How long will Xorg live?

Don't know anything about all this but saw this article today that seems linked to this conversation
To me this news just looks like one guy spreading a bunch of nonsense as part of an ego trip or popularity contest.

Rather than spending all his effort trying to undermine Xorg, he should just fix his code. Otherwise it can get in line alongside all the other incomplete ports to other platforms.

In almost the same sentence he mentions that Xorg is no longer developed for. But at the same time states that xf86-video-modesetting didn't get tear-free support implemented until earlier this year. Which one is it kiddo?

He then goes on some pedestal based tangent to say that the "modern" Apple GPU hardware is more suited to a "modern" display system from Apple which Wayland is more similar to than Xorg. However completely overlooks the fact that most Wayland "backends" are simply based on the Xorg modesetting "driver" anyway and interact with the hardware in an absolutely identical way via libdrm.

No doubt we will see people regurgitating this twits verbal diarrhea for years to come. Not impressed by him.
 
You didn't like my post then :-/
Haha. Nah, your post is fine (sharing links to "extreme" views on either side is very relevant), I was even just about to link to it myself. It is *his* post I am obviously quite passionate about ;)

By regurgitate I more mean someone confidently stating incorrect things without knowing anything about it themselves and just repeating what others have said. The classic one is "Xorg is insecure because it listens on TCP by default". The guys who repeat this are often completely ignorant of the fact it hasn't listened by default on a network socket for a decade (but they perhaps don't know what that means anyway, they just heard the FUD and ran with it by repeating it).

We will absolutely get some of those jumping on these forums asking why we still bother with Xorg when it has "all these issues as told by the fabulous Asahi Linux developer". While we are at it, they will tell us that BSD is dying, our init system is "retro" and that the BSD license is worse than being stabbed.
 
The lack of a display server is Wayland's strength. The reason Wayland was originally created was to compete with the Windows gaming performance. Windows apps write directly to display card memory while only the X server writes directly to display card memory. X apps must communicate with the X server using Xlib through UNIX domain sockets or internet sockets. This indirection, handing data to be displayed from one piece of software to another to another takes time an resources. Writing directly to video card RAM (one memory transfer) will always outperform even sending data to an X server through a UNIX domain socket (virtually a memory to memory transfer) so that the X server can transfer, again, to video RAM.

Remember X was designed when the network was the computer (a Scott McNeily statement), When large better performing UNIX machines would do the work while smaller slower desktop UNIX machines would handle display; X apps could run on remote machines and display on the local X display. (My laptop today has more power than the Sun Sparc 2000 I used to work on.)

Wayland has no such feature -- IMO a deficiency -- but then, that's its strength because playing games on Linux should be as responsive with Wayland as on Windows.

If one is a gamer -- I am not -- all this would make you happy. I run apps on my server machines in my basement, displaying to my laptop on the second floor.
 
Funny that you mention games. Wayland is performing significantly worse than X11 for non-native games on Linux (e.g. Proton emulation on Steam) and noticeably worse for native games using the NVidia drivers. That does not leave all that much of an advantage for Wayland at the moment.
 
The reason Wayland was originally created was to compete with the Windows gaming performance.

If one is a gamer -- I am not -- all this would make you happy. I run apps on my server machines in my basement, displaying to my laptop on the second floor.

Now I finally understand why this recurrent discussions about wayland and desktop, why people love these themes.
Since I am also not a gamer, I was not able to understand why all these insistent discussions --- by playing kids.

I think, cy@ said it clear: the goal of wayland is gaming, the goal of X11 is a display server. In spite of all
hardware development and its increased speed, I do think that to have a display server make sense.
 
An argument made was speed due to lack of network layers in wayland, As long as there is stuff like Node.js, Applications written using tons of "dynamic" frameworks and systems needing several GB/GHz/... to even boot I call BS on that. Someone had an ego trip to "make things better". I think he failed.
 
Oh, the usual "X vs Y" with both sides only pointing at opponents weak point and only praising own strong points. I must admit I never was unixy person WRT the desktop and never really had the need for running X networked, so I don't feel others loss here :D

What I am really wondering about, is wayland yet another "we will rewrite everything from scratch" (instead of improving the original)? Or was there a way to implement all that fancy (and required) desktop stuff in X (probably cutting down the non-essential parts via compile switches)?
 
If you look at the tests of phoronix, it seems to be only minimally worthwhile currently and in most games you probably wouldn't notice it without a comparison.
 
Hector Martin has had some posts on Mastodon about how the Xorg devs have basically moved on to Wayland because the X11 proto is too busted and they would have to break all backwards compatibility to add features people want. The backwards compatibility situation is XWayland, WayPipe, etc
 
Hector Martin has had some posts on Mastodon about how the Xord devs have basically moved on to Wayland because the X11 proto is too busted and they would have to break all backwards compatibility to add features people want. The backwards compatibility situation is XWayland, WayPipe, etc
OK, that answers my question above, thanks; wayland it is then.
 
Hector Martin has had some posts on Mastodon about how the Xord devs have basically moved on to Wayland because the X11 proto is too busted and they would have to break all backwards compatibility to add features people want. The backwards compatibility situation is XWayland, WayPipe, etc
Um, that's the same guy richardtoohey2 quoted above. I skimmed his Mastodon post, and find myself in vigorous agreement with what kpedersen said up thread.

Please, please stop using Xorg with Asahi Linux.
So some people are getting Xorg to work with your shiny "modern" project? Why are you trying to stop them?
Yes, not every random app and feature you use on Xorg will have a Wayland equivalent. Deal with it.
I have a different idea. How about I keep using Xorg and you go pound sand?
 
that's the same guy richardtoohey2 quoted above
Yeah, I didn't get an alert on this thread until today and I thought when I clicked it that I was seeing page 2, so, apologies.

I think the point of the posts is that for a new project, this is the train to get on. Nobody seems to be interested in X12. Will there be another split like xfree86/xorg. I doubt it. This seems like a "lost interest" sort of thing. Xorg gets updates, but it seems like it's not getting the same attention as Wayland?
 
I think the point of the posts is that for a new project, this is the train to get on. Nobody seems to be interested in X12. Will there be another split like xfree86/xorg. I doubt it. This seems like a "lost interest" sort of thing. Xorg gets updates, but it seems like it's not getting the same attention as Wayland?
Yeah, that's going to be interesting. Maybe Openbsd's Xenocara will become a real fork, and get more people working on it, as others have suggested. I can hope.
 
I suspect that FreeBSD could at some point develop their own version of Xorg (OpenBSD did). Our developers are resourceful. I'm not concerned about Xorg.
 
Or... FreeBSD doesn't swim upstream and just does work to make Wayland by default-able.

 
Back
Top