X12...

[reaperx7]

A lot of people have been throwing X12 as a failure of Xorg to justify deprecating Xorg in favor of wayland, but... I think we need to think about this.

What if we could start efforts towards a real X12? A real successor to X11 but still X.

[xgui4]

i agree

[reaperx7]

I think what we should do is formulate an outline of what we want the system to be first and foremost.

My approach to this idea is simple yet based on the X11 design, somewhat, but also incorporating some ideas from wayland, but in a less intrusive manor...

The idea is to have a secure server layer operating in kernel space to manage drivers for video, inputs, and monitor EDID management, and a client layer operating in userspace managing and directing applications without interfering with applications loaded in parallel or singular sich as using the clipboard or file management. These drivers allow for vesa mode (software accelerated), and DRI3 (glamor accelerated) for video (vesa is the fallback for virtual machines and non-glamor GPUs). All input can be effectively managed by libinput with a configuration file to manage buttons and instances. Same for video.

The server layer loads during boot as a precursor service to be readied. Can be loaded either by a .service file for systemd or a rc.xserver script for sysvinit like systems, or a custom script for other init systems. This readies the system for the client either local or via SSH secure connection for remote.

A full set of protocols handle all aspects of server management, behaviors, and application behaviors. As well as a library set for usage by applications to link to for handling windows and other functions within the session.

The client has a basic stacked window management system similar as part of the system (bwm - basic window manager), but can be supplicanted by external window managers started using the client layer session manager such as Xfce, Mate, etc.

BWM (in concept theory) is akin to IceWM with a simple start menu, desktop selector, clock, toolbar, and service tray, and can serve as a fallback, diagnostic, or independent WM. Windows are stackable, but can be configured for tiling mode also.

This is just a rough draft at this point...

[b-aaz]

https://www.x.org/wiki/Development/X12/

[xgui4]

still, linux on mobile could be coming an option in the future ...

[reaperx7]

Yes, in future. I'm not closed off to the idea, but the foundation for where to begin needs to be based on PCs.

[metux]

By the way, mobile devices are exactly one of the core use cases for Xnamespace extension.
So, yes, mobile devices already are on the agenda :)

[kkofler]

still, linux on mobile could be coming an option in the future ...

Why only in the future? I have been using a PinePhone as my main phone and my only smartphone for 4 years!

[metux]

Maybe I've been a bit imprecise: with "mobile" I'm not just talking about a computer happing to be small enough to fit into your pocket. But also the applications being mobile. This implies you might have non-trustworthy clients connecting your display (which is quite rare on classic desktop/workstation).

So, the idea is using similar techniques like in data center world (eg. containers, automatic provisioning & scaling, automatic migration, etc, etc). And it's getting hard to really trust all individual applications. That's where Xnamespace comes in.

[darkhog]

I don't see how we can do that without breaking backwards compatibility, i.e. another Wayland situation. X11 Xserver can't run X10 or X9 apps or environments.

[reaperx7]

Compatibility layer through an API.

Technically, it should be invisible to the application anyway unless it directly calls to specific libraries, but even then a rebuild should be the ideal fix.

The goal for X12 should be to maintain compatibility with X11 without sacrificing function over form.

[dec05eba]

If you look at https://www.x.org/wiki/Development/X12/ pretty much all of that can be done while keeping X11 compatibility. Comparing it to X10 and earlier isn't the same, those earlier X were still relatively a prototype for X. X evolved to X11 in just 3 years.

[metux]

Typical objection, but yet have to see a single thing here that really requires breaking compatibility.

[ZyxerZyxani]

Ok I read that link and I wasn't aware of X12 specs even existing.
I suppose the main issue then is ppl thinking X12 is not defined and that it would be some big upgrade/change similar to microsoft windows, which breaks compatibility all the time and relies on compatibility layers for backwards compatibility.

[reaperx7]

It's very crude, but this is the concept idea... It borrows heavily from wayland, but places functions where they need to be in a form of sanity to resource management and system security. It gives more compromise to userland and applications without allowing access to critical levels of features. Access control is implemented through the session management systems, and the xclient works with existing applications and software without a need to aggressively use compositors to do things in a non-standard way.

[xgui4]

waylandisms

what is that new word ???? why are you soo anti-wayland ????? in my use case it work pretty great , but i will like x11 to exist

[reaperx7]

You lost me at "it borrows heavily from Wayland". Sorry, we don't need no stinky waylandisms.

The idea is to borrow a few things from wayland in the design, but not the code.

But instead of being just a protocol, X12 is a complete solution that reinvents X from the ground up. Instead of using a compatibility layer, xwayland, for use with applications, we use an API solution for X11 to maximize compatibility and translate X11 calls to X12.

[darkhog]

why are you soo anti-wayland ?????

Why are wayland devs so anti-x11?

[xgui4]

why are you soo anti-wayland ?????

Why are wayland devs so anti-x11?

why are xlibre dev so anti-wayland? i am not answering a weak whataboutism argument. i am not a wayland dev.

and sadly pladma 7 will be wayland only ... i guess waayland is future ... and i am really sad even as a new linux user, the only reason i use wayland is to hyprland being wayland only and touchpad gesture + mutli-monitor setup that doesn't work well with x11

[probonopd]

Wayland is not "the future". It's one option. Been working on X11 related code for a new desktop environment all day long. Was able to implement global shortcuts in no time, something not possible on Wayland. Also I wrote some code that depends on being able to move the window, also not possible in Wayland without writing or forking an entire compositor.

[callmetango]

@xgui4 Could you please create a new discussion regarding this issue at X11Libre Xlibre On Arch Linux · Discussions · GitHub? This discussion thread is about "What if we could start efforts towards a real X12? A real successor to X11 but still X."

[ZyxerZyxani]

No. X11 with modules and updates is the best way forward.
X stopped the releases for a reason. Or we are making the versioning number not mean anything, since new version should have new things not working on older version. Meaning we would at worst need a X11 comp layer for X12 or X12 will not mean anything at all.

We do not need a successor to X11. Explain why we need it. And no, being old and stable is not an argument against it. Mona Lisa is old and no one is gonna say "Let's replace Mona Lisa because that painting is old". X11 is a piece of art as well and works pretty much everywhere. And on old hardware. By doing too much upgrading and "getting rid of old" you are just supporting planned obsolescence. Absolutely, you can complain that my 30 year old hardware is old and that I should upgrade, but this just seems like a slippery slope to making newer and newer hardware incompatible.

I want to run my retro hardware with new software. And then I don't want X to lose so much "old and outdated code" that it won't even compile or run on 10 year old hardware. X11 is great and what makes it what it is still has a usecase and is important.