This Wayland Issue Will Get You Banned

[@BrodieRobertson]

Just in case anyone thinks I said something wild linking this Artem dude to Birdie, if you notice he uses the same handle on both platforms, and I can't imagine there are that many Birdie's in the Linux space who are obsessed with Wayland

[@myownfriend23]

The birdie on Phoronix has also said that he's Artem multiple times in his posts. As Avis, an account he started on the same day his birdie account was banned, he pretends he isn't birdie. Of course when he first made the account, his profile even said "In my past life, I was a bird"

[@szymonagiewka4513]

And avis is latin for a bird 😂

[@yoshistar3141]

Hi, i saw your video of Atlas OS, security breaches and other stuff, i feel ignorant about the terminology of software and their functions, do you have some recommendation from were to begin learning, videos, books?

[@okashiromi5541]

​@@yoshistar3141for general stuff, look at code/software engineering stuff I guess? That'll teach you how stuff works. If you're interested in a specific thing, search for that thing. You can also lurk around on tech spaces, basically learn bit by bit from the environment

[@infinitivez]

Even people who been dormant for years, know its Artem. 😂Honestly, we're all just humoring him for funsies at this point. If only he'd put half this effort back into xorg 😜

[@Felinaro]

One of the main problems of wayland is its mainly not "work in progress", but "rant in progress".

[@hwertz10]

For real. I mean, it appears to be in OK shape now? In general? But, as a Linux user since 1994 or so, I followed Wayland pretty early. And it's like "This doesn't work" (early on even multiple monitors). "Why would you want to do that?" was a common response from the Wayland devs. Then they'd get on with it and implement the missing functionality. Rinse and repeat. I was shocked when I realized (just within the last year) that Wayland is ONLY a protocol and every (what in X is called a window manager) compositor must implement the Wayland protocol and effectively implement it independently. It does seem like it would be good to have at least a real reference implementation for this. I do think the Wayland devs, rather than bickering over what should and should not be included in some definitive code base,, or saying it's impossible, why not just encourage the devs of various compositors who have been found to handle a specific function particularly well to just submit some reference code. I mean, especially for some of this stuff where it's complicated, it has to be a hell of a lot easier to have some reference code to work with or at least reference compared to just having a protocol description of how it should work and writing code from that. Of course, this code already exists within kwin, weston, etc. etc., but it's be a lot easier to have it seperated out instead of havng to go through all these code bases to go find it. Agreed regarding xrandr complaint being silly. I haven't used wayfire, but complaining about complex configruations and stuff? xrandr is not simple to set up or use.

[@miigon9117]

​​@@hwertz10there is a real reference implementation, it's called weston. Don't like it? there's also wlroots.

[@miigon9117]

​@@hwertz10and yeah his configuration complaint is silly. not wanting to learn a new configuration format sounds like a personal problem.

[@azari007]

@@hwertz10 you're either full of shit or seriously misrepresenting wayland devs. i've been following wayland since early days as well, and even lurked in the irc channel for many years back in the day, *and* participated in a few of the arguments and flame wars myself a few times. there was *never* any intention to leave out functionality. the question has always been *how* to implement it. and wayland devs have never had to actually compromise on their principles unlike what people keep falsely repeating; they have always accepted solutions that *fit* the wayland model. and this is always how it was intended to be. the current state of wayland's progress is not because "people have pushed against the stubborn wayland devs and gotten them to accept things", it's rather that solutions fitting the wayland model eventually reached a level of stability and have then made it into various compositors, thus becoming standard. as the other reply pointed out, there is already a reference implementation, it's weston. the reason weston didn't get used by others wasn't because it wasn't credible, but because other DEs didn't *want* to operate within weston's constraints and wanted to do things differently, which was a predicted outcome and the very reason wayland was made "just a protocol" to begin with, because wayland devs have been in the community long enough to know any solution that *forces* a reference implementation would get zero adoption, people want to do their own thing. for those that don't, shared libraries are a thing. as for your suggestion, this is arguably how it already works. the "bickering" you're referring to is literally irreconcilable differences of opinion regarding certain functions, and the whole *problem* is precisely that no compositor *has* figured out a solution for said feature that "makes everyone happy". what you're asking for is for them to just "shut up and do what i want", which obviously they aren't going to do.

[@TurtleKwitty]

@@miigon9117 isnt weston the one "mainline" implementation that refuses to implement half the damn protocols that are non optional though XD

[@thingsiplay]

"Birdie" sounds like the "Brodie" from an alternative universe where Brodie is a X11 shill.

[@jeffreyjoshuarollin9554]

Brodie makes way too much sense to be birdie, and I'm *not* talking about the fact that English is presumably not Artem's first language. He's very good at (trolling in) it.

[@cheako91155]

I'm waiting for an Xorg module that listens for wayland clients... :)

[@immibis]

I thought about it. When I downloaded Xorg's code I got more distracted by looking at the overall code quality (it's fine) and trying to identify parts that were no longer necessary (nothing unless we can decide which things it should stop supporting or make a fork that stops supporting things). Ironically the hardest thing about it will be to implement the Wayland protocol, which is much more complicated than X11! The most interesting thing I found in Xorg is a security framework. You know how the Wayland people keep saying X11 is insecure? Well! There's a solution for that! But nobody uses it.

[@JTCPingas]

Mad man

[@uis246]

​@@immibisWow. I didn't know. Thanks for digging in sources and telling us.

[@immibis]

​@@uis246 btw the reason Xorg has so much code is that it supports so many system types that Wayland doesn't support. The kernel-modesetting driver is just one of many drivers, like a VGA driver, while most Wayland compositors ONLY support kernel-modesetting, which is fine because most systems support kernel-modesetting, but if it's to be removed from Xorg, can we really be sure nobody still cares about the systems that don't? I can't be sure of that. A Wayland compositor that supports all the platforms Xorg does will be just as complex as Xorg. The X11 protocol isn't the complex thing about Xorg.

[@speedytruck]

@immibis If it's so easy than go fix all of xorg's issues then, including multiple monitors, HDR, and enable that "security framwork" you're talking about. BTW, don't forget that every app and window manager relies on an xorg quirk one way or another i.e. non standard x11 function. Red Hat could really use your help! They're basically the only ones maintaining it and even they can't take it anymore and are dropping it soon...

[@WiihawkPL]

i refuse to switch to wayland until a compositor exists that puts an anime girl on my windows

[@umop3plsdn]

*cough* Hyprland does this *cough* Waifus4u

[@timur1228]

As has been said, Hyprland does this right now.

[@uuu12343]

This video title is a pretty good representation of the Wayland project in general "Why would you want this" - says the Wayland project, for every, single, idea

[@АлексейГриднев-и7р]

As Steve Jobs said about the connectivity problem with iPhone 4, "you're just holding it wrong".

[@DizzyAndHigh]

This is an official NACK from the Insufferable Technical Working Group to include this proposal in the xdg (or wp) workspaces. We believe that this protocol hurts our ego too much, because fuck you. Only crackheads would need this thing, and if you need this, you're just a crackhead. It is still theoretically possible to go fuck yourself and your crackheadedness, and silently weep while waiting for your compositor to implement a nonstandard protocol for this that nobody would use anyway. And in the case that any compositor actually does this, we will screw their upstream to oblivion.

[@Coopertronics]

PulseAudio was a pain to setup back in the day, but it just got better with age. I'm sure the same thing will happen with Wayland. Time and patients is all that's really required. As the old saying goes "Don't try to run before you can walk"

[@ocsanik502]

patience*

[@Coopertronics]

@@ocsanik502 Which do you prefer, patients or patience? You may have to wait a long time for the answer ...

[@TheEvilAdministrator]

I dunno, I think a lot of people still think PulseAudio is crap. Presumably that's how we ended up with PipeWire...

[@monsterovich]

It literally got killed by Pipewire lmao

[@ocsanik502]

@@monsterovich Not entirely, there are still use cases that PipeWire isn't qualified for yet, namely in the music space.

