Double click by default is a good decision. My first experience with KDE on Arch was me getting completely bamboozled as the applications opened just from selecting the icon on the desktop and I wondered what just happened. If you do single click to open, you have to remove the selection logic.
The floating panel is another one of those politics and fashion based decisions where GUI programmers keep forgetting Fitt’s law [1], where buttons that lay flush with the border of the screen are way easier to click because you cannot overshoot them, so they have a theoretically infinite size [2]. Maybe we should call this recurring phenomenon “Fitt Amnesia” because GUI programmers keep forgetting this decades old really quite basic and fundamental lesson, though it’s usually the GNOME and not the KDE devs forgetting it. It surprises me every time when seasoned GUI programmers introduce such severe usability regressions willy-nilly based on aesthetics. If you look at the task, you see a complete tragedy [3]:
Nicolas Fella: Is there a UX benefit to this? Or is it just “it’s fancy”?
Nate Graham: Just fancy, no UX benefit.
Al De: Personally I don’t find it quiet as “fancy”, as the panel in its current floating form (5.27.2) takes up more vertical space (roughly 16%), when set to floating.
Fitt’s law and ease of use are not even mentioned. Eh, as long as I’m not forced to use it and it remains customizable, I’m fine with KDE either way.
I just looked it up, they do make the clickable area larger than the button to extend it to the edge of the screen [1]. So I guess it’s just visually misleading but they kept UX in mind - at least as long as the panel spans the full width of the screen, such that the “start menu” button is in the corner.
That was 1997. 26 years ago. I am guessing that a lot of contemporary developers weren’t born yet, or were more concerned with learning to use a potty than a GUI.
So, no, probably not.
And IE4 on Windows brought in Active Desktop, another abomination. That’s the Windows model that KDE was based upon, with file explorer windows’ contents being rendered as HTML before display. A horrid design but it justified MS building IE’s rendering engine into the OS, which was its defence in the US government’s lawsuit against MS
It worked. MS got away with it… and KDE copied it.
Which is why KHTML got so good, which is why Apple used it for Safari, which is what led to Webkit and Chrome and the downfall of Mozilla and the increasingly closed proprietary 21st century Web. :-(
it’s probably not the ideal default setting for people who are migrating from other platforms, which is most of them
That’s probably not true. “most of them” are your existing users. I can’t remember if the setting is already there to switch between single/double click for opening stuff, but this default only makes sense if the behaviour they expect won’t be changed when upgrading.
Reading through the article makes me want to go back to try Plasma again. :) I am glad that the devs are still putting so much effort into the project.
To get there, we went over our “Wayland showstoppers” wiki page with a fine-toothed comb to refine what we really consider a showstopper. We decided that a lot of them are really more like annoyances rather than showstoppers, because X11 has plenty of annoyances of a similar severity too!
All platforms bring their own quirks – lots of users are baffled by X11’s two clipboards, for instance – so quirk-for-quirk compatibility is understandably a pointless goal. But these aren’t showstoppers only insofar as, sure, you can always move windows back to where they were and manually set wallpapers again, install fonts by hand, or logout and login when you switch away from that horrifying Breeze switcher (speaking of “better defaults”…).
Switching the back-end was bound to happen with some bugs still around, I mean, it’s not like the X variant is bug-free. But I would’ve hoped the bugs were less ominous. One makes session management kind of useless, one makes a whole app useless, another one makes the “it’s a cheap Windows clone” talk a lot more believable. This isn’t the kind of quality that the sixth release (okay, technically third for Plasma Desktop?) should aspire to.
Other than that, honestly, reading all this makes me happy: except for Wayland and double-click by default, I think the new defaults are terrible, but the old behaviour is still around and I can just roll it back in System Settings. If that makes the “well ackshually” crowd shut up, it’s a net win :-D.
Just to add some “showstopper” anecdote: On Kubuntu 22.04 if I use the nvidia driver blob (doesn’t matter which one), I can’t login with two displays. I have to disable the high-dpi one first or the whole system freezes (except for the kernel panic hotkey). Also if you disconnect the display and then later on re-connect it, it may not be detected anymore. It’s shown, but simply not used until reboot.
Is this a “non-showstopper” ? Probably yes, because nvidia drivers aren’t exactly their fault. Is it really annoying ? Yes. (Wayland won’t even login.)
