I think GNOME is what you get when devs think of themselves as "opinionated" and keep trying to chase the latest trends in UX.

The thing I like about KDE is that they adhere to the classic tried-and-true desktop experience while being ready and willing to accept that everyone has different preferences that they should accommodate. That is pragmatism, and that is treating your users with respect. (Whereas the GNOME philosophy is more, "our way or the highway.")

> everyone has different preferences that they should accommodate. That is pragmatism, and that is treating your users with respect.

I completely disagree. Not every product is made for every person. If you don't like it, use a different product. It's not like there aren't a plethora of desktop environments to choose from. There is obviously a sizeable portion of the desktop Linux market that finds GNOME to be a perfectly usable desktop environment.

Your comment is similar to asking why Tesla does not create cars with internal combustion engines.

The problem with GNOME is that it caters to a very narrow audience, yet it is the default in most Linux distributions.

That is not a GNOME problem. That is a distro problem. Users should petition their distro maintainers for change.

We do, but we get "spins". When a distro starts they accidentally pick a core desktop environment by simply picking one first. They call their distro Foobar Linux and everyone is happy. Later people realize that the choice of desktop environment wasn't good for everyones needs and they add a KDE spin called Koobar Linux or Foobara Linux or who knows what else worse unrelated shit they can call it.

The end result is that most users who are getting involved in Linux for the first time are making this choice long before they even know they are making this choice or being given an option simply by picking a random letter version of a distro like Xubuntu or Kubutnu or not having a letter at all and it being Ubuntu alone.

There are a few distros that have done it better and given the user a choice upon install to pick their DE. OpenSUSE has always done something like that from what I recall and I appreciate it a lot.

Where is the problem here?

The real issue should be: Fedora/Debian/Ubuntu cater themselves to a very big audience and yet chose to set GNOME as the default

My understanding[1] is that distros prefer GNOME as the default due to it's release cycle being significantly easier to package then KDE's.

[1] Going off comments from one dev behind the Aeon distro on why it uses GNOME and not KDE, despite being an OpenSUSE spin-off.

> There is obviously a sizeable portion of the desktop Linux market that finds GNOME to be a perfectly usable desktop environment

yeah the ones who don't know how to change it.

I've been using Linux since Ubuntu 12.10-ish days, and I use vanilla GNOME on Fedora Workstation while knowing how to add extensions. Why are you discounting experiences like mine?

I hope your extensions still work at the next update!

As I said, I don't use extensions, so why are you discounting experiences like mine?

alright, sure, you can have it! I'll allow it this time.

That's almost the same setup as mine and I think there are a lot of users like us out there that want to use a regular desktop experience but sometimes want a bit of spice. I think the Burn My Windows extension gives me enough fun when I need it without diverting too much of my time to "ricing".

> I completely disagree. Not every product is made for every person. If you don't like it, use a different product.

These are large systems with their own ecosystems within them. When a project does this kind of thing it destroys a lot of the value that other developers are bringing for no gain.

Honestly, it appears as laziness sometimes like taking the easy route of keeping your verbose C codebase from growing more hair. Other developers are quite happy and accommodating to add experimental features gated behind a flag or option and still maintain a cohesive and polished experience.

Every option you add increases the size of your test matrix. It becomes untenable.

You do not have to include all experimental features in your test matrix.

If you are not testing a full matrix, you are creating less reliable software and waiting for bit rot.

This assumes you keep all of your features experimental.

No, that would be like Tesla making a car on which you can’t open the windows, because Elon thinks you shouldn’t need to.

So then don't buy the car. That's an easy solution to your problem.

I tried out Ubuntu with Gnome a few years ago after a lot of experience with KDE on Gentoo. I guess I don't get it.

No task bar. Had to Google for the command to install one. It's worthless. Icons change order. I have 5 instances of the same program open, and I can't tell which is which because they change order in the task bar. Can't even enumerate them since their order changes while I'm switching through them! This is the most basic function of a window manager and it doesn't work.

Oh, am I supposed to use that full screen task switcher thing? No, all my windows contain text, and I can't tell their thumbnails apart.

Yet, so many people prefer it. What the hell am I missing?

I feel the opposite - so many people prefer KDE, what the hell am I missing?

KDE used to look ugly comparing to Gnome. Now it's fine but still worse.

KDE offers plenty of shiny things that I don't like and need - graphical effects, application menu, widgets.

Most of the time I have a terminal and browser full screen and all I need is cmd key that triggers search and allows me to enter the name of the application I want to run and hit enter. Gnome is much cleaner, simpler and logical to use.

That's not the opposite. I'm talking about functional problems that make basic tasks difficult. You seem to be talking about your personal aesthetics.

> cmd key that triggers search and allows me to enter the name of the application I want to run and hit enter.

Yes, I do this in KDE all the time.

