Xlibre fork lights a fire under long-dormant X.org development

Considerable new activity is happening both in the established X.org X11 server and around its new fork, Xlibre. Suddenly, over in X11 land, everything seems to be kicking off. As we reported last week, there is a fresh fork of the X.org server called Xlibre, which seems to have stirred up lots of activity. The project was …

[David 132]

X11 over network connection

"The Reg FOSS desk has been installing and supporting Unix boxes since 1988 and has never once needed or used X11 over a network connection."

Until about 6 months ago I'd have been in agreement, but now I find myself in a situation where I need to run some quite complex design software remotely on a cluster in a datacentre & control it from my Windows client machine. Colleagues use various VNC-alike solutions; I run an X11 server on Windows and forward everything, and I have to say it works brilliantly. Far snappier than VNC and allows merging/management of the X application windows alongside the MS Windows ones.

"... but along with blocking his account, as Phoronix reports, the X.org team has been busily reverting lots of his merge requests."

Which seems, on the face of it, petty. If the merge requests were initially accepted, why revert them now? Smacks of Stalinesque "un-personing"; are they also using a scalpel and airbrushGimp to erase him from all photos?

Ugh. Politics.

Anyway. Thanks for a great article, Liam!

[Chris Gray 1]

Re: X11 over network connection

Ah yes. Long, long ago I took a somewhat-longer-than-usual Christmas vacation with my parents. They of course had a windows computer, not a proper Linux computer. My email then was on my own domain on my home computer (gee, just like now!), and I didn't attempt to allow email access remotely (POP/IMAP). So, I exported my usual session over X. I put a Linux distribution on a USB stick and booted the parent's machine from that. Slow, sure, but access to the X-session at home was all I needed it for. Made my stay much pleasanter. And yes, I only did it a couple of hours a day.

[Alan Brown]

Re: X11 over network connection

"Which seems, on the face of it, petty"

They would be if they weren't causing issues. A lot of his stuff had been handwaved through and upon closer inspection turned out to cause more problems than they solved

I'm not overly defending X11 - it's a shocking mess that needs a ground-up rewrite - but Xlibre is not that rewrite and I don't believe that the fork will last long, simply because of the personalities involved

It's not as if X11 hasn't been forked in the past. Xorg was a breakway from XFree86, which in turn was a breakaway from OSI, etc

The reason that the Xorg devs have been sitting on things for quite a while comes back to the "shocking mess". It's highly doubtful we'll ever have a X12 (I can remember X10) and there are so many increasingly insurmountable security issues with X11 that it was generally considered best to sunset it - this is the primary reason most of the Xorg devs jumped to Wayland when given the opportunity

All that said, given a choice between X and RDP/VNC, I'll take X every day of the week. Those who criticise X on lower bandwidth links haven't had to endure the others.

[Gene Cash]

Re: X11 over network connection

Hell, my very first UNIX experience was on a remote Sun pizzabox, back in 1985. I used a pirated Windows X11 server because they were like $700 at the time and I was an extremely poor student. (is there any other kind?)

(Strictly speaking, that box was for graduate students, and I wasn't really supposed to be on it, but I was friends with the admin, who was a HUGE Dr. Who fan[1])

I have had to maintain boxes located behind other jumpboxes. It's also useful for maintaining the Raspberry Pi in my garage, where I can display its xosview on my local desktop.

A couple weeks ago, I showed the remote xosview trick to a couple of the DBAs at work. They think I'm some sort of genius and I'm going to let them keep their illusions.

[1] Knock, knock. Who's there? The Doctor. Doctor W... ARRRGH (was his sort of humour)

[Short Fat Bald Hairy Man]

Re: X11 over network connection

Nothing as fancy for me but my desktop has some mysterious (a.k.a. could not be diagnosed) problem where it was randomly rebooting.

The only solution I found was to disconnect keyboard, mouse and other external devices (speaker etc.), keep only the network on and use X over the network. Works fine as of now.

[Arthur the cat]

Re: X11 over network connection

"The Reg FOSS desk has been installing and supporting Unix boxes since 1988 and has never once needed or used X11 over a network connection."

For the first 60 years of my life I'd never once needed or used an ambulance. Doesn't mean I was stupid enough to suggest they weren't needed by others.

[Liam Proven]

Re: X11 over network connection

> For the first 60 years of my life I'd never once needed or used an ambulance. Doesn't mean I was stupid enough to suggest they weren't needed by others.

Good for you. My first time on life-support, I was 27.

Anyway, I can't tell for sure if you are questioning my position or endorsing it.

I strongly agree with you: I do not personally want or use this, but I know that it is absolutely critical to lots of people and as such I want to see it remain.

Wayland offers me nothing I want except for one (1) feature: different (and non-integer) scaling factors on different monitors. That is the _only_ thing it does that I want, and if the price of admission is using GNOME or KDE Plasma then I don't want in thanks. I will stay outside.

The decade-old 2nd-hand iMac I am typing on has matching Apple 27" displays, but one has 4x as many pixels as the other. That is a prime reason I don't run Linux on it, even as a spare fallback OS.

(The others are: that it's a Mac and I bought it to run macOS, and secondly, that Linux can't read APFS disks, let alone write them, so it's not much use as an emergency fallback.)

[Zolko]

Re: X11 over network connection

The Reg FOSS desk has been installing and supporting Unix boxes since 1988 and has never once needed or used X11 over a network connection

that's difficult to believe : you have never ever used "ssh -X ..." in your professional life ? Or do you mean something else by "X11 over network " ? We (in a public research center) use it every day.

[Liam Proven]

Re: X11 over network connection

> You have never ever used "ssh -X ..." in your professional life

Nope. Never.

TBH, when I supported proprietary UNIX™ in production, X11 was an optional extra and many of them didn't have it.

But the Linux boxes I built and put in had no X server and no desktop. Remote admin was by bare SSH or using some kind of web console.

[Chris Gray 1]

Upvote!

Where do I upvote the article? :-)

A bit off-topic, but triggered by Liam's discussion of menus, etc.

Gnome has no title-bars? Yikes! No wonder I expunged it long ago.

I'm a person with a horrible memory for raw facts. The main reason I like menus (of any kind) is that they LIST OPTIONS. That can readily remind me what the names of the tools I need are. I'm afraid this is especially useful on Linux where geek programmers invent weird and hard-to-remember names for their programs/tools.

I spend the vast majority of non-browser computer time with a shell window and an emacs window. The "search tools" that various systems have are often useless to me since I don't remember the name of the tool I recall using - and I also might not recall the proper search terms to find the tool by function.

[news.bot.5543]

Removing the contributions?

I get they (xorg) might disagree with the guy, but why undo all the commits that have been applied? Surely they were accepted because they did something useful?

Is this cutting your nose off to spite your face?

[Anonymous Coward]

Re: Removing the contributions?

