While Elementary OS commits to Wayland, the development team of the Budgie desktop is changing course and will work with the Xfce developers toward Budgie’s Wayland future.
There is general consensus now that the future of graphical desktops on Linux lies in Wayland rather than X11, but the path is still not a smooth and easy one. While in Latvia for the Ubuntu Summit, the Reg FOSS desk met with the developers behind Ubuntu Budgie, who told us that the Budgie project is charting a new course toward the brave new Wayland world.
It seems that using EFL – the Enlightenment Foundation Libraries – wasn’t the right choice for Budgie, and so they’re now exploring working with Xfce on their Wayland efforts, instead. Considering Enlightenment’s desktop Linux presence is negligible, at best, joining forces with Xfce so that both Xfce and Budgie can make progress on Wayland faster seems like the more optimal choice for the wider desktop Linux community.
I still haven’t found a good argument why we need all these desktops like Budgie, Mate, Cosmic, Unity (still alive!), LXQt, LXDE (still alive!), Cinnamon, Pantheon, Deepin Desktop, UKUI and so on.
Wasn’t the “big three” from the late 90s enough?
KDE Plasma for the power users.
Xfce for those who want a simple, lightweight desktop.
Gnome for those who prefer the iPad workflow.
Each tries to bring something new, but is invariably based on one of the others. The “problem” comes from the fact that if something is popular the USP is incorporated back into the parent leaving you with a less capable port.
The freedom and choice Linux grants is also its curse. I’ve known a lot of users who discovered Linux one way or another, only to be turned off by feeling overwhelmed by too much choice, ultimately abandoning their interest. Regular users don’t want to research 10 different desktop environments and 20 different distros. They don’t want homework and they don’t enjoy wasting their time testing a bunch of different options without any real clue what they’re supposed to be doing. They just want to install and be done with it. I think the more savvy users & tinkerers often forget that. Honestly, I’m guilty of giving that same shitty generic advice so many others offer too, `just download a few iso’s and see what you like`. The YOTLD will never come because the fragmentation problem will never be resolved, and most Linux-curious users will undoubtedly lose interest and stick with what they know.
friedchicken,
That’s a fair point. There’s a lot of different desktops and researching them is a lot of work. Just a little guidance from experienced users could save them a ton of work. The problem is experienced users all suggest something different, haha. Personally I think most windows users will feel at home with KDE, which is what I recommend for new users.
I don’t think we forget that, but the linux community is so polarized that new users are bound to find tons of conflicting information and “top ten lists” that will make them feel like they’re flipping coins. I know the feeling because I’ve been there, I think we all have.
I don’t ever recall giving this advice personally, haha. The most practical (ie actionable) advice almost has to be biased towards some specific solution just to get the ball rolling. We’ll never agree on what those specifics should be, but getting the ball rolling needs to be prioritized over trying out all the distros up front.
Users tend to be that way with all platforms, linux or not. There’s always some friction in changing platforms because it’s a whole ecosystem. IMHO the application ecosystem is a bigger barrier for linux desktops these days, although with so much existing on the web I think the transition barrier is the lowest it’s ever been.
@Alfman
I tend to lean towards KDE being the least foreign for Windows users. The last time I helped someone try Linux for the first time was about 2 months ago. KDE kept falling on its ass just trying to fetch available wallpapers and other customization choices though, which didn’t exactly make for a stellar first impression. I had neither the time nor desire to try to figure out why, or switch to something different, and the person I was helping quickly lost interest. “I can just keep using Windows” wasn’t how I thought it’d play out.
It’s definitely important to get the ball rolling but the caveat to that is picking a starting point that’s most compatible with that particular user if they’re not interested in giving a few options a test drive. There’s pressure to get it right the first time or face the risk of driving them away before they’ve even dipped their toe in the water.
It’s hard to feel solid about recommending Linux to average users knowing the fragmentation and tribalism, among other things, is going to lessen their experience and probably leave them worse off. I think Linux in it’s current state is a good fit for some people, but not most.
friedchicken,
Well, I’ve never tried pushing linux onto someone who didn’t have an interest in changing in the first place. Frankly, unless there is a reason they want to change (big emphasis on “they”), windows users are best served by windows, macos users are best served by macos, etc. This actually applies to linux as well, but being linux users, we already know where we fit into things.
I understand that, it likely depends on their motivation for changing. To be perfectly honest, my first professional experience with macos, as a windows user, did not go well. But this is not to say macos is bad, but my productivity fell because I lost the ability to lean on years of windows proficiency. Even trivial things like keyboard shortcuts were out the window. (Incidentally, watching macos users, I’m often surprised at how much they rely on slower mouse inputs. I’m curious if there are statistics on this, haha).