The amount of friction for desktop linux is really annoying once you leave the beaten path (iGPU, one display, full-hd). Sure some stuff may be fixed on Arch, but I’m using LTS for stability regarding all the other things that may break with a feature update. But I do welcome the amount of time KDE developers invest into their software, and maybe it is better for move forward and break things, than to never actually progress and possibly burn out people on fixing annoying bugs. As opposed to people asking “why there is KDE 6 when KDE 5 isn’t stable”.
The amount of friction for desktop linux is really annoying once you leave the beaten path…
Some paths seem to be more broadly beaten than others. I have an AMD discrete GPU, three displays with weird vertical layouts and one of them hidpi, and it works perfectly with minor fiddling… using sway. Kinda unfortunate that KDE is having so much trouble replicating that. :-C nvidia drivers probably introduce a lot of bullshit into the system though, alas. Once in a blue moon my AMD drivers crash, but setup is literally plug-and-play.
In my experience sway is delightful if you’re willing/able to customize your workflow to it, but it definitely involves customizing your workflow. I like a “minimal but not ascetic” workflow, so for me it works out out well with a modest amount of tinkering, and it does even better on low-end hardware than X11 WM’s like i3 or awesome, so it kinda turned me into a Wayland True Believer. And similarly, AMD drivers work great on Linux ever since the RX 480 or so. But I have a friend who does lots of 3D modelling and so their GPU choices are based on 1) what gets the best perf in Blender’s CUDA-oriented benchmarks, and b) everything else, and they just gotta put up with nvidia drivers one way or another.
There is no one-size-fits-all solution, but for me the Linux+sway+AMD ecosystem is very satisfying.
Really hope that plasma tackles its not-really-human-editable config files next. My single gripe with Plasma is that I can’t put its configs into my dotfiles repo, everything else works mor or less exactly how I need it.
Design wise, and not really related to the article, I think that the clock widget on the right of the task bar (or whatever plasma calls it) is seriously ugly and out of place - the text is way too big. I prefer the way windows and macos make the clock look - it looks less out of place IMO.
“Rather, it’s the fact that Microsoft has blatantly copied us in Windows 11, and as a result, people are starting to see Plasma as a cheap clone of Windows again”
Is there even 1 designer at MS which is using Linux ? It’s pretty clear that with Win11 they copied MacOS.
But anyway, as a KDE user what makes it special is the huge amount of customizability. Nice defaults are good for first impressions, but letting the user customize the crap of out everything is by far the best thing about KDE.
Double click by default is a good decision. My first experience with KDE on Arch was me getting completely bamboozled as the applications opened just from selecting the icon on the desktop and I wondered what just happened. If you do single click to open, you have to remove the selection logic.
The floating panel is another one of those politics and fashion based decisions where GUI programmers keep forgetting Fitt’s law [1], where buttons that lay flush with the border of the screen are way easier to click because you cannot overshoot them, so they have a theoretically infinite size [2]. Maybe we should call this recurring phenomenon “Fitt Amnesia” because GUI programmers keep forgetting this decades old really quite basic and fundamental lesson, though it’s usually the GNOME and not the KDE devs forgetting it. It surprises me every time when seasoned GUI programmers introduce such severe usability regressions willy-nilly based on aesthetics. If you look at the task, you see a complete tragedy [3]:
Fitt’s law and ease of use are not even mentioned. Eh, as long as I’m not forced to use it and it remains customizable, I’m fine with KDE either way.
[1] https://en.wikipedia.org/wiki/Fitts%27s_law
[2] https://en.wikipedia.org/wiki/File:Magic_Corners_English.jpg
[3] https://invent.kde.org/plasma/plasma-desktop/-/issues/73
GNOME has a floating panel and you can still select the items on it when you have your cursor on the gap below. Is that not how it works on KDE?
I just looked it up, they do make the clickable area larger than the button to extend it to the edge of the screen [1]. So I guess it’s just visually misleading but they kept UX in mind - at least as long as the panel spans the full width of the screen, such that the “start menu” button is in the corner.
[1] https://youtu.be/8yeruvis8I0?t=264
Doesn’t anybody remember the debacle that was IE4?
I do. But…
That was 1997. 26 years ago. I am guessing that a lot of contemporary developers weren’t born yet, or were more concerned with learning to use a potty than a GUI.
So, no, probably not.
And IE4 on Windows brought in Active Desktop, another abomination. That’s the Windows model that KDE was based upon, with file explorer windows’ contents being rendered as HTML before display. A horrid design but it justified MS building IE’s rendering engine into the OS, which was its defence in the US government’s lawsuit against MS
It worked. MS got away with it… and KDE copied it.