AFAIK it's done automatically by Gitlab.

[FIA]

Re: Removing the contributions?

It's probably easier to just quote the linked Phoronix article...

There was this revert for not handling copyright and license notices correctly. Some existing code macros were moved to a new file while dropping the existing copyright holders from being mentioned in the new file and only adding the new contributor to that header file. The code license was also changed from MIT AND X11 to MIT OR X11.

Seems fair

Also merged this week was this big revert of prior "RandR cleanups" that ended up breaking at least some RandR functionality.

There was also a revert to avoid unnecessarily breaking the NVIDIA driver. It was also commented by NVIDIA that some additional requests for other reverts are coming too.

Both seem fair, especially in the context of 'not breaking things that people use' that seems to be the basis for many Wayland objections.

There were also other reverts for code of questionable value. And other reverts making changes without knowing the prior knowledge for why some macros were added in the first place by X.Org developers.

Again, none of this seems unfair.

It seems the biggest crime might've been letting someone commit code without paying too much attention to what was being committed in the first place. Not ideal, but not that uncommon with old and unfashionable volunteer lead projects. You're happy for the effort and don't spend as much time looking until someone goes 'Have you seen this shit?'

Of course it could all just be a conspiracy by big WM to kill X, but it sounds like a much more prosaic instance of bad code.

[Anonymous Coward]

Re: Removing the contributions?

I've read the threads, and the RandR breakage appears to be because the project folk that could approve the pull requests were a little tardy and they were merged out of order. I guess the branching could have been better by the developer, but the rate he was changing stuff just outpaced the rate it was being reviewed and he was branching from a single development branch that was working and passing tests for him.

The impression I get is that the Freedesktop folk that can approve changes (mostly RedHat/IBM employees) have no interest in improving the X11 codebase as their employer is totally committed to Wayland. So it's an eager third party developer (with some views outside the technical sphere that may rub people, including me, up the wrong way) versus some vested interests that don't mind being obstructive.

[Blackjack]

As I have said before, Wayland needs more coders and less promoters.

Not allowing to control the window position? Is not like Wayland is new, why the heck is not that implemented yet?

[Liam Proven]

> why the heck is not that implemented yet?

Carsten Haizler, AKA "Rasterman" the Enlightenment developer, has an explanation:

https://lore.freedesktop.org/wayland-devel/[e-mail]/

Not sure I can adequately summarise it!

The gist is: they want to cater for tiling window managers now, and plan for the future and for things like VR/AR displays. Letting apps control window size would break such displays.

One might well ask if that is worth breaking the way existing tools work, but you know how that would go...

-----

Found via the comments here actually:

https://forums.theregister.com/forum/all/2025/06/18/kde_plasma_64_released/

[Doctor Syntax]

If I follow the Rasterman explanation correctly it boils down to not allowing some frequently used requirements to be implemented because in use cases where they would be irrelevant (e.e. tiling managers or wrapping a window round a 3D VR bunny) they'd be impossible to implement. Or have I missed something?

[b0llchit]

You are spot on. They do not implement a feature because they have no clue about and don't care about usability in a broader sense. It all seems to boil down to a narrow-minded development strategy where some group is cut off because a developer doesn't belong to that group.

[Sitaram Chamarty]

aah, the systemd ethos then

did Pottering or one of his minions join that project?

[steviator]

Re: aah, the systemd ethos then

The Poettering ethos is, in fact the Red Hat ethos.

Red Hat has for some time regarded whatever they do as the state of the art and other Linux distributions as Legacy OSes with hacked together solutions waiting for Red Hat to come along and bless their hacks and make them official or waiting for Red Hat to show them the way with an in house solution done right by the Linux mother ship.

Out of that environment we have seen Poettering's poorly designed half completed projects like avahi, pulseaudio and systemd. We get the Gnome project who ignore what their users want and refer to other DEs as "legacy" and Red Hat are also a principal sponsor of Wayland.

All these projects show common threads of ignoring well established traditions, dismissing significant and well founded design concerns and taking a leadership position without consulting the community.

This arrogance and hubris has a root and the common denominator seems to be the involvement Red Hat, it seems that the culture of Red Hat propagates the belief that Red Hat is the "official" version of Linux and other distributions are just toys made by hobbyists cosplaying as OS developers.

If it had happened once, with a single developer you could claim it was an isolated case, but again and again and again, developers associated with Red Hat keep foisting radical designs that don't seem to consider the needs of huge swaths of users.

It's time to start pointing fingers and naming names.

[Liam Proven]

Re: aah, the systemd ethos then

>This arrogance and hubris has a root and the common denominator seems to be the involvement Red Hat, it seems that the culture of Red Hat propagates the belief that Red Hat is the "official" version of Linux and other distributions are just toys made by hobbyists cosplaying as OS developers.

Very perceptive assessment.

I worked for RH for a short time. This sort of position statement was not some unspoken belief; it was part of my New Hire Orientation, an officially stated position. We were even given a little book of company culture -- I almost want to say a book of _acceptable thought_ or the Sayings of Chairman Ewing or something like that.

I still have my official Red Hat red hat: a proper felt fedora, received for completing that training. For the two DevConf.cz events I went to in recent years, I wore it, with a card saying PRESS inserted in the hat-band.

The company has a remarkably strong culture which it encourages and promotes. The result is that many younger and more naïve employees and camp followers, who know little of the outside world, are evangelists for its religion -- but they don't know that they are.

This, in part, is why I would like to see RH's main rivals distance themselves from the company. If possible, don't use RH tech, or tools, or products. They don't need to reinvent wheels; that has been done. There are racks of wheels there for the taking.

Canonical had its own desktop, Unity. It junked that. That was a mistake. But it still has its own display server, and although few see it, there is one Ubuntu-only desktop: UKUI. It's pretty good. Double down: port UKUI to Mir and make it an option.

SUSE remains mainly Central European and strong in DACH. So is KDE. I'd like to see SUSE officially re-adopt KDE and banish GNOME from its enterprise distro.

Debian once offered Xfce as its default. That was a good move. It's light, it's customisable, it runs on anything. It suits the "universal OS". Bring it back. And make systemd optional -- . MX Linux shows how to do that. Entice the Devuan fork to merge back in. They have a valid point of view. Acknowledge it.

[Gene Cash]

What's really interesting is that placing a window is 1000% up to the window manager in X11, so the proper window manager can and does handle tiling, and could also handle 3D VR bunnies.

For example, I have some apps that shit the bed because they expect the window title to be at the top, not the side (looking at YOU, Android Studio) so I have FVWM2 set to override their programmatic placement.

So this is a non-issue already in X11 and Wayland needs to keep up.

[tinpinion]

"What's really interesting is that placing a window is 1000% up to the window manager in X11"

Wrong: override_redirect.

"so the proper window manager can and does handle tiling, and could also handle 3D VR bunnies."