It depends on what they do, but I’ve long said that it is easiest to stick to what they know unless they have a reason to change.
@Alfman
I don’t think the number of people with an actual reason to switch holds a candle to the number of those who don’t. But, a number of those who don’t can still have curiosity. That describes the type of user I run into the most. I’m a big fan of use what works best for you & your needs but if someone wants to experiment a little, cool, go for it!
friedchicken,
I agree. I would say a live disk is best for the curious before doing a full install. From there, if they remain interested, then great!
I don’t think one necessarily needs to be tech savvy to use linux, I know a few people using linux who aren’t tech savvy (grandparents & kids). It’s really a matter of use cases. If they want to use windows applications, that rules linux right out, but for web browsing and libreoffice/etc, they can usually pick it up. If they are already familiar with FOSS software on windows, that’s over half the challenge.
The “big three” is what has produced KDE 4 and shortly after Gnome 3. Commercial vendor would have been slapped by investors and quickly adjusted their course. Not Linux. Here it is the community that has to step in to rescue its most popular desktops (KDE 3, Gnome 2) abandoned by the authors whilst others scramble to find other desktops (it is only then “big two” became “big three”) or propose their alternatives. Still, Linux desktop has suffered massively, it has never returned to the growth it was seeing in Gnome 2 times.
This should be a lesson to proponents of killing X11 – it is not possible to alter a living platform, the authors can only decide whether to kill the existing one and start from scratch from scratch. X11 is more important than a DE, as it cones with its own ecosystem. Not so much with X11, if it was to magically disappear, Linux desktop would have lost a major (only) niche of apps ported over from Solaris and of course a long tail of smaller apps that nobody will have resources to port. Fortunately, this is Linux, X11 implementations are only a few clicks away, so commercial vendor will simply say “use support only distribution A” instead of porting the whole thing to Windows.
My third point is Wayland standardisation. Realistically, we will have a dual stack solution for as long as Linux desktop exists. Whether it is X11 or Wayland handling the desktop is not important as long as it is a single standard similar to fd.org. Unfortunately, currently it isn’t. So far, it has mostly been driven by Gnome in an attempt to break off from standards, so unless all desktops agree on countless fine details (so it is great to see some movement here), this is not a Linux desktop. Not more than Android or Chrome OS are.
Regarding enlightened, obligatory post: https://what.thedailywtf.com/topic/15001/enlightened
This is what happens when people drag their feet. Truth is that Wayland was anointed the successor of X11 on Linux. Despite Canonical muddying the waters with Mir for a bit, it has been known for well over a decade that Wayland was it.
**Nope! Not gonna do it! You will pry X11 from my cold dead hands!** Not a literal quote, but the prevailing sentiment. Now that dead X11 is starting to smell in their living hands, now we need to scramble. This could have been avoided if the smaller projects had banded together to write a common Wayland compositor. Oh well, at least there is movement at last.
Mir was designed to be used for mobile/tablet/desktop. The crowning of Wayland by the Linux community was acknowledgment they gave up on that unified dream.
Agreed, it is the future of display servers for Linux based OSes, just as systemd is the future of init and system supervision on them. This kind of progress was inevitable in the Linux world but will leave behind folks like me who prefer a more BSD like environment when using Linux. Which is fine, projects like Void, Alpine, Slackware, and a few others will continue to be BSD-like, so for the oddballs like me out there, a path to continue using Linux our own way exists.
Have you seen Chimera Linux?
https://chimera-linux.org/
Chimera uses the BSD userland, avoids systemd, and compiles with Clang / LLVM. You may like it.
Then again, Chimera uses Wayland with the GNOME desktop environment by default.
Of course, even BSD has Wayland support these days. When Red Hat stops supporting Xorg, we will see if anybody steps up to keep it viable. We forget ( or I do ) how much support the entire Open Source ecosystem gets from Linux.
I’ve heard of it but it didn’t seem to offer me anything I couldn’t get from the Musl C version of Void Linux. I’m definitely cheering them on though, it’s an ambitious project!
r_a_trip,
I’m not going to speak for anyone else, but for me personally the argument was never about sticking with X11 and avoiding replacement. However it is important that thing replacing X11 needs to work. This has been the obstacle.
That’s exactly it for me too; I’m glad the people behind systemd and Wayland respectively are trying to modernize and improve those parts of the Linux ecosystem as it is sorely needed, but both projects are developed in a “move fast, break things, damn the consequences to the users and fuck the rest” philosophy that is simply not compatible with those of us wanting stability and maturity from such critical services. No software is ever “finished”, but I for one won’t be able to adopt systemd and Wayland until they are at parity with their contemporaries in stability and robustness. Without a change of leadership for both projects, I don’t think that will happen any time soon.