[@alh-xj6gt]

4:00 IME - Korean characterset is called Hangul.

[@KoopstaKlicca]

Just for the sake of being pedantic, it's also known as chosongul in N Korea

[@fenix849]

@@KoopstaKlicca Ok i have to ask, is there any non-north korean produced operating system out there that has that as an option and labeled by the north korean name? Perhaps chinese linux distros?

[@KoopstaKlicca]

i don't know@@fenix849

[@MarcCastellsBallesta]

​@@fenix849Somebody please create a distro called Kim Jong Unix.

[@tranthien3932]

Vietnamese character set is called Unikey

[@talkysassis]

We actually have a point here. As a dev making a wayland app. What's the base runtime I should target that everyone will use? What's the standard? Then I look at the standard and see that many features I need are external modules instead of a requirement meaning that I can't rely on them for production as I have no clue if the user can run it. My options are to choose distros with the support I need and target them, but that's not a good solution. I can see a trend for a desktop API that is universal for all distros, but it is separated in so many layers for so many different projects that is too hard to know what the desktop API really is. I'm all in for a shell api that gives me all I need (like, instead of wayland calls, we have shell calls that call wayland, the same for pipewire, portals, dialogs, gpu encoding etc.) but only KDE has something like that and is not complete as KDE need dbus for lots of things instead of a proper API.

[@quadrupledamage]

i'd guess gnome is the standard because gnome is the most broken by design out of all(wayland's core ideology)

[@talkysassis]

@@quadrupledamage Problem is that there's a limit on how minimum your system is. GNOME can't provide features that many devs want to use, and really want to force us to change our UX designs to fit their ideas. At least KDE tries to give devs all the things they want to use on the shell like tray, custom window decorations and static libs.

[@Beryesa.]

Target our freedesktop, that's what we're united upon.

[@WiihawkPL]

this was the same problem with X back in the day and ultimately what killed the unix workstation, i wonder if the wayland devs are aware of this

[@uis246]

GPU encoding is definitely not something graphical shell should do. GPU encoding is done in GPU drivers.

[@SeralyneYT]

The Korean character set is called Hangul, by the way! As for 21:26 - And you actually used to be able to do the desktop switching. Granted, under the hood, it was still DWM, but SharpEnviro was essentially a wannabe-linux shell replacement that existed for a while. It wasn't as extensive as what Linux offers, but it got you a good bit of the way there. It even had the desktop cube!

[@zlice0]

cue xkcd 927 B) besides the bickering, fragmentation is def 2nd for holding things back. it was left to the masses on how to do the extra stuff like listening for keys or sharing windows/screens and then everyone's surprised pikachu face when programs like discord go lul-nope

[@electronics-girl]

Yeah. One of the strengths of X11 (and its related specifications, like the ICCCM) was that it was built around the idea that diverse programs should be able to coexist and interoperate with each other, even if they were written with different toolkits. Wayland seems to be punting so much to the desktop environment or the toolkit that it's just not going to be as interoperable.

[@Mallchad]

In my opinion it is kind of a major issue that a bunch of things are just going unimplimented and being left completely up to the compositor. Not because it takes a lot of but but it means you have a bunch of different composition features which are incompatible with each other and leads to confusion and breakable application-side with development. Obviously that's something that should be setup as Wayland protocols but A. they aren't done yet, the spec, not even the implimentation. B, multiple implimentations leads to extremely quirky and inconsistent behaviour (although you could allow alternative implimentations), if compositors skip feature implimentations (and they will) you will just end up with most compositors being incomplete and incompatible with some apps. Which is an annoying place to be development wise

[@xipietotec]

Oh also, birdie was a bit right (but not really, it’s very hacky, and was never really popular in Windows), and maybe they did confuse a few things. Prior to vista, Windows did have a more “xorg-like” setup (superficially), what I think they’re referring to is XP used to allow, and there were a ***few*** implementations of “alternative” window managers. I used some on XP, I imagine they must have been absolutely terrible to maintain because they never took off and frequently broke applications. But I tried several out for “cool factor” back in the XP days. There wasn’t just DWM, enlightenment for instance would run on Windows XP

[@Poldovico]

I remember using a Stardock thing that would turn your desktop into a box with physics-enabled glass tiles as icons. You could shake the box, move the tiles, bounce them around. It would make my tiny EeePC cry in agony whenever I tried running it.

[@xipietotec]

@@Poldovico yep! Stardock still has the best solution to windows desktop shortcut bloat (Fences). It’s honestly a feature set I’d like to see implemented in other desktop environments

[@Poldovico]

@@xipietotec Pretty sure Plasma 4 had fence-like constructs. Might even be where Stardock got the idea. Nowadays my solution is no desktop shortcuts at all. I use a Mac-like setup with a lower dock and top bar. The icons in the dock cover over 95% of my computer time, and for everything else I use Kickoff and KRunner.

[@Poldovico]

@@xipietotec BumpTop! I found its name! Apparently it went open source at some point.

[@xipietotec]

@@Poldovico that’s pretty cool, I will have to check it out. Back when plasma launched I was a gnome user running compiz Edit: also why I stayed out of further developments in Wayland and compositors and DE’s in general is that I had moved into dynamic window managers and creating a bespoke experience for me, so by the time later versions of KDE and plasma had launched, I was using an awesomeWM setup. I’ve been away from computers for a few years.

[@matthias6933]

morale of the story: engaging with birdie in good faith is a waste of time and energy

[@BrodieRobertson]

Nah it's fun

[@honza970]

​@@BrodieRobertsonFrom outside. People like him sap energy from open source projects. Instead of making his idea reality, he demands others do work for him.

[@MegaManNeo]

Guess the road to Wayland is covered with many small stones unless your name is Fedora. Being a Plasma guy and being dependent on Barrier, Xorg is still relevant to me in 2020v4.

[@cameronbosch1213]

Legitimate question, what is Barrier?

[@MegaManNeo]

@@cameronbosch1213 It's a fork of Synergy which is a software based KM switch.

[@WiihawkPL]

doesn't input-leap (the follow-up to barrier by the same people) support wayland now?

[@MegaManNeo]

@@WiihawkPL it should but from what I have understand in another video, it's still under development and not yet available for public use or something.

[@notuxnobux]

11:22 you know exactly what it has to do with wayland. You had a discussion with a kde dev about it. The kde dev mentioned that the compositor can be restarted but only for applications that use qt or other major toolkits. So the complaint here is that reloading is a feature of the toolkits and that wayland itself (or compositors) cant handle it automatically so it doesn't work in all applications (and never will).

[@fuseteam]

Wayland is a set of protocols. It only defines how features can implemented ;)

[@DaraelDraconis]

​@@fuseteamAnd when those protocols don't cover a feature, it's very convenient for Wayland developers that they can then say "it's not part of the spec so it's not our problem", _even if it affects everything that uses the protocol._ Yes, Wayland has declared that compositor-restarts are not part of the protocol and therefore are for all clients to work out by themselves. The obvious response is "that _should_ be covered by the protocol because so long as it isn't, compositor-restarts will continue to break X, Y, and Z".

[@fuseteam]

@@DaraelDraconis orrrr the missing protocols can be added, so it is covered :D

[@Abu_Shawarib]

That dude is crazy. No wonder he got banned on Phoronix.

[@fu886]

21:12 on vista they changed the backend renderer and splitted compositing from rendering while maintain backward compatibilty they did it so good you didn't notice.(or noticed with a lot of drivers failing near release,,,,)

[@ocsanik502]

ah i thought the guy was referring to the DOS to NT transition where the system became a lot less modular with core drivers

[@evildragon1774]

is it safe to assume that microsoft's "wayland" has matured?

