TL;DR: the author needs to do a better job of citing sources and building an argument.
The author’s argument from self-appointed authority tone aside, I dug into the only two verifiable pieces of evidence cited. These are almost impenetrable to the outsider, and even with plenty of coding experience behind me, I’m having to go deep to make sense of any of it. After all, sometimes, bugs and design decisions are the result of a best effort in the situation at hand and not necessarily evidence of negligence, incompetence, or bad architecture. There’s also something to be said for organizing labor, focusing effort on what matters, and triaging the backlog.
The original author really needs to pony up a deeper digest of the project, with many more verifiable links to back up the various quality claims. If anyone is going to take this seriously, a proper postmortem is a better way to go. Cite the version reviewed, link to every flaw you can find, suggest ways to improve things, and keep it blameless. Instead, this reads like cherry picking two whole things on the public bug tracker and then making unsubstantiated claims that’s a part of a bigger pattern.
My personal take on what was cited:
I’m grossly unqualified to assess this codebase as a Wayland or GUI programmer, but work plenty in the Linux space as a cloud practitioner and shell coder.
The first article smells like inadequate QA for cases like placing Wayland programs in the background, which is not typically done for GUI apps under normal usage (IMO).
The second article is a two-line change that I suppose highlights how ill-suited C is for this kind of software. Developer chatter on the MR suggests that the internal API could use some safeguards and sanity checks.
162 open issues, 259 closed, oldest still open is five years old. Not great, but not terrible.
None of this is particularly egregious, considering the age of the project and the use it enjoys today.
You aired my frustrations really well. He spent a lot more time making claims and discussing his own background than demonstrating Wayland’s alleged issues and showing that they’re egregious. It’s an entertaining rant at best, but that doesn’t make his points valid nor does it make anything actionable.
I have been using Wayland on void for a while and have no particular issue with it. There is screen sharing on stuff like zoom that isn’t working at the moment (unless you use gnome) which is a bit annoying but not really serious enough to force a change to xorg. Also Wayland has more clean code then xorg and I do like the potential it has, specially when it comes to security.
Nothing against xorg, if you can use Wayland its better imo but otherwise xorg is fine as well.
I think adoption is king. The best solutions often fail when it comes to adoption, though. And starting a new solution when one is already adopted is not at all easy.
I understand that this author is working at a much lower level than the gamers and other casual users, so they will be much more likely to have to deal with the repercussions of poor design choices and edge-case bugs and missing functionality. But if they can make things work well enough when they are paid to do so, then adoption will continue. (On the other hand, they will also be the among the first to hit any showstopper issues if they do exist.)
I don’t think this kind of community is the best place for discussing nitpicky technical details because to most of us it is effectively whining about issues we will never have to deal with. (Nor is it a bad place, per se.) I think the comment would find a better home being digested by the technical experts who work on display solutions and other interoperating pieces of the larger environment. They are in a good position to weigh the criticism’s merits, and if any concerns are highly merited then they are the ones who would decide whether and how to design and and implement improvements.
The guy’s language is the exact same hand wavey magic he opened up criticizing. “Appeal to my claimed authority. Wayland bad. Missing parts. Analogous to poop.”
I practically never come across anyone who addresses a point directly and don’t think I ever will. Everything is tribalism and politics and sunken cost fallacies.
I don’t really have much of an opinion about Wayland but it’s still funny to me whenever somebody using Wayland shits on X11 and then tries to share their screen on Zoom or something. If Wayland ends up being great I’ll be happy, but for now X11 just kind of works, so I don’t understand why people are so eager to switch? This isn’t to say I don’t understand the desire to build something better and more secure than X11, I’m just not sure what the end user gets out of Wayland right now. I don’t have VRR monitors and stuff, though, so maybe I’m not running into problems I would be if I wanted fancier features. Plus, I use xmonad and some other stuff right now that won’t work on Wayland, so I don’t have much incentive to try it. Hopefully everything gets Wayland updates eventually.
With those however, they never ran on Linux. This situation is different because it did run. I’ve only used Zoom once, so no clue if it worked excellently or if it was “meh”, but it sounds like it did the job before.
Regardless, it doesn’t matter if Zoom hasn’t updated their Electron to account for the Wayland changes - all people will see is that it doesn’t (or did, but no longer) works on Linux and will blame Linux instead.
Which, that is fine if we want Linux to always be a hobbyist operating system. However if we want Linux to be more accessible to people then unfortunately the ball is in our court to try to not break something as simple (or rather, what most regular users would define as simple) as this.