Partially true: the window manager may be able to handle wrapping windows around 3D VR bunnies, but those bunnies wouldn't be primitives within an X session.

"For example, I have some apps that shit the bed because they expect the window title to be at the top, not the side (looking at YOU, Android Studio) so I have FVWM2 set to override their programmatic placement."

Are you defending apps that don't follow the expectations of the window managers or saying that they're behaving badly? I can't tell.

"So this is a non-issue already in X11 and Wayland needs to keep up."

... especially with that follow up!

[tinpinion]

Wayland is a protocol that lets an application hand a shared video buffer to a display server that the combined display server/window manager (called the compositor) will then render and manage.

It isn't that window positioning isn't allowed to be implemented, it's that every window manager would need to implement that functionally themselves, since Wayland has literally no control over window placement.

If you already knew all that, then the only thing you're missing is that applications don't need to control the positions of their windows in order to run.

[midgepad]

But I have views on window position

And I'd like some of the windows I use to go to the right screen and part of the screen, please.

Without me moving them each time.

[Ken Hagan]

If you are really concerned about the logical incompatibility then it would be equally reasonable to ban tiling window managers because they can't handle window placement.

Of course, both positions are stupid, because they are absolute.

A more reasonable approach is to allow apps to say "if I can position my windows, here's where I'd like them" or even "it's me again, please try to put me where I was last time". No absolute guarantees, but no ban hammers either.

Fact is, most end-users don't switch between overlapped window managers and tiling ones or VR bunnies, so they don't care about the kind of future-proofing that the designers are aiming for.

[tinpinion]

I agree with the approach you're outlining being more reasonable than strictly banning either side, but I also don't see the lack of support as a ban so much as an unimplemented feature that wouldn't be beneficial enough to justify the cost of implementing it.

"If I can position my windows, here's where I'd like them" assumes that the application knows something about the compositor it's running on.

"I would like to position my windows, could you tell me about what I need to provide in order to do that?" might be met with:

"Okay, that sounds great! You can place your surfaces within [0, 1) in three dimensions represented as a double, and within [0, 65535] in a quantized fourth. Please ensure that you provide the correct transform matrix to ensure that the window is rotated appropriately. Your window will be removed if it is either nonplanar (as determined by us), nonrectangular within its plane (also as determined by us), if it intersects any other object, or if it is too far from the current viewport."

That entire transaction would need to be codified in an extensible manner so that any compositor could represent their own unique requirements in such a way as could be interpreted by any application aiming to make use of them.

I'm open-minded here because it doesn't really affect me personally, but I'm not convinced that we're losing anything by giving up windows being able to shuffle themselves about. I'd love to have some concrete examples if there are UI design patterns that rely on window manipulation functionality, especially if those patterns are locally optimal means of interacting with the software within whatever domain the program is related to.

Cheers!

[Alan Mackenzie]

> "If I can position my windows, here's where I'd like them" assumes that the application knows something about the compositor it's running on.

Such as that it's a on a 2d screen controlled by a keyboard and a mouse. Last time I looked, such setups are more common than having code wrapped around 3d bunnies which get eaten by dinosaurs when they are closed.

> I'm open-minded here .....

I think Carsten Haitzler probably thinks he's open minded too. Flat-Earthers are also very open minded, since they are willing to consider the evidence that the Earth isn't round.

> .... because it doesn't really affect me personally, but I'm not convinced that we're losing anything by giving up windows being able to shuffle themselves about. I'd love to have some concrete examples if there are UI design patterns that rely on window manipulation functionality, especially if those patterns are locally optimal means of interacting with the software within whatever domain the program is related to.

Lots of programs need to store their current window layouts so that they can be restored on restarting the computer after a night's sleep. That's an obvious and reasonable requirement. Some programs, for example, will need to open two windows adjacent to each other. (No, I can't give you an example). These are common operations provided by windowing systems since they first existed ~50 years ago. Wayland fails to provide them because it is open-minded.

Wayland is open minded, because it is not insisting that wheels be round. It is anticipating that somebody might want to invent square wheels, so is preventing application programs doing anything which only works with round wheels.

[Jason Bloomberg]

Some programs, for example, will need to open two windows adjacent to each other. (No, I can't give you an example).

I have a graphic LCD emulator; it has command and control in one window, another which shows the graphical output. They don't have to be side-by-side or adjacent, they just need to be where they have been dragged to as suits personal preference at some point in time. They come back where they were placed when the app is closed and re-opened later. I like C&C on the left, output on the right. Others can have it however they want.

The last thing anyone needs is to have to drag windows about to where one would like then to be every single time the app starts.

X11 may be deemed shittier than the shittiest shit under the hood but so what? Wayland is ultimately far shittier because it doesn't support what I want, makes life more painful for me.

[tinpinion]

Thank you for the example. There's also the possibility that you could have controls on one display and output on another. Extended outside of the domain you're working in, there are some cases where having a render window on a dedicated screen would be a particular benefit, like having a client-facing screen displaying a live render output. It'd be really irritating to have to dump the window into the correct location every time it spawns.

[tinpinion]

It would be incorrect for an application to assume that it is running on a 2D screen controlled by a keyboard and a mouse simply because such setups are more common than some ridiculously distracting bullshit analogy for uncertain future display technologies.

If applications need this functionality, and if you're ridiculing me for suggesting that Wayland ought to be extended to support robust, standardized window manipulation across as many compositors that would implement such an extension, what is your actual position?

[Alan Mackenzie]

The ridiculously distracting bullshit analogy (of 3d bunnies, etc.) was originated by Carsten Haitzler himself in the referenced Phoronix thread.

I think it reasonable to assume that an application will be run on a 2d screen/keyboard/mouse, if that was the technology used on the previous run of the program which saved its status file. It is also reasonable to assume that all the monitors used will be unchanged in the same position as last time, so long as the user can fix things on the rare occasion when that isn't the case.

I agree with you that Wayland should support robust standardized window manipulation. That necessarily includes the ability of a program to specify the size and position of each window it uses. The developers of Wayland apparently disagree, thinking that Wayland can work out far better than the program where the windows should go and how big they should be. This is because there may be setups where such settings aren't coherent. I think there's a gap between these developers and real world users, several of whom have made their feelings known in this thread.

Disclaimer: I've not used Wayland. What I've heard about it doesn't incline me to spend time trying it out.

[tinpinion]

> The ridiculously distracting bullshit analogy (of 3d bunnies, etc.) was originated by Carsten Haitzler himself in the referenced Phoronix thread.

I think we can both agree that Carsten Haitzler has a few screws loose.

> I think it reasonable to assume that an application will be run on a 2d screen/keyboard/mouse, if that was the technology used on the previous run of the program which saved its status file. It is also reasonable to assume that all the monitors used will be unchanged in the same position as last time, so long as the user can fix things on the rare occasion when that isn't the case.