[@fu886]

@@ocsanik502 it is not the removal of the 9x codebase its the move to vista. its not only matter of maturing, when they moved it around they moved it while keeping stabs for the previous system, with wayland they decided backwards compatability doesnt matter to beginwith

[@igordasunddas3377]

​@@evildragon1774you mean "dwm"? That's I think what it is replaced with and yes, it works very well, but we didn't have a choice. However: WinForms/GDI was still supported even though WPF/UWP was introduced.

[@0xc4ae1e5]

Yeah, and the compositing engine is called DWM :) But the actual windowing system and rendering is in the kernel. @@evildragon1774

[@thedoctor5478]

Didn't the wayland project start in like 2008, and it's still not stable? Sounds like the sort of project that will never be a good idea to rely on.

[@BrodieRobertson]

Replacing 30 years of workflows takes some time

[@qwelias]

wake me up when entire screen sharing starts working on web under Wayland

[@thock_enjoyer]

another 25 mins wayland video less goo

[@Qohist]

me sad

[@MenaceInc]

Gotta farm dat content 🙃

[@orbatos]

Not even watching at 3x.

[@rizkyadiyanto7922]

tldw

[@ivailogeimara]

"Linux users don't run random applications". This is the dumbest thing I've ever heard of. Of course they do. But the problem is not just with "random" software. It's with every closed source software. Running only open source software is very unrealistic in post-2019. And every closed source software can be considered a malware/spyware because you really don't know what it's doing. But you absolutely need to be able to run it in order to have a fully usable system. (I for example run Viber, Discord, Intellij Idea Ultimate, games, and others). So some kind of security not allow apps see and do whatever is absolutely needed in modern Linux desktop.

[@__christopher__]

If you run the applications locally, can there be any security anyway? I mean, the processes have direct access to the computer, restricted only by the permissions of the user. Well, unless you configure something like SELinux or AppArmor, but how many Linux users do that for downloaded stuff?

[@djentledjosh]

​@@__christopher__fedora has SELinux enabled by default. flatpak+portals also works pretty good and is continually improving. steam runs games in a bubblewrap container, which is the same tool that flatpak uses to sandbox applications.

[@uis246]

Why 2019?

[@Mankepanke]

How do Window users even survive, then? They run almost 100% closed source, yet it's not mad max out there.

[@speedytruck]

Well, I use Fedora which is only open source so...

[@knightrider585]

I remember really enjoying using Word Perfect on DOS back in school. This version of Word Perfect was probably my favourite word processor ever. Sadly it was written in 8086 assembly language and now can only be run as a curiosity in a DOS emulator. Software dies.

[@rovano10]

The local (Czechoslovakia, 80286?) most effective text editor I have ever seen and used. kzbin.info/www/bejne/gGLXc62Lf8Rpl9U

[@spatiumowl]

Great thing about Wayland being a protocol is if something like wlroots or kwin becomes an "unmaintainable mess", it's much easier to write something more maintainable and not be worried about a major migration since it's still the same protocol

[@thingsiplay]

Because a protocol is basically a standard, a documentation, like HTML is. X11 is an implementation without a standard and without exact documentation. I have to say, the more I learn about Wayland, the more I like it. And now I understand why my old thinking of "Wayland should be named X12", because Wayland is fundamentally different. And that's good.

[@KettLovahr]

Sounds like it's a little more future proof, too, at least in theory.

[@obake6290]

same thing applies to X. X11 is a protocol. Xorg was an implementation of the X11 protocol. Just so happened to be the only implementation that mattered, but that's beside the point. This is why I always thought the "Wayland is just a protocol!" arguments to be incredibly stupid. Support Wayland all you want, like I do. Just use arguments that make sense.

[@electronics-girl]

@@thingsiplay Are you kidding me? X11 is a protocol, designed from the ground up for multiple implementations to be able to interoperate. And it's incredibly well-documented. Back in the day, it was a multi-volume set of printed books, but today all of those books are available online. (Well, the protocol itself was Volume 0, and then the other volumes documented higher-level things.)

[@electronics-girl]

@@obake6290 That's one of the things I hate about Wayland. Its supporters make arguments that don't make any sense. Basically, they're saying "I don't know the slightest thing about X11, but trust me, I know how to do it better!"

[@alexanderdelguidice4660]

xdg-placement has just been moved to the ext namespace. I'm slightly disappointed but I also get it, it is very controversial. I just really hope that it becomes an actual protocol and is imlemented by more than one compositor.

[@iodreamify]

What are the consequences of this? Does this mean not all compositors will be required to implement it?

[@BrodieRobertson]

Even if it's in xdg not all compositors will implement it anyway, they're supposed to but even Weston skips some

[@alexanderdelguidice4660]

My logic is that if it's in xdg, compositors will be more likely to implement it. I never said that all compositors implement them.

[@hwertz10]

@@iodreamify They aren't required to implement it. xdg namespace is described as "established for protocols letting clients configure its surfaces as “windows”, allowing clients to affect how they are managed" and ext "established as a general catch-all for protocols that fit into no other namespace." So from the description, window position is definitely xdg. But the developers that decide what goes in xdg namespace were quite adamant that since some compositors use tiles (like 4 items each taking 1/4th of the screen) or other setups where you don't have draggable and resizable windows, that therefore specifying window position doesn't make sense. Period. What's the use case for this anyway? If your app requires window position to be set, it's old and obsolete like xorg, so suck it up. So anyway, no window positioning in xdg namespace. (In my opinion the more sensible view would be to just have compositors doing conventional movable and resizable windows to allow positioning and size to be set, while those that do tiles or whatever don't, if your app requires it to work it simply won't work in that type of environment but works fine in the more conventional ones.) So it was not going to be put into wayland specs at all since it was not going to be allowed into xdg namespace; it was allowed into ext namespace as a compromise.

[@taukakao]

You will probably use a compositor that will implement it. Gnome has a different philosophy for stuff like that but you don't have to use Gnome.

[@myownfriend23]

Oh no. You gave birdie publicity. Now he's never gonna shut up :-( BUT I do think it's worth mentioning this issue that he opened because a lot of people repeat them and it's good to have someone respond to and correct them publicly.

[@keit99]

That birdie? From phoronix? *runs and hides*

[@myownfriend23]

@@keit99 Yup lol I can't believe his second account hasn't been banned for banning evasion yet

[@Beryesa.]

Imagine such a bird that you don't want to hear lol

[@softwarelivre2389]

Long live birdie. Btw: he wouldn't shut up regardless

[@splitprissm9339]

The "Xorg like architecture in Windows 3.0 to XP" probably refers to the message queue and damage-event based architecture (Windows sees that some of your application window was painted over by eg a window overlaying it, and will queue messages into your applications event queue telling it to draw that area again as soon as whatever covered and overdrew it gets moved away. The GetMessage/TranslateMessage/DispatchMessage/DefWndProc loop that even most Windows GUI toolkits (eg MFC or Delphi VCL) abstracted away from the developer.). BTW, did you know every GUI control (eg a text input box, a button, a set of radio buttons...) that can have keyboard focus is technically its own window in classic windows architecture? .... I still think Artem is right - if "community rules" are used to silence user feedback, even downright brigading of bug trackers and forum would be a very justified response. Respect users if you want users to respect you. He seems to be from a time where compatibility/portability was considered a responsibility of developers especially in FOSS. And I can agree with that view.

[@BrodieRobertson]

I've so far heard 3 different answers to the windows comment lol

[@lucas7061]

birdie is being "silenced" because he's purposefully trying to stir up drama and create problems using an issue that was closed because it's been repeatedly explained it's outside of the project's scope. That's without taking into consideration the fact that not everything he mentions is a Wayland responsibility. Acting like a douchebag online won't take you anywhere. Hell, even in Phoronix he is the forum's punching bag because people always try to explain him what's wrong with his arguments and he proceeds to completely ignore it and repeat the same tired stuff.

