X is certainly overdue for replacement, and I can easily believe that the price it paid for high cross platform portability has exceeded the benefits.
However, relying as much on Linux specific features as the article describes strikes me as a step too far in the other direction. Are the benefits from these kernel specific features really awesome enough to outweigh the benefits of being portable to at least, say, the *BSDs, and maybe Mac OS X and Solaris?
Linux does not have a huge desktop presence. It has, of course, a huge cell phone presence, but Android does not seem likely to adopt a new window system.
It sounds to me like they wouldn't be against accepting patches that makes Wayland work on other platforms, just that they aren't going to do that work themselves. At least that's what I get from this quote,
> It's certainly possible to port Wayland to other operating systems, but they'll have to provide the same level of infrastructure as Linux does.
> It sounds to me like they wouldn't be against accepting patches that makes Wayland work on other platforms, just that they aren't going to do that work themselves. At least that's what I get from this quote,
As far as I can tell, they are not willing to add abstraction layers to accommodate other operating systems. (Or even GFX drivers that don't implement the KMS interface). If you wanted to port Wayland to BSD or OSX, you'd effectively have to implement most of the modern Linux graphics stack on them.
The way I read it is that they want other operating systems to adopt the same infrastructure, e.g. kernel mode setting, instead of implementing an abstraction layer in Wayland itself.
> But I guess they'll be forced to add some layers sooner or later. Having too few of those can be just as bad as having too many.
The main thesis of Wayland is that they won't add abstraction layers. If you want to support the Wayland protocol on another os without Linux facilities, or if there is new hardware or changes in the drivers that don't fit the Wayland model, Wayland will not expand to those needs, but instead a new display server should be built to serve them.
Small projects to specifically fit the needs of the users, instead of one large one that expands to serve all badly.
> Small projects to specifically fit the needs of the users, instead of one large one that expands to serve all badly.
For something at the bottom doesn't the opposite principal apply(not fully/all bad but to a certain extent)? Isn't this why linux is popular for so many embedded device instead of some proprietary embeded os for instance.
I cannot stand reading anything on Phoronix. The person who writes those articles really, really grates me. It seems all finger pointy and blamey all the time. I've not seen too much constructive or useful information come out of that site in a long, long time.
And as for Wayland. I keep coming back to one thing for it...input. I don't know if anyone is going to pick that up. Seems like the weakest bit of X and someone will need to redo quite a bit of that work for Wayland and I remember when X was struggling to get people to work on that...can Wayland get someone to work on some of the most boring bits ever?
Maybe I'd care more about the language if this was just some random blog... But these guys are a bit different. There's always some data and process explanation behind each of their posts. They might not be great programmers and OS contributors, but they're doing something really important anyway and are doing it in a pretty transparent way. There were many situations where I learned something from Phoronix or found a performance-related answer there. They're also a bit different from other blogs in that they follow up on a story. While others report "this is broken, whaaa!" and find a new topic the next day, Phoronix will write up posts about ongoing progress to fix the thing (typically).
For the effort they put into it, I can completely ignore the language they're using. And if they blame a specific thing/group... as long as they're correct - I don't mind :)
The only time I've ever seen Phoronix is either faulty benchmarks comparing Windows and Linux or for "Steam is coming to Linux! No it's not. YES IT IS! Wait no, I guess it's not... OMG IS THIS A LINUX STEAM CLIENT?!"
It's a new display server protocol that uses EGL and a whole lot of the modern parts of the Linux graphics system, while throwing away most of the old parts nobody uses. The reference compositor Weston can be used either as a replacement for Xorg, or run in Xorg. In it, you can even run a rootless Xorg server so you can run native Wayland apps alongside X apps (a bit like Mac OS X does).
This was in the article, and summarizes it reasonably well.
"Wayland is a new window system architecture aiming to be a good fit for everything from embedded and mobile devices to full-blown desktop environments. Wayland builds on most of the graphics driver, desktop and UI infrastructure we have today, but distils out just the display server functionality we actually use today. The toolkits, device drivers, compositors and desktop environments we use today all play their parts, but the X server is essentially reduced to an awkward, 500kloc IPC mechanism."
From what I understand it's a replacement for X, or rather for the X display protocol. It's rewritten from scratch to take advantage of modern hardware with GPU acceleration.
From what I understand it's a replacement for X.Org also providing compositing and such things. It tries to provide a clean codebase without being quite a monolithic as X.Org is right now.
It's currently quite immature, you can only really use demo apps on it, and while gtk and such have been ported I know of no real window managers that support it.
Anyone know if wayland will support remote "clients" and local "servers" the way X does today? What about xdmcp? Losing either would be quite disruptive for me....
There is currently no network transparency (thus no xdmcp either) in Wayland. Although the primary developer has left the possibility open that in the future some remote display support could be added.
Of course it is something that could be shimmed into another part of the stack by (NX, VNC, etc) as it has been on Windows for instance.
Until either of those happen, X will still happily coexist with Wayland.
While wayland-native applications would not be network transparent (as wayland does not support this), you would still be able to run X11 applications over the network. You can run X11 applications in Wayland (like OSX) so you can also run remote X11 applications in Wayland (like OSX).
Has anyone tried running wayland as their desktop? With features like GTK+ and clutter it looks like they're targeting gnome 3 - is it possible to get a build of it up and running or is anyone providing a live cd?
You can if you like running all your applications from a terminal... There aren't any desktop shells that support it apart from a really simple demo one.
After watching the video, being able to rotate my windows N degrees as the main showoff feature does not seem like a big selling point to me. Maybe I'm old school, but sideways and upside down windows seems like a bit of a waste of time to lose the stability, drivers and network transparency of a working window system.
Yeah, it can be hard to explain the value of Wayland given that it has similar features as Xorg/Compiz but implemented without kludges.
I think Mike Paquette's classic post is relevant here (since Wayland is inspired by Quartz Compositor): http://developers.slashdot.org/comments.pl?sid=75257&cid... "There doesn't appear to be much code left from the original X server in the drawing path or windowing machinery, and it doesn't appear that apps relying on these extensions can work with any other X server. Just what did we gain from this?"
However, relying as much on Linux specific features as the article describes strikes me as a step too far in the other direction. Are the benefits from these kernel specific features really awesome enough to outweigh the benefits of being portable to at least, say, the *BSDs, and maybe Mac OS X and Solaris?
Linux does not have a huge desktop presence. It has, of course, a huge cell phone presence, but Android does not seem likely to adopt a new window system.
[Disclaimer: I work for a non-Linux Unix vendor]