How Important Is The Wayland Display Server?

[mirza]

X rant

Remote X is useful when ever you want to run the application in remote machine.

Every OS, including Windows can access remote desktops and I think users are quite happy with how it works (at least from what I heard so far). On desktop, it is *very* rarely used feature so Apple and MS created architecture that doesn't really care about forwarding graphics over TCP/IP and god, that was really good idea. They desinged it with Joe Sixpack in mind: tearing-free video, overall experience (resizing window instant and nice, without gray artifacts). Also, codesize and memory usage is always issue. If Wayland can run my Gnome desktop with much smaller memory footprint, codesize and startup time ... I will use VNC for remote accessing other computers, good bye X!

I am using Ubuntu right now, another user also logged in different session but overall only one application is currently running (Chrome web browser). Xorg is running 2 processes "Xorg" 82+61 MB of virtual memory, plus x-session-manager process (2x25Mb). Total = 200MB for X, give or take, and I am NOT even running Compiz (guess why?). Thats Insane. Let's continue, Alt-Tab, switch between 2 windows repeatedly... LOTS of gray artifacts. No, this CAN'T be ment seriously for desktop usage, let alone embedded. I am all for Wayland and I never saw it running. Hows that possible? Because I am running X every day and know all about that.

[RealNC]

"Remote X" is hopelessly outdated by now anyway. Virtually everyone uses VNC or something NX-based. Plain old X over the network is pure suckage today.

[josvazg]

X.org is dead, long live X.org?

I think its quite obvious that Allen Lowe complain's are reasonable. Why could't the X.org project work as well as, for instance, the kernel project? I don't think the reason is that the developers are not as good as in other FOSS projects, but that X legacy makes their job TOO difficult and releases get TOO slow. There might also be project organisation problems, but I think he weight less than the technical ones due to legacy.

I believe this will not need to last much longer. Both GTK and QT/KDE are able to run on Windows (and MAC OS X?) so they already provide some translation libs that DO NOT use the outdated X lib. So it is possible to migrate 90% of apps (100% of the real new ones) to a NON Xlib infrastructure ON LINUX just by replacing GTK and GNOME/QT lower libraries.

This can be done for X.org, for Wayland AND for whatever Google OS writes instead of X.org.

Remote X on 90% of the devices is useless and in the other 10% is less efficient that plain VNC and other alternatives. I think MAC OS and Windows show the way to go here.

Google Chrome OS WILL BE COMPLETELY OPENSOURCE, so if they do something good enought there is nothing to prevent the FOSS comunity to use it or fork it into the next generation Linux Graphics stack. And keep in mind they will have to be able to RUN things like OpenOffice and other regular Gnome or KDE Linux Apps (otherwise they will not meet the date to market NOR will provide a minimal feature set that will be able to attract anyone to the new OS)

As of Gallium3D... I hope to be wrong on his one, but usually in FOSS when you keep on hearing that something will be great for too long BUT you never see it working anywere, you end up stopping hearing about it at all after a few more years, because it just disappeared.

Good FOSS improvements and incremental, simple and done relatively quickly. Like Git replacing BitKeeper, XGL/AIGLX and the Compiz stack, etc.

[beniwtv]

"Remote X" is hopelessly outdated by now anyway. Virtually everyone uses VNC or something NX-based. Plain old X over the network is pure suckage today.

Do you realize that NX uses the same "remote" X protocol?

That being said, VNC is slow as hell. It just doesn't cut it anymore after you have used NX. So until Wayland supports something like NX, I'm not interested. Just think about all those thin clients...

[MostAwesomeDude]

Hey, guys, when I said that we should stop running around in circles with bloody spatter flying from the stumps where our necks used to be, I didn't imply that you should just spin in place instead like little beheaded chicken dervishes.

X uses local sockets when possible instead of going through the loopback network interface. You don't lose any speed or gain any latency that way. It's not a bad thing, having a protocol that can instantly go out on the wire instead of always using IPC. In fact, I'd say it's the exact opposite of a bad thing.

VNC can't possibly beat X in speed because it has to transfer more data, excepting certain pathological cases in certain extension libraries. (Oversized display lists in GLX are one thing that immediately comes to mind, for example.) NX is faster because it compresses certain parts of the wire protocol, and caches others, but it's not always faster.

The X server manages memory for all the client processes' pixmaps, which makes it appear to take up lots of memory. A tool like xrestop can properly distinguish Xserver's internal memory usage from the memory belonging to other programs. Similarly, applications that request lots of rendering cause Xorg to suck up CPU cycles.

Gallium is moving at a disturbingly fast pace. I'm sorry that we're not moving faster; maybe if some more people joined in, we could go at that "speed of sound" pace you're expecting. Or maybe we could adopt the Vista release cycle, and wait a few more years?

APIs don't compete for the hardware. This isn't 1990; we have the DRM in-kernel and DRM arbitrates hardware access. Not that hard to grasp.

X12 is not theoretical; there are a lot of very real problems with X11. However, X11's pretty damn good, considering we haven't had to change it in a couple decades. The big overview list is http://www.x.org/wiki/Development/X12 ; if you don't understand it, don't feel bad. Or feel bad, whatever.

[R3MF]

"how important is wayland?"

he who code wins.

i'm not exactly hoping to replace X.org, but if X.org cannot keep up with the pace of development then i won't cry when distro's make the hop.

[rohcQaH]

The big overview list is http://www.x.org/wiki/Development/X12

That was actually the most interesting piece of information in this thread. Thanks.

[Kjella]

Hey, guys, when I said that we should stop running around in circles with bloody spatter flying from the stumps where our necks used to be, I didn't imply that you should just spin in place instead like little beheaded chicken dervishes.

Well, this is starting to turn into one of those open source sessions I hate where it's "pin the tail on the donkey" because you can't even figure out which module to blame. So let's say I get one of these triple-head cards AMD has been showing off.

I start with two hardware accelerated composited desktops.
I launch a 3D accelerated game in one.
I launch a 1080p tearfree video in the other.
I plug in a third screen (hotswap) to get a third desktop.

Bonus credit:
Let me plug in another graphics card and run CF/SLI.
Make that video hardware accelerated on shaders or fixed function.
Let me move all those to different screens.
...on different graphics cards.

Now is any of this at all related to X? X drivers? Mesa? DRI? DRM? Or maybe I'm going about it the wrong way, if you were given a blank sheet and those requirements, would you create X? Or what would the ideal graphics stack look like? I'm struggling to get the picture of what does what, and what should be doing what...

[89c51]

@MostAwesomeDude

will the new implementation of X (X12) make the codebase cleaner, smaller, easier to comprehend and easier for new people to jump in the development process??

[RealNC]

@MostAwesomeDude

will X12 provide an API as easy to code for as OS X Cocoa? Last time I looked into the X API, I couldn't eat for 2 days