> I feel the opposite - so many people prefer KDE, what the hell am I missing?

Many people use their computer for a use case that's not "single application taking up the entire screen.", and KDE works better for them than GNOME. I'm not sure what's hard to understand about that. There's no objectively correct way to use a computer, it's a preference thing.

The way I like my KDE is probably very unlike how other people like their KDE.

KDE is extremely customizable. A recent update added a floating taskbar (on a new install - upgraded installs keep the old behavior), like everyone else has. However, you can change it easily, and I even discovered some more customizability I didn't know existed. I can have an icon-only taskbar like Windows, or I can have the full title.

Everything in KDE is like that. You can follow the trends, or stick to the tried and true, or come up with your own style.

Ubuntu runs a heavily modified GNOME, and your issue is entirely in their taskbar extension. Vanilla GNOME on Fedora works fine for me.

GP said:

>No task bar. Had to Google for the command to install one.

>Oh, am I supposed to use that full screen task switcher thing? No, all my windows contain text, and I can't tell their thumbnails apart.

It looks like they installed Ubuntu with vanilla GNOME (considering the "few years ago" part perhaps even the now discontinued GNOME flavor) rather using Ubuntu's interface. And their issue seems to lie in GNOME itself (no taskbar, activities overview-oriented interface) rather something Ubuntu is doing.

I’ve always wanted to like GNOME—it’s polished and has an undeniable appeal—but I just can’t get behind its UI metaphor.

For my workflow, simplicity is key. That’s why I prefer desktops like XFCE, Budgie, or Cinnamon. Even MATE with the "Redmond" skin feels more intuitive to me.

Why? It comes down to how I manage my work. I often set up one browser window per project or topic, labeled accordingly (e.g., Project1, Project2). I even have a window called MAIN for my daily essentials like Mail, Calendar, WhatsApp, and Signal.

With a macOS-style UI, I only get one taskbar icon for all these windows, making it difficult to find what I need quickly. On the other hand, with the classic Win95-style interface, I can easily and intuitively spot my open projects at a glance. No guessing, just muscle memory.

Just my thoughts—your mileage may vary. Curious if anyone else feels the same way?

I feel exactly the same way. Back when KDE was a bit messier and perhaps buggier, I tried to muddle along with GNOME for a number of years. Lots of little daily annoyances especially around the file management windows, multiple screens, audio configuration, etc. I prefer focus-follows-mouse, which required some kind of add-on plus changing a hidden setting somewhere and it still never worked properly. Switched to KDE two years ago and ALL of those things just worked out of the box. And everything else, is easy to customize without jumping through any hoops.

Yes, KDE6 surprised me in a good way. I still prefer the other desktops, but I know I can use KDE just fine.

I believe it was the KDE4 release that went down in history as being horribly buggy.

KDE 4 was pretty bad and we had to use it via Whonix for years.

> Just my thoughts—your mileage may vary. Curious if anyone else feels the same way?

I sure do. The default MacOS interface, to me, feels like it's for children. It's too much work to use it for serious work.

These days I've settled on AwesomeWM. I've been slowly building my ideal desktop since the pandemic. It's super lightweight and completely, completely customizable. I can add code in to do whatever I want pretty much anywhere and it is setup to make that super easy to do so. It's a true joy.

I have a super lightweight, efficient and attractive version of the way I have my Windows desktop setup, except there I have to use a bunch of add-ons that take up memory to achieve the same result. Things liek adding a titlebar button to be on top or to send to a different screen with just a click, or to have a quake style terminal dropdown.

I also think the taskbar showing only icons is confusing when we have the same app opened multiple time. I have a similar organization as you for work: a Firefox window on desktop 10 for Calendar, Mail, another on desktop 9 for company Chat, another (main) window on desktop 1, another on desktop 2 for a different project, ... By default on Gnome they would be all grouped into 1 Firefox icon. We can change the settings to not group apps, but a bunch of Firefox icons next to each others doesn't help either.

I recently discovered in the Fluxbox edition of MX Linux the taskbar Tint2. It was configured in a way that split the taskbar into dedicated and fixed workspace areas. It's an efficient way to see quickly what app is on which desktop, and clicking on one app will bring me to the desktop where the app is. I can also move apps to different desktop with the mouse by dragging them in the bar (for instance drag terminal of desktop 2 in desktop 3 next to the file browser opened there).

It looks like this: https://imgur.com/a/FGNfL7e

I currently use this taskbar with Openbox, but it should work with other DE/WMs. It has some bugs in some edge-cases so it's not perfect, but I like the concept.