[@guss77]

I can't shake the feeling that the reason Wayland is not a server is mirroring the decision of GNOME developers deciding to stop building their desktop environment as a set of libraries that other desktop environments can build upon - Wayland is not the server project because the GNOME developers didn't want to build a generic implementation that will have to support other desktops - they don't want to be forced to support server side decorations, notification areas, desktop icons, etc.

[@MrGamelover23]

Yes and no, you're right for the wrong reason. Basically, the problem with X-org is that it does too much. This makes it really difficult to maintain.

[@guss77]

@@MrGamelover23 that's not a very good reason to not build any central server for anything - Apache still works with a very lean core where all the complicated stuff gets relegated to modules. Same with nginx. Compare with all of the small servers (lighttpd, varnish, etc) that do very little and as a result are very good at what they do but are only suited to specific use cases. And no one talks about abandoning Apache or nginx - that are very successful projects and are very popular. I can see a timeline where there's a core Wayland server and a set of modules that are optional that implement protocols. That would allow the people that are interested to cooperate, on say - server side decorations, to do so while still maintaining a small core with a lot of functionality overall. It sounds like the problem with X11 was a monolithic software design that didn't scale but there are other ways of doing it and successful software projects exist that are either highly modular (as discussed above) or manage a monolithic design in a scalable way (the Linux kernel comes to mind).

[@fuseteam]

Consider this, despite how often this issue comes up; why has no one stepped up so far to keep xorg going? It is not for a lack of trying ;)

[@guss77]

@@fuseteam I'm not saying "lets revive xorg". I've been fighting in the trenches of xorg as well and I know how that feels. I'm saying there was a chance to make something better - but still a shared code space that reduces effort duplication, and I feel that chance was wasted because GNOME devs didn't want other people to play with their toys.

[@guss77]

@@MrGamelover23 This can't be a reason why no one should ever make a server. I understand the Xorg is unmaintainable - I had my share of those, but there are successful very large servers out there that have managed complexity well - take for example Apache or nginx: these are huge projects with tons of features where the core is very small and they use plugins to extend functionality and keep the code maintainable. I can envision a Wayland core server that implements the barebones minimal way to display windows and plugins can provide protocol extensions. I'm sure KDE and XFCE devs wil love to cooperate on the server-side decorations implementation and hammer all the details between them and not bother GNOME devs. There are various ways to manage complexity, and some are very successful, either by highly modularizing the code base (like the above mentioned Apache and nginx) or other ways (the Linux kernel is another good example). Saying that the X11 project failed to manage complexity so any other way to build a shared code base for display servers is doomed by definition, is patently incorrect.

[@rhysperry111]

Good ol' birdie. I'll never understand that man, but it's fun to watch the chaos he causes

[@DePhoegonIsle]

Things like this make me glad to stay with windows. Also, there are more real-world examples of spector & meltdown being used on Windows than there ever was for Xorg security flaws being exploited. Seriously though, there are merits to having more 'security' in computer systems... but there is a point where it breaks feature sets and just becomes aggressively anti-user. How exactly are they handling redrawing/overlaying of applications in wayland? I know there are android apps that I use that use it, and when it comes to anti-cheat .... how exactly are you going to give the increased access to those toolkits?(Or are you going to build a bypass just for anti-cheats, or further break them meaning ... you set linux back)

[@fuseteam]

That's not what wayland is, it is __just__ a protocols :p Android does not use xorg nor wayland, but 'overlaying windows on other windows' works fine under that is what brodie meant with the lock screen example Anti cheat also has nothing to do with wayland, it just works when the developers want to support linux (except kernel anti cheats, but we don't talk about that) see the website 'are we anti cheat yet)

[@Souls4Roca]

Brodie password managers like 1password do write into the password prompts of other programs

[@BrodieRobertson]

The issue isn't that it's redrawing over it to pretend to be the password prompt when it's actually another application

[@MinaSchloch]

I love how since that one ToT you discovered the word "bikeshedding" and I have still no idea what it means.

[@blinking_dodo]

[originally BSD, now common] Technical disputes over minor, marginal issues conducted while more serious ones are being overlooked. The implied image is of people arguing over what color to paint the bicycle shed while the house is not finished.

[@conarius13]

​@@blinking_dodo To expand an your explanation: It stems from Parkinson's law of triviality. His thought experiment was a committee of nuclear engineers to plan and design a nuclear power plant. While most of the discussions on bigger issues will take some time and resource but will eventually be resolved in a civil manner, the more trivial it gets the more resources they will waste on it and everybody wants their solution and nothing else. They would then argue about a bike shed for the employees a lot of the time. Basically the law is: the more trivial something is to discuss and reach a conclusion the more people will argue about it since they really wanna show that it was their idea that was chosen.

[@davidh.4944]

@@conarius13 It's not just because people want "their" ideas, but because the more trivial the topic, the more people feel emboldened to participate in the debate. Decisions on the design of the reactor core take little time to resolve because there are usually only a small number of experts who know anything about it, and everyone else shuts up and defers to their expertise. But bring up the design of the bike shed, and suddenly everyone is an expert, or at least thinks they know enough to offer up their opinions. Everyone wants to feel like they are contributing, and they jump on the chance when they think they can. Note that this also can have adverse effects in the opposite direction, too. Critical decisions are sometimes not given as much scrutiny as they really should.

[@jeremyandrews3292]

I'm thinking the best solution might be to just bring in other libraries into applications and compositors that can do what Wayland won't do. Wayland won't implement it? Well, then use this other library that does. I mean, if we have dbus for IPC, Pipewire for audio, libei for input devices, and more... then why not just keep going until we have all of X11's functionality split across multiple libraries? Then all the crazy people who are happy with the base Wayland API can just build against Wayland, and everyone else can use it with a ton of extra libraries so they can have a good experience. Maybe at some point we'll have enough different libraries that reimplement pieces of X11 that at some point Wayland itself won't even be needed as a framework.

[@GrzesiekJedenastka]

You pretty much described what Wayland is supposed to be. It's separation of concerns. The only thing you are wrong about is Wayland being "the framework" that won't be needed at some point. It's not a framework, it's just one more part of the equation - specifically one that's supposed to be used to _draw stuff on the screen._

[@electronics-girl]

Yeah, the problem is pretty much the exact opposite. The problem is fragmentation. Everyone going off and doing their own thing is a mess, and makes it hard to write programs that target Linux/UNIX as a whole, since every desktop environment will be doing something different. X11 provided a much better baseline to allow programs to be able to interoperate.

[@necuz]

Still on Xorg because Nvidia. Yeah, it's gotten better, but still worse on Wayland and no benefits.

[@FrankHarwald]

Still on Xorg because Raspberry Pi 3 being too slow on Wayland.

[@stickfigure31]

@11:43 just felt the need to say "I use iceWM BTW!", I tried it out on my old laptop because gnome started becoming unresponsive and ended up switching my gaming rig to it. I also like the default config of TWM when you install openbsd with just it, but I can't figure out how to replicate it on Linux. In both cases they work really nice when displaying on remote X displays over the network. I'm now interiged by the others listed and will probably try them.

[@Scoopta]

Fun fact...dwm isn't even the windows WM, it's just the windows compositor, the windows WM is actually in the kernel...for a reason...

[@dsihacks]

You mean the Win32 Subsystem? Pretty sure ntoskrnl.exe does nothing related to being a display server.

[@bes12000]

My concern is things not working correctly that were working fine on X11 once they fully switch to wayland if something needed was left out of the code..just because people think it's not needed.. I struggle enough to get certain apps working on linux that were made for windows if there are no alternatives, once I get them working it stays working even through updates, what if wayland breaks that stuff?..

[@jeremiahgriffinphotography9436]

This is exactly why I unfortunately had to quit linux for personal use. I still use it for headless stuff, but I just got tired of stable distros breaking my software and worklflow without warning every other update. Wayland was the last straw. I don't consider myself to have an esoteric workflow in which case I would excuse it. I love the performance and lack of shitware as compared to windows but in exchange for an unshittified OS, I would have to either be stuck with rolling or standard distros that break without cause or LTS with software so out of date they're incompatible with my work. It's a pain.

