KDE is now my favorite desktop

Major changes aren't even _desirable_ in UI. People kind of emotionally enjoy novelty, however when it actually comes to using a computer consistency is superior to absolute excellence. Figuring out where settings and buttons are just because you ran software updates is a total waste of time on both ends; it wastes the user's time, and was a waste of time to develop. Maybe I'll switch from gnome to KDE this weekend, this looks promising.

Glad to know I'm not alone. As I grow older moving stuff around just to make it prettier doesn't do much except make me angry that 'they' changed things for no reason again.

Any organization that doesn't have backpressure against UX breaking changes is vulnerable to this.

The root cause is that UX folks almost never use a product as often as their users.

So what's an "oh, left instead of right" minor change for them is anathema to someone with muscle memory.

Ergo, IMHO, all breaking UX changes should be required to clear a high bar, with the default being status quo + tweaks.

I think it's perplexing that UX has generally gotten worse subsequent to multiple developments which you might expect would make UX better:

  - We now have a plethora of UX logging and can see real time where users struggle.

  - There are dedicated UX teams whose sole focus is to improve UX.

  - More people are using technology than ever, and so we have a more representative sample of data to work with.

But despite this, UIs have consistently gotten worse over the past 10-20 years. I think there are a few possible culrpits.

  - Mimicking mobile UIs, as eloquently called out here: https://news.ycombinator.com/item?id=45290812

  - I suspect there is something of a race to the bottom WRT To UX teams; they're always designing around any pain point, which has a few knock-on effects:

    - There will _always_ be pain points, and so there will _always_ need to be UI changes.

    - Designing a product so that the bottom of the bell curve can use it well probably does make an objectively worse product.

    - There's nothing wrong with needing to learn a UI, and this "learning" could be mistaken as pain point.

    - UX teams can't exist if there aren't things to constantly change, which increases the UI churn.

In concert, you have a UX which is constantly changing, and never really getting better, and often getting worse.

Is this always the case?

I prefer what Windows 11 has done with settings being a simple two panel window with categories on left and scrollable settings on the right, with a search/filter bar at top. As you drill deeper you have a breadcrumb at top allowing you to see the levels you are in and click to go back up. This also allows space for descriptions of what each setting does. It could even be improved by allowing users to pin commonly used settings.

This seems overall more simple and cohesive compared to the old Windows control panel with icons and nested settings being popups within popups within popups. It also allows easier scaling and viewing depending on DPI, screen size, resolution, etc.

This is an opinion stated as fact. Not every user is the same. That's why there are loads of apps with UIs that have different user modes, for power users, etc. KDE is most suitable for power users of Linux desktops, probably who use it as their daily driver. If you aren't in that category, you may not like it or may find it to be not worth the time investment in.

Not every user is the same, but it's absolutely valid to discuss whether broadly for most users constantly-churning UI is a net positive or net negative. I think your case, ie customizable UIs, is something of an edge case, and I do agree with you that expert versions of simple UIs can be a really positive move.

> There are dedicated UX teams whose sole focus is to improve UX.

Imho, this is a big source of the problem.

Granted: there are some amazing UX designers and teams out there.

But my experience with UX teams has been that in most middle-market companies they're usually less that sort and more the {huge designer ego} + {management consulting political skillset} one.

And it's a tough problem to solve! Because ultimately you want someone who can argue very hard for their approach to improving UX (usually against opposition from others). But when someone's ego exceeds their skill, that leads to disaster.

And without a strong Jobs-esque "this sucks" arbiter over them, their changes make it to prod.

Windows 11 settings are worse than they were prior to Windows 10. Before I could have multiple windows open for settings to monitor progress (like of windows updates) or check settings against each other. Now it's a monolithic interface that forces me to back out of something I'm looking at to look at something else, like a website that doesn't let me open multiple tabs to browse it. Terrible UX IMO.

A Linux user is, and has always been, an edge case

The builder vs maintainer mindset probably plays a role too.

Mature products need the maintainer mindset a lot more than the builder mindset.

It's hard enough to find devs who are good at maintainer-mode. I think it's even harder with other roles.