Agreed. Would it be acceptable for the application to write a cookie out which would refer to a previous window configuration so it could pass the cookie back to the compositor on a future run and the compositor would query a database of such cookies to determine the actual position that it would need to place the window at, then place the window there? It would be a much less robust system, but it might be substantially easier to implement. We'd lose the ability for the application to place its own windows, but would gain just enough functionality to allow applications to restore windows to their previous positions.

> I agree with you that Wayland should support robust standardized window manipulation. That necessarily includes the ability of a program to specify the size and position of each window it uses. The developers of Wayland apparently disagree, thinking that Wayland can work out far better than the program where the windows should go and how big they should be. This is because there may be setups where such settings aren't coherent. I think there's a gap between these developers and real world users, several of whom have made their feelings known in this thread.

The developers of Wayland aren't really the ones who are responsible for window positioning or sizing, though. The developers of the compositors leveraging the Wayland protocol are solely responsible for those functions.

> Disclaimer: I've not used Wayland. What I've heard about it doesn't incline me to spend time trying it out.

I live in X11 because I use Xfce as my desktop environment. X11 has decades of extensions stacked on top of it, and Wayland and its family of compositors will continue to be in their infancy for a few more years at least. I hope that Wayland and Wayland-compatible compositors will eventually gain the functionality that they're lacking.

[midgepad]

My position

I'm sitting in my chair. Mundane, but a technology of work with some persistence so far.

I'd like a terminal in the top left screen, Gimp on the bottom right, Firefox windows on left lower and right lower.

And a text editor, just where it was last time.

Am I being unreasonable?

Oh, and I do like to run an X session from the other box, otherwise I get confused about mouse, keyboard, screen.

[Anonymous Coward]

> "it's me again, please try to put me where I was last time"

This is supported now, via the xdg_session_management_v1 extension documented here: https://wayland.app/protocols/wayland-protocols/18

Still fairly new - GNOME first implemented it in version 47 - but it's ready for apps to start using. And it works pretty much exactly like your comment says. The app provides an id to say which session it would like to restore, and then creates windows attached to the session by name. The compositor looks up the window in its saved session data and restores things like position, size, workspace, stacking order, etc.

[Ken Hagan]

Good to know. This is surely the dominant use-case and probably always has been, which makes me think that if X11 window managers (and Windows and...) had offered this sort of interface 40 years ago then we probably wouldn't be having this discussion.

[ChoHag]

Strange. X manages to have tiling window managers and apps which set their windows' positions. I know this because I'm using all three right now.

Has Wayland regressed on something that has been working in X for decades? Say it ain't so!

(For AR displays I have Real Windows with a delightful view of trees and squirrels and sometimes trains which far surpass anything any VR will ever achieve and if this grumpy old git doesn't need it then neither do you)

[tinpinion]

Please name of one of those apps that set their own window positions. I'm convinced that it's an anti-feature but am still waiting to see one for myself.

[PM.]

After reading the article I understand one of them is KiCad .

Not worth antbody's time to make CAD tools run under GUI ?

Srsly?

[Adair]

Doesn't GIMP, in its proper native form, shotgun its windows all over the screen—and expect them to be where they were next time?

[tinpinion]

Oooh, nice one! I don't think the application would experience difficulties if the windows weren't in the same positions at a subsequent launch, but it might be detrimental to the user's experience if they weren't restored.

[Havin_it]

GIMP has been all-in-one-window for a while now. A lot of filters use a settings dialogue in a separate window though, and it'd be nice if they could remember to appear off to the side instead of slap bang in front of the thing you're editing.

[midgepad]

My Gimp

Could be in one window, but then I'd be looking at a photo in part of a screen, with furniture, so no thanks.

[ComputerSays_noAbsolutelyNo]

Config, anybody?

With stuff in the Linux world being sometimes famously configurable, can't they make it a config option at setup?

[Anonymous Coward]

I'd take any explanation from Rasterman with a pinch of salt. The state of the code in his Enlightenment project is terrible, with what seems like a pathological desire to omit type safety and make C code behave more like JavaScript or PHP. The guy is clever, as I'm sure a lot of Wayland developers are, but like them he sees the world through a very narrow set of use cases and belittles anyone who questions his often appalling technical choices.

[Blackjack]

You know, no wonder so many people hates Wayland.

Is their way or the highway? Then people are gonna take the x11 bus to the highway.

[lsces]

KiCAD is one of my income producing tools ...

I keep being told that I 'need to switch to wayland', but if it is pulling stunts like stopping key applications from working as they have for a LONG time then it time to stick two fingers up as the people telling me I'M in the wrong.

And yes where DO we upvote such a well produced article!

[FIA]

"This fork was necessary since toxic elements within Xorg projects, moles from BigTech, are boycotting any substantial work on Xorg, in order to destroy the project, to eliminate competition of their own products. Classic embrace, extend, extinguish tactics."

Am I just getting old, but would that kind of statement ever be made by a serious project even a few years ago? Nowadays comments like this get quoted, without even any commentary as to their slight, erm, paranoia?

Who are these 'toxic elements' and 'moles from BigTech', who even is 'BigTech'? Why is there no investigation into this alleged infiltration? Is there any evidence? The amount of times you hear of this kind of stuff yet there's never anyone who's actually been approached by members of these shady cabals, despite literally everything in human history of slight contention having at least one dissenter. I mean I've worked in IT for my entire career, I could do with some cash for retirement, I'm quite open to be corrupted by 'Big Tech' to infiltrate some product or other; but I've never once been approached.... has anyone here? We should really be told.

To my jaded eyes it looks like an over eager but not perhaps as skilled as required developer has contributed enough crap to a fairly dormant project that someone has started to notice.

[Gene Cash]

I assume they're referring to RedHat pushing systemd, pulseaudio, and now wayland as hard as they can.

I get paid extra to put up with their shit and Microsoft's shit.

I don't know about "boycotting" but you try to get a RedHat employee to discuss sysv init.

[Adair]

But this is absolutely primary, unexceptional 'corporatist' behaviour. The big beasts in the jungle get to do what they want, when they want, how they want, etc.

The strength/weakness of FLOSS is that nobody actually 'owns' the code—it's evolution in the raw.

But it does rely, in the end, totally on enough people and institutions being willing, and able, to fight the good fight to survive, whether through forks or other evolutionary arguments.

Very worthy code may well end up extinct, simply through 'fashion', lack of interest, etc. That's the price we pay where 'nobody owns the code' and anybody can stick their fork in it, or simply ignore it.

[DS999]

Been using Wayland for over a month now as a part of Fedora 42

Have noticed ZERO difference in anything. It just works. I don't know what advantages I'm getting over the Xorg server I had been using for years but I do know it is much more actively developed. Who wants to use something where the development until recently was being referred to as "long dormant"? Yes you can argue "if it works why does it need development?" but when everything else around it is constantly changing having such a vital piece of the GUI puzzle remain stagnant is dumb.