[@LostieTrekieTechie]

​@@jeremiahgriffinphotography9436Linus has that famous line about "not breaking userspace", but it turns out user programs rely on software that isn't in the kernel.

[@myownfriend23]

It's not that Wayland and it's developers are saying that applications shouldn't have certain functionality, just that Wayland shouldn't be the thing to provide them. For example, in X11, any client could see the contents of any other window so the mechanisms for doing that were used for screen and window capture. Clients could also see all keyboard input regardless of whether it was meant for them so clients used that to create global shortcuts. Both of those things are against Wayland's security model but since Wayland replaces X11 in the display stack and people were used to doing those thing via X1, they felt like Wayland developers were taking the stance that applications should not have screen or window capture or global shortcuts. That's not the case. They just saying that Wayland isn't the right place for them. The display server protocol isn't the only thing that determines what an application can do. Client's can and do interact with multiple APIs and protocols to use or provide different functionality. Wayland and X11 exist as ways to allow clients to run alongside each other in a windowing environment. The windowing environment exists as a way to both manage windows and make sure that you input gets sent to the active applications, If you think about how window capturing and global shortcuts work, they don't really have anything to do with windowing at all. Global shortcuts kind of circumvent the input routing of a windowing environment. It doesn't matter if the client's the in-focus window, a background window, or has no window at all, it just gets input. The process of window capturing literally requires viewing another client outside of the context of a window manager. It doesn't matter if the client is hidden, minimize, or even full-screen (not in a window at all), all it's really doing is using another client's front buffer as a texture for some process. Because they have so little to do with windowing, someone else can provide that functionality in a way that works in both Wayland and X11 and something else does: Portals. I don't think what functionality you're talking about specifically, but I wouldn't worry about it. The things that Wayland's devs omit from the protocol are omitted because they don't think Wayland is the best place for those things to be implemented, not because they're saying that applications should never be able to do something.

[@PeakKissShot]

Nothing was ever working fine on x

[@damianateiro]

​@@myownfriend23They have been scratching for 10 or more years and refusing to implement basic things and crying if they are told what needs to be done. Wayland breaking is the complete responsibility of its developers.

[@rigen97]

Referencing Windows in this discussion is wild because the reason Microsoft can just drop a single "universal" implementation for window manager and compositor etc is because Windows users literally had no choice but use them if they want to use Windows. Most people that use Windows suck up the complaint about its quirks because they usually don't know that things can be better.

[@abitterberry2149]

I've been running Hyprland on my laptop for about a year (It's your fault btw). At first, switching from X11 was kind of annoying, but now I'm really enjoying it now. The only thing I miss from my previous Qtile these days is its generous choice of layouts. Wayland is definitely good in 2024, enough to use it on my work computer.

[@RandomGeometryDashStuff]

11:46 icewm comes with opensuse tumbleweed

[@vanodon2257]

These issue andies are so funny. I ahted wayland found a minimal compositor and just hacked in fixes to my problems. System wide Portal, no more needing developer to implement wayland support everything has access to everything like on x11. And later I added apparmor config so this wasnt completed opening my system. If you have a problem the code is there fix it....

[@Kneedragon1962]

So that's Birdie. I've not had the joy of reading Birdie's posts myself, I have avoided looking for them, but as a long time user of Phoronix, I've seen an awful lot of people refer to him. He seems to be just as loud a persistent and negative on Phoronix. I get the impression that like Richard Stallman, Birdie has a way of being difficult. --- by the by, there's a question. How is Mr Stallman's health? Last I heard, Stallman had declared cancer. Since then, King Charles has forked the project, but I haven't heard anything about the head GNU.

[@rflsms]

It seems people just don't get what Wayland is.. (even in this comment section after many many wayland videos)

[@Soitisisit]

I understand perfectly! Wayland is a mistake. :)

[@rflsms]

@@Soitisisit if only there were another way...

[@blinking_dodo]

This will be a feature i will *not* be implementing in my PHP display manager... 😂

[@kebien6020]

Why do you need a display manager specifically for P*rnHub Premium?

[@그냥사람-e9f]

birdle try not to beg Wayland devs to turn Wayland into X12 for 30 Planck time challenge (literally impossible)

[@Jmvars]

I'm using KDE Plasma on Wayland but also XWayland. Works wonders, never had issues except mpvz likes to open a second window, what seems like a non-Wayland window since one of them has the Wayland icon while the other has the mpv icon, and only the Wayland window has smooth playback.

[@ssokolow]

By "Well-established competing desktop OS'es work exactly this way", he's referring to how, for example, Windows's compositor, dwm.exe, is separated from the OS's graphics subsystem, can die without taking down your Win32 apps, and yes, according to the newer columns on Wikipedia's "List of alternative shells for Windows" page, it CAN be replaced by third-party options. (I ran Litestep on my Windows XP before switching fully to Linux.) As for "none of them support the kind of custom desktop environments that...", Windows has always supported that. Windows 3.x's architecture was literally the same "the desktop session runs as long as this anchor application does" as X11 session management. It's how Windows 3.1 had stuff like Packard Bell Navigator, TabWorks, Norton Desktop, and Microsoft Bob running instead of Program Manager. Microsoft used it themselves in Windows 9x to allow companies to delay retraining by bundling copies of Program Manager and File manager that IT could swap in until Windows XP... SP1, I think? From what I remember, the reason the flood of alternative shells dried up was partly that Microsoft demanded that OEMs ship Windows 9x with the stock shell if they wanted to retain their access to the discounted licenses and partly that Windows Explorer was mac-like enough to be "good enough" that people wouldn't pay for complete shell replacements like Norton Desktop or Central Point Desktop anymore.

[@whtiequillBj]

I use X11. because nVidia doesn't have good support for Wayland. I usually only get a single screen of my two screens working under Wayland. Today I tried again and my screen was mirrored. which was an improvement... usually there is no recognition of my second screen at all. It just isn't much of a improvement. Also it only shows a single 1024x786 display while I have two 1920x1080 screens. I'm using KDE Plasma on Gentoo.

[@sukidable]

You might have a better experience with hyprland

[@cameronbosch1213]

3:54 *cough cough* GNOME Shell, *cough cough.*

[@no_name4796]

He said "sensible bar", not just "any bar" lol

[@cameronbosch1213]

@@no_name4796 I know, just getting my two cents out there.

[@cameronbosch1213]

@@no_name4796 I replied to your comment, but it seems to have been culled by KZbin. What you're saying is true though; GNOME isn't really sensible for former Windows users or even macOS users because of the lack of AppIndicator support without an extension.

[@enemixius]

It can be enabled. I do think it's stupid to have it disabled by default though.

[@taukakao]

@@enemixius I think it's very sensible to not have it enabled by default. Tray icons are only really used for closing an application and displaying some state. The first issue is solved very well in Gnome 45 and the second one is solved for many states. (But yes, not for all states, which I can see is annoying some people) I think it would be sensible to have it as a setting instead of an extension though, since for some people it can still be a valid use case.

[@MonkeeSage]

Brodie giving Wayland a reach around again, they should really pay him.

[@lithiumflower31337]

I really don't' understand Linux developers, "Maintaining X is impossible!" while also "Let's just re-write all the graphical code for 15 years and force everyone to re-write their graphical programs or be rendered unusable! So much better!" And if X is so unmaintainable, then how does something like Linux get maintained? Surely they're not trying to say painting the screen is harder than writing a whole operating system+necessary utilities, are they?

