstream patches work flawlessly.
Drivers do in fact need an interface to X. In mesa, that's the DDX. (check out: http://yangman.ca/blog/2009/10/linux...ack-explained/) The NVidia driver and AMD's Catalyst also install their own X drivers (http://wiki.debian.org/NvidiaGraphic...Installation-1). While I'm not intimately familiar with Wayland/Mir architecture, it is likely that these drivers would need something written to interface with Wayland/Mir.
But isn't that, in effect, "choosing" Ubuntu? Ubuntu users could quickly download Stream via the download manager. For a short time the first official release of Steam "on Linux" wasn't initially available to other distros unless you were knowledgeable enough to manually install it. If what JDShu writes is correct, then even with MIR it will be business as usual.
Interesting. So, the question becomes: Are these companies going to write drivers for X, Wayland and MIR (if a choice has to be made). They only just barely cobble together Linux drivers. Am I wrong in thinking this could turn into a Blu-ray versus HD DVD style "battle"? If so, I would expect Canonical's MIR to become the defacto standard. The other distros, even if they were to all go with Wayland, are utterly irrelevant outside the Linux-verse -- including RedHat.
Anything can happen, but prominent developers from within AMD and Red Hat claim that the majority of their customers that use Linux and propietary drivers are corporate workstations, implying that the desktop market - ie. us - are not targetted at all. This is the justification for keeping Catalyst around despite it being a buggy POS. The fact that it even works for us is basically luck. Ubuntu is only really a powerhouse on consumer desktop Linux, so if I was to bet money I would bet on whatever Red Hat and Novell (the main corporate Linux providers) choose.
Btw, I am unfamiliar with exactly how difficult a DDX or Wayland equivalent is to write. If it's not hard then maybe it doesn't matter on the driver side.
Bookmarks