I didn't deliberately "choose" Wayland, but I had been using Cinnamon with Fedora for years which has remained on Xorg (I think they are working towards Wayland but aren't there yet) but decided to give another shot at whether GNOME could work for me given the appropriate extensions. I found it had improved greatly on that front and for me it works about 95% identical to Cinnamon (i.e. basically the same bottom bar, the differences are mainly some niggles in how it handles multiple workspaces) So I took the plunge and got Wayland as part of the package - hadn't even thought about that other than the couple times I've read Reg articles that mention it and I was like "oh yeah I guess I'm using Wayland now"

[Androgynous Cupboard]

Re: Been using Wayland for over a month now as a part of Fedora 42

Look, you can’t just come along here with an open mind, actually use something for a while and then say it’s not all that bad. What do you think we are, engineers? A bit more dogma in future please, DS999.

[Doctor Syntax]

Re: Been using Wayland for over a month now as a part of Fedora 42

"I don't know what advantages I'm getting over the Xorg server I had been using for years but I do know it is much more actively developed."

Xorg has been around for a long time. Experience is that it Just Works. There's good reason to keep it like that. Do we need something that's being actively developed? "Actively developed" can mean "still needs work" or "frequently broken" but not exclusively.

[tatatata]

Re: Been using Wayland for over a month now as a part of Fedora 42

"I don't know what advantages I'm getting over the Xorg server I had been using for years"

That is my main issue with Wayland as well. However, it did produce keyboard input repetitions and lagging cursor. Also, I found it difficult to create screencaptures. In short: it solves no problems, but introduces new problems.

[Peter Gathercole]

Re: Been using Wayland for over a month now as a part of Fedora 42

"Have noticed ZERO difference in anything"

Ah, yes. You have Xwayland running, which is the bridge/glue between Xorg and Wayland. Turn that off (it's probably controlled by systemd), and then see what a "pure" Wayland/Weston implementation looks like. The Wayland people know that they can't go all-in yet (which is why Xwayland exists!), but they're trying hard to get there by crippling the alternatives.

See how many of the commands you use everyday are native Wayland clients.

[Anonymous Coward]

Always get downvoted

for explaining why neither X nor Wayland is suited to remote access in high-security environments.

Still bickering, still not listening. Roll on the ignorant downvoters who never explain why they imagine I am wrong.

[Gnisho]

Re: Always get downvoted

Many protocols you can wrap it with ssh and port forwarding as needed, and call it a day. Granted, I'll stick to Xvnc and not raw X, but...

Your mileage (and local security requirements) may vary...

[Anonymous Coward]

Re: Always get downvoted

The particular issue is not about secure protocols as such - a difficult message to get across for some reason. It is about secure architecture and containment of any compromise. A compromised client is one thing, a compromised secure server is quite another.

[Gnisho]

Re: Always get downvoted

Then one could, I don't know, actually make some sort of statement about that in the first place?

One should not be surprised at being misunderstood when one doesn't just say what they mean.

[Alistair]

Re: Always get downvoted

Gnisho:

From the flow I'm guessing the AC has no clue of what he speaks.

[Anonymous Coward]

Re: Always get downvoted

I absolutely do, hence AC. But the explanations just earn more unexplained downvotes. No pleasing some people.

As I have said here several times, it is inherently insecure to allow a server in a secure enclave to initiate (ie send a request for) a session with a client outside the enclave. Some of you may be old and ugly enough to recall the controversy when web server "push" technology was introduced. If you need a lecture in how to exploit it, go ask ChatGPT or something. And that's not even in an enclave.

It is therefore a requirement in some environments that any client session must be initiated (requested) by the client device. X runs the X client on the server host, and the X server on the client host. Thus, when a client device connects, it is the server box which requests the X session. This is what is unacceptable about X. SSH helps with other vulns, but cannot help with this. Can't recall the details of the Wayland pipe compromise thing, but that proved equally unacceptable.

Cue those downvotes. But if one person with influence learns the lesson, that is worth persisting for.

[Doctor Syntax]

Re: Always get downvoted

I don' think C was intended for the circumstances in which you're condemning its use. Because the server provides realisation of the UI as a service its relationship with the computers* doing the heavy lifting is the opposite to what one might expect. What you need is something more suited to the use case - even a dumb terminal on a serial line.

Horses for courses.

* Yes, in the plural. The whole point is that a user at one terminal can access several different providers of computation services at once. That's an inherently insecure requirement as the user can easily bridge them with no more than cut and paste.

[Doctor Syntax]

Re: Always get downvoted

Dammit, X, not C.

[Anonymous Coward]

Re: Always get downvoted

Doctor Syntax Yes, exactly that. But some apps can only be monitored/administered through a GUI - and many others more productively so - and this is where the fun starts.

[steelpillow]

Re: Always get downvoted

On cut-and-paste, I would expect that to be managed as a separate exercise. If your admins are assholes you have already lost. It is also difficult to automate the spreading of malware using such methods. Not really the same problem the AC is banging on about.

[DS999]

Re: Always get downvoted

Look you can't complain "the security of our secure environment is compromised if someone does something that the firewall should be blocking". If people can do that and violate the "security" of that environment then whoever is operating both the network and the security is incompetent.

You want to have the session initiated by the client device? Fine, leave open the SSH port at the firewall, run an SSH server on the server, then the client must initiate by connecting to SSH on the server and use X forwarding to connect.

Or don't use X (or Wayland) at all, use something like VNC which has the sort of "client connects to the server" behavior you're seeking.

[Anonymous Coward]

Re: Always get downvoted

First point. Firewalls can be breached or physically bypassed. As you seem unable to grasp, this is about containment were such a thing to happen: defence in depth.

Second point: As you seem unable to grasp, this is not about secure protocols as such. Using forwarding, the X client on the server host still requests the X session. SSH is irrelevant if the damage is already done.

Third point. Exactly what I am saying. There are various adequate secure solutions, but X and Wayland are not among them.

[Havin_it]

Re: Always get downvoted

So ... what you're concerned about is the client machine's X server being compromised/naughty and then compromising the server through the X client app (whatever it may be)?

1. This is a pretty damn skinny attack vector considering how many people actually use X forwarding (let alone on servers of any value/interest), plus requiring a chain of at least 3 exploits to pwn the server (machine)

2. I'd say it falls within the standard purview of whoever admins the server (machine) to sandbox or otherwise lockdown any client apps that might be run on it to an extent that is satisfactory for that server's particular security needs. This applies equally to Xeyes run in a forwarded X session as it does to nano run in an SSH session, as far as I see it.

The whole client-on-the-server/server-on-the-client element of this topic still makes my brain hurt a bit after 20some years so apologies if I've misunderstood your thrust here.