[@BrodieRobertson]

Maintaining X is impossible because nobody actually wants to do it

[@alpheusmadsen8485]

For the record, I use XOrg KDE and am trying to come to terms with Awesome WM (which is, necessarily, XOrg right now). I have explored the possibility of doing Wayland with something like Sway, but I'm not convinced that it's "ready" for use. At the same time, I'm a little nervous over the effort I've been putting into Awesome WM, because I'm aware there's currently no interest in porting to Wayland, and I realize now that there may be good reasons to just start creating something from scratch. I am not entirely convinced that Wayland will "win out" in the end -- legacy standards are a big hurdle to overcome! -- but I'm also not entirely convinced that sticking with XOrg is the "way forward", either, so I'm going to keep an eye on how things progress. And yes, that means a lot of "duplication of effort". Having competing standards isn't necessarily a bad thing, though! In the end, either XOrg or Wayland will win out, unless someone comes up with a fantastic third option that edges out both of them in their quests for users -- and in the end, to the extent that all these things improve, we *all* win!

[@night_h4nter]

it's not about winning or losing. xorg shall be abandoned in a few years, and that's it. it'll probably keep working for a while, but won't be supported, so everyone shall have to switch to wayland if they wanna use *nix

[@Spencer-wc6ew]

I'm curious, what do the Wayland devs actually "develop"? Like what coding do they do if there isn't something like this? I never really thought about before watching this.

[@Poldovico]

A lot of them develop various bits of desktop infrastructure in other projects, but as far as their contribution to Wayland itself, the main product is protocol specifications. So, a bunch of talking about how things SHOULD be done. When they reach an agreement, they know how to do those things in their other projects in a way that is standardized, compatible, and unavailable in GNOME. Unless they get NACKed by Weston because car radios don't need it.

[@softwarelivre2389]

​@@Poldovico"and unavailable in GNOME" lol

[@night_h4nter]

@@Poldovico lol this is an exceptionally well-written answer, got nothing to add to it xD

[@didjidks]

Aside from Wayland specification, they also maintain the standard Wayland library (libwayland) and few Wayland related tools and libraries. Also Weston I guess?

[@XoaGray]

Honestly, I'm still using X11, because when I use Wayland I have too many issues. But I have used Wayland and can see why people do like it. I've also been told that KDE and Wayland really don't get along right now and that's a lot of the issues. So hopefully that's better in the future.

[@CharlesGriswold]

I'm running XFCE on Linux Mint, so I'm pretty sure that I'm still on X11. Honestly, I'll just use whatever gets shipped and not worry about it too much.

[@DanielNerd]

personally i use X11 because i like playing games. but I definitely wait for wayland to be good enough for games so I could jump ships! maybe when I get a thinkpad i'd install wayland on it. some of the wms look really fun to mess around with! and some have features that sounds really useful.

[@rogo7330]

I'm sorry, but "people running random stuff from AUR or pipe curl into sudo bash" is stupid. I can do that without graphics, Wayland will not stop root-priveleged program from directly reading anothers program memory. Also, I think Wayland protocol in it's core is just too schizophrenic. All I want in Xorg is the ability to restrict apps by methods that already exists, like file groups, which are POSIX-available everythere and don't need some stupid another-one-server-that-everyone-need-to-run-(with-root-priveleges). If Xorg is unmantainable, just start with brand new X11 server from scratch, throw away stupid stuff (like assumptions that server will be running other network, but still let it happen through 3rd-party hacking).

[@speedytruck]

so birdie isn't just annoying on phoronix...

[@thingsiplay]

* Brodie: famous, Wayland shill * Birdie: infamous, X11 shill On the other alternate universe, I probably named Birdie first in this list.

[@Desmaad]

4:05 The Korean character set is called hangul.

[@thecrow3461]

Currently on XFCE so on X11 for now, will switch to wayland when XFCE releases a version that supports it

[@klaudyw3]

I don't even run linux, i haven't in a couple years. Still along for the ride/drama though!

[@cameronbosch1213]

What do you run then? Microshaft Winblows? CrApple macOSn't?

[@MegaManNeo]

@@cameronbosch1213 Well, this person has to run something if it's not Shitnux. Before you blame me, same treatment for everyone.

[@klaudyw3]

@@cameronbosch1213 I run Bighard Wall. While i technically need it for work, the main reason I use it is gaming, although i don't even do that too often anymore. It's simply more convenient for me. I'm not in the mood to dual boot or run a VM just for games, some of which don't like VMs (rootkit cough cough). I have faith in Bighard that they're going to do something stupid in the near future that will make me switch, so that's something.

[@orbatos]

​@MichaelDustterhis to tell everyone you never really used it. X has always been a pain who's open devs have hated it since the early 00s

[@JTCPingas]

@@cameronbosch1213 I heard about Microshaft Winblows 98 through LGR.

[@sweetbabyalaska]

I think its funny how writing an X11 window manager using 99% of the behemoth that is the X11 library is perfectly fine for some people, but using wl-roots in a similar fashion is suddenly unheard of and outrageous lol

[@damianateiro]

It will be because x11 gives you more freedom, and you have a solid base, in Wayland you have to create all the libraries from scratch, or use the same ones as someone else. That's why all composers are the same and most of them are abandoned, add to that there is not a day in which Wayland falls apart and its developers are crybabies

[@myownfriend23]

@@damianateiro FREEDUMB!

[@ocsanik502]

I think it's mainly just developers not wanting to learn new tooling, which is perfectly understandable because I don't either. It's a chicken and egg problem, for things to get better you need a driving force, that driving force is usually users that want it, to get users onboard you need it to be getting better; It's like a exponentially increasing curve and we're all stuck at the bottom waiting for it to reach that tipping point when it both attracts developers and users enough to reach a point where it can reach every need presented.

[@notuxnobux]

Sounds like something someone that has never written a X11 window manager or used wlroots would say. All X11 client library does (xcb) is to connect to the server and send data. The x11 library itself is 0% bloat. It doesn't contain any logic. You can write a x11 window manager in around 20 lines of code and all x11 software will work under this. Screen recording, global hotkeys, etc. Wlroots does 1 million times more things than that and so does wayland compositors that use wlroots. You need to write way more code with wlroots to get the fraction of functionality of a very minimal X11 window manager.

[@myownfriend23]

@@notuxnobux I think the smallest Wlroots compositor is a few hundred loc which isn't all that bad considering it's a window manager and compositor. I thought the smallest X11 window manager is just over 100 loc.

[@AlexandruVoda]

Hey @Brodie, care to make an episode about niri at some point? I really like the scrolling-tiling desktop concept and think it deserves a bit more visibility. Currently niri (Smithay compositor), Karousel (Plasma) and PaperWM (Gnome) are the only such active projects with previous projects like Cardboard (wlroots compositor) being archived.

[@night_h4nter]

mind elaborating on what's the workflow like on such wms/compositors? like to me it seems quite weird and quite slow

[@AlexandruVoda]

@@night_h4nter well, just like on traditional tiling WM you make use of keyboard shortcuts both to move around and to move stuff around. Movement however is conceptually simpler because horizontally you move between apps and vertically between workspaces. Here is an example from how I use it. I like having several messenger apps (4+) open because I don't like having to look at my phone. Because of the way these apps are made I can at most have 2 at the same time on screen. Making the windows smaller is not feasible. But I want them all out of my way. On a traditional tiling WM they can not all share the same virtual desktop unless you use tabs like on Cosmic. On a floating WM they can all share the same virtual desktop but then you have to mess with the Alt+Tab switcher or some launcher. On a scrolling-tiling WM, they can share the same virtual desktop, be as large as needed and I just go left and right to switch between them and up and down to get to the right workspace. It is not an optimization in terms of key presses, it is an optimization in terms of cognitive load. It's kinda like putting everything on a grid. Grids are easy to comprehend.

[@PauEstalella]

That we are still talking about pixels at application level is sad. Then we have all the problems with HiDPI and combining monitors of different resolutions.

[@alirahimi92]

13:00 the idea may work but it's not efficient.

