Red Hat Expecting X.Org To "Go Into Hard Maintenance Mode Fairly Quickly"

[kgonzales]

I despise a display protocol which is not network-transparent. It's a huge step backward.
Also, I want it to have built-in support for HiDPi displays (support for content resizing done properly).
And support for screen grabbing. This one should be easy to add. Except when developers are clumsy.
But the first two things are not happening.

Looks like its time for you to step up and support the code base of your choice, or to pay someone to support that code base on your behalf.

[smitty3268]

There is nothing diminishing user experience in a network-tranparent protocol.

The point is, it's reasonable to say you need to be able to remotely run applications from other machines. What's not reasonabe is saying it has to be done with a particular technology - one that time has proven to be quite unoptimized, at that. The user experience is what should matter, not how it's accomplished.

It's like refusing to run a desktop based on the fact that it's written in C instead of C++, or C++ instead of C. You should just base your decision on whether the desktop works well.

[Paul Frederick]

The basic premise of the FLOSS ecosystem is that once the code is open, people can step up and integrate/build upon what is already there to the benefit of everyone. If NVidia doesn't want to take advantage of that because they feel they can gain a competitive advantage by keeping their code proprietary, that's their prerogative.

Now, if NVidia wants to support Linux and the BSDs with a closed driver, it only seems fair that they carry the burden of making sure it works.

As an aside, I decided to stop supporting NVidia a long time ago and now my money benefit AMD instead due to their willingness to engage with the FLOSS ecosystem.

Has it ever crossed your pea brain that maybe Nvidia licenses technology and it not at liberty to share what isn't theirs?

[areilly]

The point is, it's reasonable to say you need to be able to remotely run applications from other machines. What's not reasonabe is saying it has to be done with a particular technology - one that time has proven to be quite unoptimized, at that. The user experience is what should matter, not how it's accomplished.

It's like refusing to run a desktop based on the fact that it's written in C instead of C++, or C++ instead of C. You should just base your decision on whether the desktop works well.

The funny thing is that there is a networked application interface protocol that renders locally while operating remotely. It's defined in terms of its protocols, not it's implementations. It operates in a space somewhere between the "raw text and shapes" of X11 and the "push the UI layer to the client" of Java's original UI use-case. It is, of course, web tech. Yes, there's a lot that's more awkward than the old-school X11 approach. I haven't come across any mechanism to allow me to trigger a browser to open on an SSH-remoted HTTP socket, yet, but surely that can't be far off.

[finalzone]

Network-transparent display protocol also stabilizes graphics APIs, so lots of benefits for other developers & maintenance.

It is just a superior technology, from a theoretical point of view. The problem is, noone had, so-far, been able to properly design and implement such a protocol. Because it is hard.

The implementation of network transparent for X server was good on its era. In the age of security vulnerability, the current iteration is prone of exploit as no safety mechanism were implemented due to the complexity of codes from even Xorg developers.

In reality, network transparency on xserver is fundamentally broken beyond repair for very long time. For such purpose, toolkit from each desktop environment with a common protocol from say freedesktop needs to be set with security in mind. Both spice and vnc are among the most used as seen on Gnome Boxes as an example.

Let remind Wayland was intentionally made to be minimal as possible.

[R41N3R]

don't see how this discussion is in any way related to the 32 bit abandoning of Canonical. If there would be no multilib suddenly thousands of games wouldn't work anymore, so this was really stupid! With xorg we are just now at a point where the main maintainers will switch focus, this is natural and they overlapped with the Wayland developers before. So xorg is in maintenance mode, nothing new or unexpected. Wayland together with xwayland is running all applications expect some special cases where the application developer resisted so far to spent effort. So the switch to Wayland can be done by every user when he feels confident and where it is seemless.
I did that step with KDE plasma on Wayland on all my systems long time ago, not everything is perfect yet, but it wasn't with xorg either, but my use cases are covered including gaming. What is missing on the KDE side is just some focus and attention to the detail on Wayland, it is overdue but I have the feeling that if they mark one of the next releases as the switch to Wayland, it could be done!
My only worry with xorg being less maintained would be xwayland getting no release, hope there is one soon.

[cynic]

Another RH forced change when the replacement is half-baked. X+KDE is rock-solid on Fedora so long as you're not using one of AMD's broken driver forks. Already several apps choke on Wayland, including NX clients or X2Go. Reinventing the wheel and forcing everybody to use hovercraft when hovercrafts are experimental is almost like forcing SystemD on a world when it's full of bugs and untested.

"RH forced"?

Nobody is stopping other to take care of X and maintain it, if they want.
RH is not forcing anyone to use wayland either. It's open source, remeber?

They are just making choices they think are the best for they customers, and that's all.

If you want to keep X alive, just do it by yourself, RH will have nothing to argue about.

[bogdanbiv]

So far the dominant discussion that I see is around people complaining about the state of affairs, claiming that RedHat pushes Wayland to people. Here's how I see things.

RedHat is only responsible to their paying customers and then (perhaps) Fedora users. They only said they will stop paying for patches to X, it's their right to choose where their contributions go. They aren't forcing anything on Debian, Devuan, Arch... even more so on AMD, NVidia, KDE/Plasma, Gnome, XFCE.
As an enduser, even if one is using RedHat 8 / Fedora, they can still hold out the change from X. And then there is porting and forking - check this port of KDE3, called Trinity Desktop (screenshots) or Devuan (related to SystemD).

I would love to see all people complaining about this state of affairs make contributions to either fix the experience on Wayland or step up as X contributors. X won't die if there are enough people developing it. Then again I'd like to point out the mess that is the Python 2 to 3 transition - 10 years later - people are still dodging it and would still hold out on it (my assumption) if someone would still pay for releases on 2.7. People would still release software only for 2.7 if it wasn't a hard deadline for it in 2020 (again, my assumption)...

Ok, KDE, Gnome, others may not be ready now for the transition, if you say so, but they'd better be ready by the time distros decide they should make the jump. I guess this is the nudge / friendly reminder for users and contributors to make their choice and to be happy with it.

Please, if you really care, vote with your wallet or feet, not just on a forum. NB: by voting with one's feet I mean to say make your bug reports and patches, by voting with one's wallet I mean pay some one to do it in your place.

-----
TL DR; I just don't understand how even with the openness in our free software culture people still find it easier to complain about the state of affairs instead of contributing their own vision to projects or forking their own.

[Mike Frett]

They better solve any issues before forcing people from X Org.

[Guest]

Time for you to step up and start support Xorg code directly, or pay for people to support Xorg on your behalf. If you won't do either, stop your bitching.

In that case, are there any developers willing to get crowdfunded to maintain X.org?