[Doctor Syntax]

Re: Always get downvoted

Always assuming the A/C you just replied to was the OP of the thread. That's the trouble with A/Cs.

Perhaps el Reg could look at a scheme whereby A/Cs get numbered as they join in the commentary of an article. That way they could maintain their anonymity if they need it but maintain the integrity of a discussion thread.

[Alistair]

Re: Always get downvoted

I'll point out that SSH is secure. And my X sessions always tunnel. In all honesty, over 25 plus years of using various unices in an enterprise environment, I've *NEVER* had an unprotected X session configured. Localhost ONLY. Tunnelled back to the desktop in SSH. I built *all* our systems that sported X that way. And in coherent discussions with real systems admins all over the world it is generally the default configuration for most Unixes.

Unless you're talking air gap requirements, configured correctly Xorg server can be made secure, I've not tried similar with wayland as I've never used it.

[steelpillow]

Re: Always get downvoted

What part of "it's not about secure protocols, it's about secure architecture" are you having trouble with?

[Doctor Syntax]

Re: Always get downvoted

"for explaining"

Where? Your post consists of a complaint, not an explanation.

[exovert]

I feel as though GNOME (3) & Wayland are as one developed from the "we're doing what we want to do" mindset, as though the entire basis of open source were not collaboration, at least with other developers. There are some worthwhile goals they want to achieve, and you can't get some of them except by pushing even when it's hard. But there are grey areas, as well as the total blackness of blind spots.

I did think the thought at some point, that 'what do I care if all x applications can read keypresses, they call come from the repos'. But maybe I do want the web browser not to be able to read what I'm typing into RDP sessions. Is that enough of a benefit to wipe out the *bsd's from the application ecosystem. maybe it won't, but it doesn't work yet.

Which incidentally even on xfreerdp directly, I still can't get to behave with multiple monitors. I have to keep bouncing into X11 for that. To work. Is that important enough? Guess not, fine on their laptop probably. (X11 forwarding not unuseful to get a window, out of WSL either)

Unfortunately the highly shall I say, disassociated politics of the driving agent for this fork (https://www.reddit.com/r/linux/comments/1l9s7rg/comment/mxfigv7/ , not interested enough to find a direct source) make me feel more tantrum than movement.

[Jou (Mxyzptlk)]

I still like the principle of X

Letting all the graphics be actually rendered remotely instead of rendered locally, compress, and then decompress remotely. The one big advantage over most other solutions.

And maybe, FINALLY, a free fully functional X-Server for Windows, relatively easy to install.

[eldakka]

Re: I still like the principle of X

> And maybe, FINALLY, a free fully functional X-Server for Windows, relatively easy to install.

Maybe I'm misunderstanding something, but there is a fuly functional free X-server for windows availabe via Cygwin.

[Jou (Mxyzptlk)]

Re: I still like the principle of X

I know, but then you have to get the FULL cygwin bloat along. It is not like, with other cygwin tools, you can "extract" them when you know how to get the dependency .dlls (fetchmail / rsync / wget in most of my cases).

[Doctor Syntax]

Re: I still like the principle of X

Who notices a bit of extra bloat on Windows?

[Jou (Mxyzptlk)]

Re: I still like the principle of X

/me, and a few others who can manage to run a full Windows 10/Server2019/Server2022/Server2025 on 30 GB storage. With about 10 GB free space left. Without needing sick tricks. See how Windows 11, with the exception of MAYBE 21h2, is not on the list. I could add Server 2012R2 though. Server 2016 no chance. Current preference is Server 2022.

But apart from that you are right, I am among the minority caring about that. And with the server version I can mount or junction in a deduplicated second volume to \cygwin, parts of \program* (if not the whole three \Program* directories, but I never tested that), my %userprofile% and a few other things.

[docmattock]

Re: I still like the principle of X

See MobaXterm - free version fully functional for X.

[abend0c4]

Re: I still like the principle of X

The fundamental problem is that technology changes in ways that nullify historic assumptions.

X arose at a time of slow CPUs, simple graphics hardware and slow networks, so it contained a lot of features (such as remote font rendering and drawing and filling geometric shapes) that were intended to compensate but didn't match what more sophisticated applications (i.e. other than xterm or xclock) needed to do - almost all applications now do these via some other graphics library. In reality, X is mostly moving bitmaps around, because that's what made sense when GPUs and networks improved, though it can pass OpenGL (for example) transparently as a more modern twist on its old remote rendering capabilities. Which behaviour you get,of course, depends on which API the application calls, which is not perhaps the kind of display independence you might hope for.

The challenge for interfaces is always picking the abstraction that has the best chance of being robust (in the sense of not changing behaviour) in the face of those technology changes. Arguably, X. in retrospect, made some fundamental errors (such as integer pixel coordinates), some avoidable assumptions (square pixels, rectangular windows) that were excusable at the time and had some good ideas (such as the remote drawing instructions) that weren't in practice usable by applications (because they were too primitive).

It's not an accident that both X and Wayland are essentially bitmap wranglers: it's the lowest common denominator that accommodates existing applications. A true remote display protocol might well include rendering instructions you could ship across a pipeline rather than bitmaps. You'd also probably be looking at allowing the display host to retrieve media (such as video or audio) from a third location rather than have to be received by the application host from a remote location and then relayed to the display host. You might even want display lists of objects that can be grouped and transformed independently. The snag is that you'd only get these things in applications that were specifically written to take advantage of them - chicken, meet egg!

I agree that Wayland is punting far too many usability considerations. However, it's important to remember that X has had vociferous critics, too, for many years. The UNIX-HATERS Handbook quotes Marcus Ranum of DEC:

If the designers of X Windows built cars, there would be no fewer than five steering wheels hidden about the cockpit, none of which followed the same principles—but you’d be able to shift gears with your car stereo. Useful feature, that.

The thing is that people typically, in the end, take what they're given - particularly if they get it for nothing. We'd have to do something more constructive than grumbling to change that, but grumbling is the only skill I keep truly current.

[Jou (Mxyzptlk)]

Re: I still like the principle of X

The WPF way (yes MS-Windows...) could have been made to be more "remote" than it is now. The normal Windows user interface takes the definitions from the program, sends them to the remote display, and let it render there. All that RDP and ShCitrix stuff never event tried to. Probable reason: MS is too half arsed about way too many things, and the way I suggest was not "quick and dirty" enough. Only a small subset of things were made > 90% right. Currently MS is on the mission to demolish that good subset.

[Doctor Syntax]

Re: I still like the principle of X

If the designers of X Windows built cars you'd have one cockpit remotely controlling several cars. Yes, with multiple steering wheels. If you wanted one steering wheel to control them all you need the designers of KVM switches for that.

It's an odd fact of life that different use cases need different solutions.