[@dexterman6361]

The one thing I definitely dislike is that without a uniform framework that interprets gestures, every app has to reimplement it, and everyone is going to do it differently. And here we go again with the inconsistent desktop

[@nmjerry]

I would love to see a version of Linux that would take up as a fork i386SX linux, and be updated as much as possible, and made to run on bare metal on as many machines as possible. Of course it would have to rely on task switching, unless it could turn on and off co-processing as needed which would be more ideal. I think for most application code forking a process is the same whether it is forked to a task or a co-processor.

[@rancidbeef582]

I will continue to use X11 for as long as it will run because it does everything I need it to. I got X working on Linux way back in the early 90's and it's only gotten better since. One thing required for me is the ability to remotely run an application over an ssh tunnel (-X). Does that work in Wayland? Not unless you also run something like XWayland, so I've heard, which just adds another layer of bloat.

[@GrzesiekJedenastka]

There is a Wayland-native solution for this, I don't remember the name though, sorry.

[@eliminmax]

I've hear about waypipe, but am still on X11, at least until Xfce 4.20 makes it into Debian stable or my plans change, whichever comes first, so I have not tried it yet.

[@electronics-girl]

There is Waypipe, but it seems to be a second-class citizen, and even it doesn't work if the two machines have different endianness.

[@Lambda_Ovine]

Maybe Birdie could implement the X11 Wayland compositor frankenstein if he dedicated the same energy to development instead of trolling Wayland devs

[@Psychx_]

Birdie has been deanonymized. Excellent!

[@BrodieRobertson]

He uses the handle on both it's not a hard problem to solve lol

[@Psychx_]

@@BrodieRobertson No need for false modesty. You earned this success! Enjoy.

[@AlexandruVoda]

I would argue that, contrary to what you said, remote desktop is not solved, it is very much still in development. For example KRDP is still in alpha with the last release being from august 2023. You also no longer have much choice in the protocol. AFAIK, Plasma and Gnome went with RDP after some switching back and forth, wlroots went with VNC and if you want something else like Spice you are SOL. Broadway feels like a possible future 3rd option. Also AFAIK, there is no easy way to get Ogon Project going for RDP. All of the puzzle pieces are available but they are not yet fully assembled.

[@supercellex4D]

Really, RDP? What the fuck is this, Windows?

[@AlexandruVoda]

@@supercellex4D Yes, RDP. So far, for Wayland the only protocols that had any work done for them are RDP and VNC. Of the two RDP is by far the better protocol. Both RDP and VNC have FLOSS implementations for both server and client and have good clients on all platforms, but RDP is clearly better. NX, X2Go are not applicable since they are tied to X. SPICE could become an option, but no work has been done for a Wayland compatible server and the client situation is just sad. Stuff like RustDesk does not serve the same role as RDP or VNC. And no, SSHing in and doing things in the terminal is not an alternative to RDP. So what alternative to RDP do you propose?

[@Thanatos2996]

I’m pretty sure when you released this video, keyboard LEDs weren’t yet implemented in DWL, and it still hasn’t merged in DPMS support. It also needs patches for touch and tablet support, because those are not at all free with wlroots. Yes, wlroots eases the pain of implementing wayland to an extent, but compositors still have to do way more than WMs, and leaning on wlroots adds a whole new set of issues since they make breaking changes with every single release. The amount of duplicated work among the devs of various wlroots-based compositors adjusting to wlroots changes alone is enormous, and that’s after they’ve all duplicated the work of setting up all of the stuff that the X server would handle under X, all to support the couple hundred lines of unique code that give the various compositors their unique behaviors. The model is terribly cumbersome and inefficient compared to X, and I wish there were a more modular option.

[@CharlesGriswold]

Telling a bunch of volunteer coders "you need to create and maintain this massive project because I want you to" is entitled, to say the least. No. No, they don't need to do that. They don't *need* to do any of the coding they're doing, and demanding that they do even more on top of what they're already doing is really, really not a good look.

[@NFvidoJagg2]

wouldn't the reference wayland compositor be Weston? sounds like if birdie should just fork Weston and implement his wants. problem solved

[@BrodieRobertson]

It's more that it's something he wanted upstream to do and not just a reference compositor but something akin to xorg

[@mysterynad]

Did this guy seriously compare Wayland development to the Windows compositor? He must have forgotten that Windows is built by a company with a market cap in the trillions, while much of Wayland is the result of what people do for free in their spare time.

[@supremesonicbrazil]

Still using X11 due to reasons I explained before but don't remember in which video, but mainly due to not knowing if my Steam games will work on it without any additional MacGyver-esque workarounds. It's not like I don't want to either, I just don't know how things are nowadays and/or if it's gonna be just like Pipewire where I had to wait until it was 1.0 to consider it "properly usable" (or for all the programs I use to natively use it, which might be "never because there'll always be this one outlier in the end to which you'll have to use XWayland or something" if I'm looking at this through the right angle).

[@GrzesiekJedenastka]

Just try if it works for you lol. I can definitely say games I tried work perfectly. It's supposed to be the whole point of Xwayland, that you can switch and programs will just continue to work.

[@robw3610]

To be honest, I feel like Wayland being "not ready yet" is precisely because it is a protocol, with no reference implementation. And honestly I think that ship is long since sailed. If it was going to be done, it should have been done more than a decade ago. Instead we seem to have different projects working on a bunch of implementations in parallel with different stages of "maturity", many of which refuse to implement the entire "standard". (Looking at you GNOME) As an application developer this makes it really hard to target Wayland because I have no assurance of what parts of the spec are going to be implemented because of what seems like a fragmented mess.

[@Wicked0547]

Whoa whoa but they said Wayland is the future. How could this happen?

[@Chafmere]

I wish I could run wayland. My main tool still not 100% supported. But in a few months will be and I will hopefully make the final switch.

[@Rand0081]

I used Sway for 2 years, but now I use Plasma Wayland because I'm lazy. I will restore my old config someday. I couldn't make qt virtual keyboard under Wayland no matter what. So my sddm runs under X. My suggestion is if you can run Wayland you should, and respect people who still need X for whatever reasons.

[@thingsiplay]

17:09 That's what she said.

[@Bigoukun]

I have to PCs, a laptop with Gnome, and a desktop with KDE.

[@opuser1]

I just wish freerdp would not crash the kde taskbar or restart the gnome session when using two monitors and you click out of it.

[@ocsanik502]

They the guy mentioned Windows I think he was actually meaning to reference the DOS to NT transition with Windows which for most people was between Windows 98 and Windows 2000 (even though Windows NT was actually already available for companies all throughout the lifespan of Windows 3 to Windows 98). Prior to Windows NT they used a really strange architecture where they basically tried to successfully multitask a standard DOS system and in the process they made something kinda like Xorg in the sense that you can easily do open heart surgery on it using components from other projects at runtime with minimal effort (ex: in Xorg you can use compositors separate from the window manager, in DOS Windows you can swap out your EMS (emm386, jemm, qemm) and XMS (himem, himemx, xmgr) drivers to something other than what Windows shipped (including even hybrid XMS + EMS drivers like JEMMEX), and often in both cases this isn't just simply stuff that works, it's often encouraged or preferred to use over the default), whereas after the change to NT they switched to a one size fits all architecture. This is all to say, that guy is wrong, one of the primary driving forces for Linux users is the fact that Windows employs a one size fits all architecture for users, and there are people that simply don't want to deal with it anymore. So trying to suggest re-implementing the same concepts is a incredibly uninformed stance to hold as Linux isn't trying to just be "Windows with better idle performance".

[@igordasunddas3377]

Perhaps he meant the transition from Windows Forms up until Windows Vista to DWM in Windows Vista (which enabled translucent dialogs, animations etc.), but it was mostly backwards compatible and except for the drivers due to WDDRV or whatever it was called and then failing frequently back in the days, it wasn't noticable.