Which is why KHTML got so good, which is why Apple used it for Safari, which is what led to Webkit and Chrome and the downfall of Mozilla and the increasingly closed proprietary 21st century Web. :-(
But no, they don’t remember.
That’s probably not true. “most of them” are your existing users. I can’t remember if the setting is already there to switch between single/double click for opening stuff, but this default only makes sense if the behaviour they expect won’t be changed when upgrading.
Reading through the article makes me want to go back to try Plasma again. :) I am glad that the devs are still putting so much effort into the project.
Setting is there, yes. Workspace Behavior>General Behavior>Clicking files or folders: Opens them | Selects them (open by double-clicking).
That was kind of a bleak read…
Some of the non-show stoppers include:
All platforms bring their own quirks – lots of users are baffled by X11’s two clipboards, for instance – so quirk-for-quirk compatibility is understandably a pointless goal. But these aren’t showstoppers only insofar as, sure, you can always move windows back to where they were and manually set wallpapers again, install fonts by hand, or logout and login when you switch away from that horrifying Breeze switcher (speaking of “better defaults”…).
Switching the back-end was bound to happen with some bugs still around, I mean, it’s not like the X variant is bug-free. But I would’ve hoped the bugs were less ominous. One makes session management kind of useless, one makes a whole app useless, another one makes the “it’s a cheap Windows clone” talk a lot more believable. This isn’t the kind of quality that the sixth release (okay, technically third for Plasma Desktop?) should aspire to.
Other than that, honestly, reading all this makes me happy: except for Wayland and double-click by default, I think the new defaults are terrible, but the old behaviour is still around and I can just roll it back in System Settings. If that makes the “well ackshually” crowd shut up, it’s a net win :-D.
Just to add some “showstopper” anecdote: On Kubuntu 22.04 if I use the nvidia driver blob (doesn’t matter which one), I can’t login with two displays. I have to disable the high-dpi one first or the whole system freezes (except for the kernel panic hotkey). Also if you disconnect the display and then later on re-connect it, it may not be detected anymore. It’s shown, but simply not used until reboot.
Is this a “non-showstopper” ? Probably yes, because nvidia drivers aren’t exactly their fault. Is it really annoying ? Yes. (Wayland won’t even login.)
The amount of friction for desktop linux is really annoying once you leave the beaten path (iGPU, one display, full-hd). Sure some stuff may be fixed on Arch, but I’m using LTS for stability regarding all the other things that may break with a feature update. But I do welcome the amount of time KDE developers invest into their software, and maybe it is better for move forward and break things, than to never actually progress and possibly burn out people on fixing annoying bugs. As opposed to people asking “why there is KDE 6 when KDE 5 isn’t stable”.
Some paths seem to be more broadly beaten than others. I have an AMD discrete GPU, three displays with weird vertical layouts and one of them hidpi, and it works perfectly with minor fiddling… using
sway. Kinda unfortunate that KDE is having so much trouble replicating that. :-C nvidia drivers probably introduce a lot of bullshit into the system though, alas. Once in a blue moon my AMD drivers crash, but setup is literally plug-and-play.Well that sounds promising. I maybe should try out sway and if I get around to buying a new GPU eventually, I’ll switch the team again ;)
In my experience
swayis delightful if you’re willing/able to customize your workflow to it, but it definitely involves customizing your workflow. I like a “minimal but not ascetic” workflow, so for me it works out out well with a modest amount of tinkering, and it does even better on low-end hardware than X11 WM’s likei3orawesome, so it kinda turned me into a Wayland True Believer. And similarly, AMD drivers work great on Linux ever since the RX 480 or so. But I have a friend who does lots of 3D modelling and so their GPU choices are based on 1) what gets the best perf in Blender’s CUDA-oriented benchmarks, and b) everything else, and they just gotta put up with nvidia drivers one way or another.There is no one-size-fits-all solution, but for me the Linux+
sway+AMD ecosystem is very satisfying.Really hope that plasma tackles its not-really-human-editable config files next. My single gripe with Plasma is that I can’t put its configs into my dotfiles repo, everything else works mor or less exactly how I need it.
Design wise, and not really related to the article, I think that the clock widget on the right of the task bar (or whatever plasma calls it) is seriously ugly and out of place - the text is way too big. I prefer the way windows and macos make the clock look - it looks less out of place IMO.
Is there even 1 designer at MS which is using Linux ? It’s pretty clear that with Win11 they copied MacOS. But anyway, as a KDE user what makes it special is the huge amount of customizability. Nice defaults are good for first impressions, but letting the user customize the crap of out everything is by far the best thing about KDE.