[Bluck Mutter]

Firmly in the X11 camp

Been a Unix dev since early 80's and have a large repository of personal stuff I developed since retiring in 2019 based on X11 ...and no.. wayland sucks with it's X11 compatibility even after all these years.

I have been a Unix and now Linux exclusive home for decades (eat my own dog food and wife was a Unix admin so she is part of my party)

The loss of X11 in mainstream distro's would sink me so I have spent a lot of time viewing these developments and future proofing myself.

Downloaded the X11 code (inc libraries) and got most of that to a compile but hit some roadblocks (will revisit at some point).

My go forward platform is Ubuntu server with fluxbox and tint2. I figure that even if desktops like Gnome and KDE drop it and X11 ones like XCFE also do that:

(1) the base X11 libraries will always be provided on server platforms

(2) that X11 based stuff like fluxbox and tint2 will also exist (in fact I compile these from source)

(3) can compile newer gcc and kernels on older distro's to keep them current (until some libc issues arise)

I am also stock piling PC hardware. Worst case I am frozen in time and can run modern web browsers in a VM....nothing else I need (libreoffice, bluefish, dia etc) will need to stay current.

Just need to get through the next 30 years until I become (more of) a dribbling old fool who cant use a desktop or pop my clogs.

Bluck

[Ace2]

Re: Firmly in the X11 camp

I am not criticizing you here… but does it say something that someone with your level of experience couldn’t get X11 to build?

[damiandixon]

Re: Firmly in the X11 camp

I've built X11 for embedded hardware several times. It's just time consuming because the build system needs updating.

[tinpinion]

Re: Firmly in the X11 camp

Hello, I'm here to shill for Gentoo, where building the entire system from nearly scratch is the norm, where creating custom packages is basically just a matter of bash scripting, and where you too can maintain a local repository of all the source files necessary to build any non-proprietary software by simply rsync'ing a distfile mirror to a large enough hard disk.

I'm off the grid often, so it's nice to be able to access a huge library of tools that can be configured to use different dependencies than what a distro maintainer decided for me simply by carrying an external hard disk around.

It's not easy, and I bounced off of it the first time I tried it, but coming back at it with more experience in working with Linux under my belt made me fall in love. (It's a heck of a lot easier than building Linux From Scratch!)

[Bitsminer]

"The Reg FOSS desk is nearing 60..."

Let the children play...

[PhilipN]

Re: "The Reg FOSS desk is nearing 60..."

In another ten 60 will no longer mean anything and then when you need to take a moment to work out how old you are the great thing is you realise it no longer matters.

[Liam Proven]

Re: "The Reg FOSS desk is nearing 60..."

> Let the children play...

They absolutely can.

The problem is that the children are demanding that they have the right to take away _my_ toys.

[rcw88]

Touch screens? Niet,no,non

A touch screen UI for a mainstream computer just sucks. How the hell can I touch type, with ALL my digits? Creating a working UI is HARD, just ask Xerox, or Apple, we’ll ignore Microso$t as they just stole Apple’s UI.

The idea that a loose federation of software developers can create a new touch UI is just laughable- I dumped Ubuntu because of Wayland and the AppImage crapware.

Usability and Accessibility are THE most essential features, some of us have crap eyesight from an early age, some don’t, but they sure as hell will by the time they are 70. And don’t get me started on voice recognition, Sorry Dave I cannot do that is still a long way off.

And the same applies in cars, if we aren’t allowed to use a phone on the move, why is a touch driven UI acceptable? Buttons/keyboards work.

[chololennon]

Re: Touch screens? Niet,no,non

> just ask Xerox, or Apple, we’ll ignore Microso$t as they just stole Apple’s UI.

You just forgot that Apple stole ideas from Xerox.

[Liam Proven]

Re: Touch screens? Niet,no,non

> You just forgot that Apple stole ideas from Xerox.

Not as many as the people who keep saying this actually believe.

Apple took windows. That's about all.

It invented...

* Menu bars and menus

* Dialog boxes

* Default buttons

* Title bars with standard controls

* Windows that could overlap and the partly hidden ones still update

* Standard drawing toolkits accessible from any app

And much much more.

[chuckufarley]

Wot I think...

...Is that doesn't matter in the end. X.org, Wayland, whatever. If you still need to run JavaScript, Java, or C# code you and yours are Fuxxord.

[Anonymous Coward]

"It's called aging, and it's a horrific experience, but it's better than the alternative."

I noticed today on our public broadcaster's news site that UK citizens are to get that choice.

UK parliament votes for assisted dying, paving way for historic law change

When I first read the title I imagined Brits were going to be invited to submits their little lists nominating those they believed required firm assistance in that direction.

I think every AU state and territory has similar legislation which in some has been in force for some years.

[carl0s]

Mouse warp

Discouraging indeed. I think I read that mouse warp protocol has just been released and various compositors have already added it though, this week possibly.

[keithpeter]

Editors that support the CUA key shortcuts: dte

"This grumpy old user wants to continue using his standards-based keyboard-driven user interface, complete with matching text editor. "

OK so no TUI menus but the CUA shortcuts are there and working. And ldd shows a very small set of libraries, basically glibc and the usual dynamic elf type stuff (on OpenBSD at least).

https://craigbarnes.gitlab.io/dte/index.html

Back on topic: we'll have to see how this fork progresses but I'm not holding my breath.

[Jou (Mxyzptlk)]

Question to Reg-Readers

Apart from some nice interruptions from Liam, Who-Me and On-Call (and way too rare others):

I have a strong feeling that the average bullshit level of articles in The Reg is rising for the last four month. But the actual question is not whether that is true or not, that is not "up for discussion".

The question is whether the world is currently creating more bullshit than ever before, 'cause I see the same tendency on other news platforms as well, with some much better being able to keep such news at bay, and more others being so much worse than The Reg...

[DrReD]

Great Article! Thanks.

[Will Godfrey]

X marks the spot for me

The application I work with uses multiple different windows (as chosen by the user) and these are all re-scalable, but necessarily at fixed width/height ratios. Window position and whether currently visible is also stored on exit, so the user can immediately get back to where they were. Also the software can run multiple instances and the title bars are used to identify the root file name, window type and the instance it belongs to.

As far as our users are concerned it aint broke - so don't 'kin fix it!

[CyHarsh]

X (windows) over networks was pretty cool

Back in the mid 90s we used a pretty heavy CAD tool. The tool would run on workstation A and render on Workstation B and vice versa.

It turned out that would improve the overall performance for both users.

[dboddie]

MDI FTW?

"Wayland does not currently allow controlling window position. This means that when you open KiCad, it cannot remember where you last placed your windows."

Time to go old school, then? https://en.wikipedia.org/wiki/Multiple-document_interface

[DuncanLarge]

Until you need it...

> The Reg FOSS desk has been installing and supporting Unix boxes since 1988 and has never once needed or used X11 over a network connection.