As a developer who has worked on similar things, that interface can prevent a lot of trouble of the kind "What if the user edits a setting that updates something in another part of settings, which might also be open at the same time?" - or even the same settings screen opened multiple times. It is rare that such cases come up for users, but they can be very annoying and time-consuming to deal with. Perhaps that was the motivation.

This is an opinion stated as fact. KDE is mostly for dads that like a mouse oriented Windows/mac like OS but with buttons to customize.

Sway, exwm etc are for power users.

If you aren't in that category, you may not like it or may find it to be not worth the time investment in.

10/10 gatekeeping buddy.

I get it, but instead of putting effort into redesigning the whole interface to avoid this possibility, they could have put that effort into rendering the same UI in immediate mode and problem solved.

I actually think the real motive is that they wanted to move to a more unified mobile and tablet friendly UI code base, which centers more around full screen windows.

> - There are dedicated UX teams whose sole focus is to improve UX.

They don't work like the UX teams of yesteryear.

In the early 2000s, companies did user studies. Put a potential user in front of the product, let them use it while the UX team observed. Ask questions to the user afterwards. Make changes, repeat.

But that kind of research is expensive, so it's thrown out to instead just collect tons of metrics that likely don't tell a whole story. They think "Users must love feature X because they keep clicking on it!" when the reality is that they're trying to find something else, but the label for X looks related to what they want.

I agree with all your points regarding the race to the bottom. I think that's why UIs hide so much information. Older interface designs are considered "confusing" or "cluttered" because there's so much there at a first glance, even if all the buttons elements are grouped by functionality.

I dislike how people seem to recommend KDE for power users when GNOME is actually.. great?

I'm an old power user/dev and I used to absolutely love KDE 3 for its take on 90s OS UI, I went into v4 thinking it was a major downgrade (I used KDE 3 as far as the KDEMOD maintainers could push it) and it never got as good as the old v3 days. Somewhere by the end of the KDE 4 life, GNOME 3 formed into something kind of usable and I started noticing some advantages to it, even tried it for a while. Fast forward to now (including a few years where I rolled my own LXDE/XFCE hybrid setup, I was desperate lol) and now I pretty much only use GNOME. I consider it a fine DE for power users... or whatever use you have really. It's great on a notebook, it's great on a desktop and it's great even as an HTPC interface. You do have to wrestle with it for some advanced functionality (dealing with extension isn't always fun, digging into dconf isn't fun...) but the OOB defaults and basic functionality are actually the best there is, maybe even among all desktop OSes.

I mean, if Linus Torvalds out of all people uses it then it must be at least decent for more advanced users, right?

Now whenever I try KDE it feels like an uncomfortable car where every single adjustable thing needs to be tweaked for it to feel minimally usable, except many adjustments are finicky and leave you with a half assed solution. It won't resonate with me anymore...

> We now have a plethora of UX logging and can see real time where users struggle.

No you can't. That telemetry gives you view into how users are experiencing the software is a myth because it doesn't include the actions users don't take and it doesn't include the reasons for actions taken.

Making the developer experience easier at the expense of the user experience is the heart of the problem with modern software.

There isn't only one kind of power user, and not every kind of power user will like KDE. I use GNOME almost daily, with KDE being my primary daily driver. I just really don't see how GNOME could ever said to be great for power users. There are so many ways you can tailer KDE to specific things you may find yourself doing heavily, that would be impossible with GNOME's UI. Being a power user and using GNOME, to me just means you're using the terminal for everything, which is totally valid.

Linus also may not even really be a power user. He says himself that he rarely writes code anymore, and primarily just sends emails and reviews code.

I didn't say KDE is the only option for power users, or that there aren't multiple types of power users with different use cases and individual preferences. I'm not sure what I'm gatekeeping.

Is it really gatekeeping to say that KDE is for power users? Setting it up in a way that really meshes with your use case and preferences is a process that you'll spend many hours or days of time on. That's not something that makes sense for grandma's computing workload.

> This is an opinion stated as fact. KDE is mostly for dads that like a mouse oriented Windows/mac like OS but with buttons to customize. Sway, exwm etc are for power users.

So you're saying that prefering a highly customizable GUI means you're be a power user, but instead you're a gasp dad? This isn't Reddit, buddy. Grow up.