Isn’t screen sharing working since some time? Works even on WebEx from Firefox, can pick any window to share. Granted a few years back it didn’t work, but now it does. Maybe it’s a zoom bug… 🤔
Probably, but my exposure to Wayland has just been people complaining about how much X11 sucks and then proceeding to have more problems than everybody else.
No,I just had to deal with this myself. Most you can do is share your entire desktop in Wayland, and it’s shaky. For the first time, I had to switch to Xorg and bingo, zoom works. Fonts are actually antialiased and kerned properly for certain applications that weren’t… Really surprising.
Again, none of that is a failure of Wayland, it’s a failure of Zoom to run on Wayland. One day, and this is in the next 5 years, Wayland-only apps will refuse to run on X.Org and the situation will be reversed.
You can share screen perfectly fine under Wayland. Many apps use it fine, and even in case of Discord if you use it with a browser it’s doable.
No Wayland dev can fix an issue that originates from lack of app support. There has been many Wayland issues through the years and trust me, I know, but how do you expect them to fix Zoom? Acquire the company and take it behind the shed?
how do you expect them to fix Zoom? Acquire the company and take it behind the shed?
I mean, you could - for example - implement the interface these apps expect to exist and use with your amazing new compositor™.
This is precisely why companies just say “fuck Linux users” - instead of supporting a single operating system where everything kinda “just works” across versions for decades you have to checks notes support 20 different compositors across 2 vastly different display servers and dozens of various desktop environments and such… All for an OS that’s used by maybe 3% of your users if you’re lucky.
The interface exists. It’s up to zoom to support it. Why are you under the impression there is a technical issue? THERE IS NONE.
It’s up to Zoom to support the aforementioned interface.
Wayland’s display handling in this manner is for security, the user will be shown a permission request dialogue to let the app access the screen only if you permit it, it’s also disallowed from accessing anything except what you’ve given it permission to. This is not even a new concept, just not doable under X.
It’s also possible to create the lawless model of X under Wayland through a protocol if you desire to make one, but it makes little sense to throw away this better model just for the sake of some shitty proprietary apps who don’t care for Linux anyways
You will never be a real display server. You have no hardware cursors, you have no xrandr, you have no setxkbmap. You are a toy project twisted by Red Hat and GNOME into a crude mockery of X11’s perfection.
All the “validation” you get is two-faced and half-hearted. Behind your back people mock you. Your developers are disgusted and ashamed of you, your “users” laugh at your lack of features behind closed doors.
Linux users are utterly repulsed by you. Thousands of years of evolution have allowed them to sniff out defective software with incredible efficiency. Even Wayland sessions that “work” look uncanny and unnatural to a seasoned sysadmin. Your bizarre render loop is a dead giveaway. And even if you manage to get a drunk Arch user home with you, he’ll turn tail and bolt the second he gets a whiff of your high latency due to forced VSync.
You will never be happy. You wrench out a fake smile every single morning and tell yourself it’s going to be ok, but deep inside you feel the technical debt creeping up like a weed, ready to crush you under the unbearable weight.
Eventually it’ll be too much to bear - you’ll log into the GitLab instance, select the project, press Delete, and plunge it into the cold abyss. Your users will find the deletion notice, heartbroken but relieved that they no longer have to live with the unbearable shame and disappointment. They’ll remember you as the biggest failure of open source development, and every passerby for the rest of eternity will know a badly run project has failed there. Your code will decay and go to historical archives, and all that will remain of your legacy is a codebase that is unmistakably poorly written.
This is your fate. This is what you chose. There is no turning back.
I remember having this realisation about Mir, but only after we collectively ran it off the cliff wall. The main reason everyone piled on Mir was that it was thought that Canonical would be priming Linux desktop for fragmentation with two competing standards.
But in fact, Mir was providing a solution to the fragmentation Wayland was bringing. Now we have 3, 4, 5 Mir-s, all with slight incompatibilities. Want a feature? Better hope all of them decide to implement the extension after someone proposes it. We know how well that worked in the past.
This is also ironic because the detractors of Xorg constantly talked about the issues with Xorg extensions and how many of them there were. But I never really had to look up which extensions Xorg supported, while I have had to do that with Wayland compositors.
The main reason that I piled on Canonical was that they kept on spreading FUD about Wayland to try to promote / justify Mir rather than discussing in good faith.
Add comment