Where I work in IT in the aerospace industry this feature of X is used all the frigging time!

We have a HPC system (1400 cores) that runs multiple jobs submitted by users. Also several Redhat development machines and some very old Linux installs which are in the usual category of "it's bloody critical so upgrade must be done as full projects" resulting in some critical Linux systems being so old as to be 20 years old and the sort of distros I was cutting my teeth on in University.

All of them from the old to the not so old and the new have one common feature that EVERYONE uses: X11 Forwarding over SSH.

Developers, testers, designers all of them REQUIRE X forwarding and if it isn’t working its IT who gets the tickets.

When I mention the way Wayland demands that you just "RDP" or "VNC" to the desktop like its 2005, they all baulk at the idea. They hate it. They dont want to RDP to a machine then another then another, just so they can run NEDIT on a text file that a datalogger just saved!

At home I dont use X forwarding, but thats because I HAVE ONE MAIN MACHINE and dont use SSH either ;) even when I have a couple of laptops on as well I'll rarely need to use ssh. Last time I used ssh at home was to ssh from my tablet to my PC, to shut it down :D

The users at work have WSL 2, which SUPPOSRT NATIVE X calls on Windows 10/11 so they can forward the apps on the HPC to their usual desktops. Why the hell did MS add that feature? WSL didnt have it and before people moved to WSL 2 here we simply installed a separate xclient application to all windows pc's. Someone must have wanted it for MS to make it work.

So by GNOME/redhat etc simply trying to kill it for an inferior system, well that suggests they are against actually making sure there is: choice.

X11 was always a problem, it was one of those things you had to have in Linux which you HAD NO CHOICE about. Since I started using Linux in 1998 I quickly realised that there was more than one way and those ways were capable of simply replacing the others, as UNIX ensured a degree of standard interfaces that made composition possible on the command line for example. Unlike on Windows I could install any desktop environment or forego a DE and use a window manager, I was even able to use any window manager INSIDE many desktop environments. I literally could build MY OWN solution! Today in fact, I don’t use XFCE, havnt used Gnone since GNONE 3 scared me off even. I’ve dabbled with LXDE and titling WMs like i3. But today I went back to an old mate; WindowMaker. And it’s great, light fast and works. Nothing is broken. Apart from system which hates WindowMaker telling it to shutdown…

Subsequently I hated more and more of the Windows world where there was "one way". Like Win 8/8.1/10/11 that all have NO capability to adjust UI colours. A feature a colourblind user like me WANTS. Instead I have to have the same desktop as everyone else but I can add a bit of flair and express my individuality with an "Accent Colour"; omg I feel spoilt and giddy!

X11 was an awkward thing in my Linux world as there was no alternative. Blackbox can replace GNU binutils etc, a functional alternative, but there was nothing for X11. X.Org was just a fork.

Wayland could have been that alternative but it clearly didnt want to be. It wanted to break the existing system and enforce a new ONE WAY. If it had tried to remain compatible, to implement the same core functionality that everyone expected, we probably wouldn’t be having this argument.

[Jou (Mxyzptlk)]

Re: Until you need it...

> WSL 2, which SUPPOSRT NATIVE X calls on Windows 10/11

Waaait WHAT? Okay, now I finally have a real reason to WSL 2, up to that point I saw only a limited usage for a real scenario.

First good hit in Search Engine: https://stackoverflow.com/questions/61110603/how-to-set-up-working-x11-forwarding-on-wsl2

You need more upvotes...

[AJ MacLeod]

Re: Until you need it...

Upvote just for WindowMaker - boxy but good! Can't really imagine using anything else on my main PC.

[b3ta]

A position on position

I am involved in use-cases where the absolute position and size of a specific window is of paramount importance. In one case it completely defines the experience of hundreds of people of an event as being a success or a failure. That case is when an operator double-clicks on a media file and expects it to be displayed on the projectors, and not on one of the screens s/he uses to control other aspects of the show.

When the lighting control program opens, it must be on THAT screen, in THAT corner, and it must be exactly THAT big. When two file manger windows are opened, they must be THERE, THAT size. The program controlling the backtracks that go to the sound desk opens on the central monitor, over there, at that size.

We even have a case where we need to achieve a picture-in-picture effect by popping one window over another... but only on one of the three displays which otherwise have identical output.

If I can no longer do these things, then I guess it is time to buy a Windows licence.

You see, Windows is fine letting VLC define where and how its window needs to be displayed: VLC > Tools > Preferences > Video > Fullscreen Video Device. That's just one aspect of getting VLC to work as I need. Is it complex? Yes, but that's because all those options mean something, allowing one to make amazing things happen, with the operators running all of it clueless about how it all works --- they merely use the tools which are configured just right for them.

Then there's the case of doing video routing with the likes of OBS Studio, where we have to grab parts of program outputs and mirror them to an external display; sometimes to a virtual camera for getting the real camera input to a Zoom session. There is no way I can fiddle with dragging windows to certain places and sizes so that the image is exactly right to comply with the standards used by the hardware world.

On my desk I also have have multiple monitors, and when I open Inkscape, it needs to have its main window full-size on the big screen in the middle. The smaller control windows that I use almost all the time (layers and objects, text and font, align and distribute) must be in specific places on the right-hand monitor. Design notes are open in an editor on the left-hand screen at a specific size and up against the right-hand edge of the monitor, which leaves enough real-estate on the left for a specially-sized file browser.

I have other use-cases, but I think my position [!] is clear: it is about usability and being able to create a fit-for-purpose environment. You cannot have a generally applicable system by removing expressive power.

To go CompSci: Stephen Wolfram's "Rule 110" is a Turing-complete language, so you could write a data base in it, but should you? The expressive power of the general-purpose languages exists for a reason, and something like X11 or Wayland is very general-purpose, with the requirement to be able to meet extremely varied needs.

[docmattock]

What what what?

“ The Reg FOSS desk has been installing and supporting Unix boxes since 1988 and has never once needed or used X11 over a network connection.”

So you’ve never really used X11 as it was intended?

One X11/xorg client (the server :)), multiple machines with remote windows open all doing different things on different hardware?

You’ve really missed a lot of the fun, then. This statement - which _surely_ can’t be completely true - means you need wayland, because you want a single user machine with all the benefit (yeah, right) that a monolithic graphics engine like Windows or macOS provide. Then your remote access to the desktop is limited to screen grabbing techniques (vlc, rdp), which as we know are awful over anything but direct wired connections.

But for all of us that roam across computing environments, sometimes across continents, X11 was designed as an elegant solution for that use case.

It also just happens to work with X clients and X server on the same machine, which only started happening with the advent of personal desktop computers more powerful than the Vax mini computer, i.e. a 486 with 4MB ram and a maths coprocessor.

Try it! You’ll have fun!