For those who enjoy depending on moving targets and the rush of living on the bleeding edge, more power to them and I wish them well, seriously. As for me, I’m going to stay safe and secure in my little corner over here.
Before the obligatory posts about how Wayland doesn’t work with this or that obscure DE, the XWayland Rootful mode allows just that.
kurkosdr,
This is the only source of info I can find on it…
https://www.phoronix.com/news/XWayland-Rootful-Useful
(my emphasis)
I agree it sounds like an interesting project, but honestly I don’t think it can add back the features that wayland took away. It’s more like creating virtual X11 desktop and running that inside a window on wayland, which is neat and all, but it doesn’t replace wayland’s missing features such as network transparency and remote desktop.
https://ofourdan.blogspot.com/2023/10/xwayland-rootful-part1.html
Eventually I hope all the desktops can independently re-implement wayland’s missing features, until then though I’m stuck. These problems are literally decades in the making. What a shame wayland has been managed with an “our way or the highway” attitude without caring about user needs.
Just to clarify, even if these features work within the virtual X11 desktop, a virtual X11 desktop running inside of a wayland window might not fit the bill. It’s like telling someone to load up a VM and use the VM’s screen share functionality. It’s great that the VM has that feature, but it’s not necessarily a substitution for the lack of screen sharing on the host.
Network transparency as implemented in X11 is what makes it hard to add new features to X11 (since you have to commit to a very detailed protocol and everything needs to be streamable through a network), so goodbye and good riddance to network transparency. Eventually, X11 holdouts will migrate to Wayland mode when their favorite apps become Wayland-only (unless they are fine with xeyes, xterm and xclock as their app collection).
Wayland does it right by rendering the desktop and just streaming the pixels like a VM’s screen share functionality.
kurkosdr,
That makes wayland officially less capable than bsd/unix/windows/macos/etc.
Screen sharing/vnc is not so niche that it can be ignored for a modern desktop. Granted remote work and collaboration is less important than it was a few years ago for remote work, but the fact that so many users could not use wayland then highlights how it was not (and still isn’t) ready for prime time for some critically important use cases. I understand it may not be important to you, but wayland is still regressive for others and you’re asking them to reduce their expectations for what computers should be able to do. I say that’s a mistake and that the problems should be fixed first, otherwise it’s only harming linux in the eyes of users.
Unfortunately that’s not the case, wayland doesn’t let you do that but I agree it should! If it had that support then clients including xvnc would be able to work just fine under xwayland and I for one would have no issue using it. Nope instead wayland omits such features thereby forcing each compositor to independently implement it.
Wayland compositors do not need to implement network transparency. Waypipe works with all of them:
https://www.mankier.com/1/waypipe
For remote desktop, most compositors that I am familiar with already support either RDP or VNC.
tanishaj,
It does not look like this is in debian stable yet. It progressed from unstable to testing in june.
https://tracker.debian.org/pkg/waypipe
Hopefully it is genuinely on track to address network transparency, that would be good news. I’d like to hear from people using it before assuming it solved.
Unless things have changed fairly recently, the only desktop I could get working using remote desktop tools was gnome. Maybe in time more compositors will fix the regressions, but for the time being it feels so wrong to have to switch my preferred desktop environment on account of wayland’s breakages.
FWIW, this has been my experience as well. On KDE Plasma with Wayland, RDP is a no-op (crashes as soon as I connect), and NoMachine has a black screen that only updates the viewport in a small square around the mouse cursor. On GNOME these issues don’t happen, but I don’t care to use GNOME in general. In X11 I have zero issues with any remote desktop protocol, under any DE or WM.
This is the kind of thing I was talking about up-thread in my other reply to you; for people who can live with daily breakage of essential, basic services like RDP and VNC, more power to them. As for me, I need to get shit done so I need stability and maturity in my OS.
I will eventually trickle down to stable and we can leave the whole issue behind.
kurkosdr,
Assuming that it actually works out, then sure, most of us are pragmatic in our choices. I do wish we’d learn to improve the process though, this tendency by project leaders to ignore users is troubling. As any experienced software engineer knows, it’s helpful to design code features around well defined requirements up rather than hacking them in later. After all, X11’s complexities mostly stem from feature extensions being added after the fact. So I find it kind of ironic for wayland to be guilty of it too; we’re already seeing how it is fragmenting the user experience across desktops.