[@jeffreyjoshuarollin9554]

You’re right, though if you read his posts on Phoronix,* to listen to him you’d think Linux was somewhere between its state in 1994, and “a complex pile of interrelated dependencies which barely work,” to quote - if not verbatim - the guy who wrote Ars Technica’s review/hatchet job on the Pi 5. Incidentally, this makes me wonder whether he has used Windows lately/at all. Indeed given OS X and iOS are written on a Unix-like codebase with, at this point, *many* proprietary additions, you could delete “which barely work” from that sentence and still be right - and some would say you would be, even if you didn’t delete it. *which might be hard to find, since he’s apparently been banned in the time (a few years) since I stopped reading (much less contributing) to the comments section on Phoronix, precisely because guys like Birdie make it a dumpster fire.

[@supercellex4D]

No, he’s talking about Aero. But Compiz and Picom exists so he just schizohallucinated there being an issue in X

[@GibusWearingMann]

Calling Windows Aero a "beautiful success" is the deadest giveaway that someone has never used Windows that I've ever heard

[@pauldunecat]

Quartz land is happy land. 😂

[@puknut]

You can customize windows to your liking. That is a thing, you can pretty much do what you want if you really want to.

[@khoidauminh]

I'm a frequent Phoronix reader and I can confirm that every time an article mentions one or more of the following: GNOME, Wayland, Rust, Ubuntu/Canonical/Snap, etc, the comment section becomes a circus

[@burein_ita]

I use TWM with xterm running in a Xwayland session, inside a Wayland Hyprland session.

[@GaryGreene1977]

Birdy's arguments about Windows absolutely make zero sense.... Unlike X11, Windows doesn't, and never has, done a network styled client-server architecture for graphical systems rendering. This is why porting Gtk+ and Qt apps from Linux to Windows relies HEAVILY on the toolkit implementing the calls to drawing; doing anything else is madness since the primitives for GDI/GDI+ versus X11 are totally different.

[@breadmoth6443]

stilll running x11 , it is not very reassuring that wayland is 15 years old and we have such petty issues slowing down progress.

[@myownfriend23]

These aren't petty issues. Designing a protocol isn't like designing software. You can't have something work one way in this version, then change it in the next and applications can ship with the old version to maintain compatibility. Things need to be thought through a lot because these protocols are expected to be used for decades across a bunch of compositors, applications, and toolkits... and versions of those compositors, applications, and toolkits. Also it's not right to say that Wayland is 15 years old. It's just been in development for that long but only in the sense that the first commit to it's repository was in 2008. It wasn't anything that was meant to be used yet at that point. It had no implementation, no infrastructure to support it: nothing. It wasn't until February 2012 that it and Weston had their first release with version 0.85. It wasn't until July with version 0.95 that the API began it's stabilization process. In October, it reached version 1.0 with a stable client API and version 1.1 in April of 2013 the server API was stable. It wasn't until 2016 that compositor's were really considered to have support for them because none of them were ever made to support any other back-ends besides X11 so they needed significant changes in order to facilitate that. That's when clients started to have more compositors to actually test the Wayland support on. Considering Wayland requires a certain amount of acks and implementations before a protocol is merged, it takes awhile. X11 also didn't have all of it's functionality by it's first release, just most users weren't there before many of them were implemented. The core X11 protocol was designed in about a year (X through X11 all came out in a four year period) but it also wasn't meant to be used for as long as it was and I don't think many of it's designers expected it to be used by as many users as it was so backwards compatibility and longevity wasn't really considered. Wayland is intended to be used for awhile and not replaced every few years so things that may seem minor require a lot of consideration.

[@breadmoth6443]

@@myownfriend23 i can't be bothered to read all that blocko-text , but on the other hand Brodie already brought up a similar issue with the icons in another video, hence why i say petty... also if t his is what to show for, in 15 years of development... pretty sad as i expected more progress.

[@myownfriend23]

@@breadmoth6443 If you can't be bothered to read a response if it's more than a paragraph then this isn't a topic you should have an opinion on.

[@breadmoth6443]

@@myownfriend23 lol you are not arbiter of what i can and cannot have an opinion on , and you didn't post a paragraph but a wall of text , unless you can't get to your point in just one to two paragraphs.. i can't be bothered.

[@softwarelivre2389]

​@@breadmoth6443if you can't bother to read 30 lines of text, then don't bother sharing an uneducated opinion.

[@te41157]

From my understanding in wayland almost everything MUST be implemented in the window manager (and seperate tools such as xinput don't work and will never work again because of design choices). Is this true? From my point of view this whould look like a horrible design choice that leads to fragmentation and prevents problems from getting resolved efficiently in the "unix way". I'm also not so shure if it's really a good thing that wayland doesn't act as a display server anymore like xorg does.

[@haisulii]

Wayland so far is basically just a more modular approach where there's not a single server handling everything, this is good because trying to support the common denominator leads to tech dept and not every use-case needs a fully featured server as the required functionality can be very different between a kiosk system, phone, desktop, embedded, etc., as well as between DE and WM. The con with this approach is that it takes time for services and programs outside of Wayland to develop and complement the functionality of Wayland compositors, like input handling, screen recording, etc.

[@te41157]

@@haisulii I think I see the positive points you're highlighting. But I still see this aspect of fragmentation and I would basically say it's probably two steps forward and one step backward (at least I hope it is). The aspect of fragmentation is underestimated and not enough highlightet imo. I don't understand why every window manager has to do it's own input handling and provide tools and settings for it (even though with existing libraries but still...). I think it should be possible to develop tools outside of the compositor to provide these kinds of functionalities, especially input handling.

[@fotnite_]

I'm pretty sure this person is making the comparison to Windows because dwm consolidates the compositor and the window manager the same way Wayland does. Of course, this has absolutely nothing to do with the issue they made, which leads me to believe this person's only understanding of the architecture of X11 and Wayland was a brief glance at their Wikipedia pages.

[@SuperSmashDolls]

10:38 I have the feeling that this guy isn't going to be happy until there's a well supported protocol for allowing clients to arbitrarily modify the compositor window tree, read all keyboard and mouse events, and send arbitrary events to other clients. I say we should call that the Malware Extension (wayland-malware) as it allows complete control over your session to an attacker.

[@WiihawkPL]

birdie will stop at nothing until he can once again launch xroach in his colleague's wayland session

[@night_h4nter]

lmao, another "you shall be secure or else" take

[@PS3PCDJ]

This video completely pushed me from ever even trying wayland

[@GrzesiekJedenastka]

Why...?

[@e8root]

I don't think we can take Wayland as serious platform without TWM

[@Daktyl198]

Lmao, Avis is very well known on the Phoronix forums. He references this issue all the time like he's not the one that wrote it.

[@QuantenMagier]

And that is part of the reason I am still on Xorg-X11 and skipped this Wayland trash. Especially I miss xwd, xdotool and the X11 protocol to use programs remotely over SSH.

[@zxuiji]

Sounds to me like he's just not happy that wayland is aiming (or at least seems to be) to be just one module targeting just one interface and thus is making programming for it hard. I'm personally glad, it means that dedicate teams can focus on different modules for the other features that are desired. Once a community favourite is collectively decided over time (like I assume Xorg was) then the simplicity of programming for it will be a thing like it sorta is for xorg.

[@framegrace1]

Wayland is just aiming to be DE agnostic. They just manage the protocol and make sure on the app side everything works. They don't want to (don't have to) spend time creating any sort of reference server. Anyone is free to implement the Composite side of the protocols as they want. As an app developer you don't really care what DE you use, as soon as they implement the protocols you uise.

[@disieh]

The problem with Linux has always been that it's basically the refuge for people who really, _really_ want to do stuff their way. This culture makes it extremely hard for some types of projects to exist in. Even if you make a good alternative, you'll still be hated for it. Systemd is a good example of somebody just making an infinitely better version of sysvinit, it being so successful that distros replaced sysvinit with systemd. The amount of butt-hurt I keep reading about systemd to this day is mind-boggling.