I went on a quest to configure the same behavior on different DEs. I couldn't reproduce it with the default bars of Budgie, Cinamon, Gnome, Mate. KDE was the only one where I was close to achieve this. In the default KDE bar, it's possible to sort the apps by their workspaces. But it only sorts them, it doesn't split clearly by static desktops like you can do in Tint2. Still, KDE showed once again it was one of the most customizable :)

What I understand is that people that use multiple desktops they do because they might have 2 apps not fully maximized in desktop 1, another 3 in desktop 2, etc. But for me, I have maximized maximized apps 99.9% of the time, so I can not see and advantage on alt-tab to another app vs shift-alt-tab (or whatever the option) to switch desktops. Or am I missing somthing here?

For me, a desktop is a logical group of programs for something I'm working on, not necessarily a specific layout of non-maximized programs. This also allows for less pollution of programs when Alt+Tabbing, which you might benefit from.

In the desktop-centric organization, many people also have 1 app per workspace most of the time (I think). In a tiling WM, the app will take the full screen estate if it's alone there, so it's also maximized.

The difference with alt+tab is that switching to another workspace (which represents a window if the workspace has only one app) is deterministic, given the right keybindings setup and if we have some habits regarding the placing of windows.

So 99% of the time I have the same placement of windows in workspaces. At the very least my main Firefox on destkop 1, Code Editor on desktop 2, a terminal (related to my coding task) on desktop 3, and then things get more "dynamic", maybe some extra term or other stuff I may need for my task on desktop 4, 5, ... With the bindings Super+<number> (number row on top of the keyboard), I jump directly to my workspaces(windows). With my left hand I hit Super+1 and it will always show Firefox, Super+2 vim, etc...

I prefer it to cycling through alt+tab, hitting Tab multiple times until I find my window. Here's an example of a flow I was doing just earlier today:

win+2 (editor) : I edit code

win+3 (term1) : run command to build or run tests or deploy...

win+1 (firefox) : refresh the app I just built, click somewhere, test...

win+3 (term1) : see that the build actually failed

win+4 (term2) : check a quick solution in another term, use a CLI tool, do some tests in a repl...

win+2 (editor) : fix code

win+3 (term1) : build

win+1 (firefox) : refresh, prepare the page (input some text or something, ready to click a button)

win+3 (term1) : check if build finished

win+1 (firefox) : click the page button to test my change

The idea is that each time I switch to a different desktop/window, I just go there directly, without thinking, as I know where they are. The example I gave is the natural way I use my computer (with i3 or dwm, but can be configured with KDE, Cinamon...), so it's not a far fetched example at all (in my case). Switching back and forth is extremely fast that way. A long time ago, a colleague even told me I was a bit hard to follow when in pair programming sessions so now I try to slow down a bit. With Alt+Tab it's not as smooth, as we'd have to cycle through 4 windows. With the default implementation of most alt+tab out there, it's the opposite of deterministic, there's some logic (that I never fully understood) to go back to the windows in the order of last used/focused windows. But I know that in KDE at least it's possible to configure the behavior of alt+tab to make it loop in a "dumb" predictive way (1->2->3->4->1->2...), so in the end, it's again just a matter of personal preference.

If the bindings were less optimized (shift+alt+<number> or something) it would get uncomfortable to use. I use the Super modifier ("Windows logo" key) as the basis for all shortcuts related to my WM, so it doesn't conflict with the shortcuts reserved by the apps themselves (apps may interpret the modifiers Alt, Option, Shift, but not Super). It's a bit of finger-stretching to reach desktops higher than 5 on the number row, and at some point I need my right hand, but it works fine for me.

You're also correct that workspaces allow for more windows (very useful the 1% of the time I need it), and in that regard a workspaces organization is not comparable to a alt+tab based flow.

Organisation. Personal windows on one desktop,work ones on another, for example.

> Why? It comes down to how I manage my work. I often set up one browser window per project or topic, labeled accordingly (e.g., Project1, Project2). I even have a window called MAIN for my daily essentials like Mail, Calendar, WhatsApp, and Signal.

This is very similar to my workflow in some ways and I was just talking to another developer about this. I keep most of my projects on separate workspaces. I feel very confident in my ability to leave each workspace in whatever state I need. I'm on an AMD system running an open stack of GPU drivers, which means if I want I can pop open a new workspace and play a run of Darkest Dungeon if needed and leave that running on the workspace too.

Until earlier this year, KDE let anyone upload a "global theme" to the official KDE Store. When a KDE user browsed the themes, he received no warning that the themes could run arbitrary code and that no one at KDE was vetting the themes. Then in March 2024 a KDE user installed a global KDE theme which did rm -rf to his home directory:

https://www.bleepingcomputer.com/news/linux/kde-advises-extr...

So, although I agree with most of the criticisms of Gnome I see here on HN, I use it because I found a way to work in it without being annoyed too much by the overconfidence and lack of skill of Gnome's UI designers and the lack of customizability and because at least the Gnome project seems to pay enough attention to security to avoid a security hole as blatant and obvious as the one I just described in KDE.

Here is an example of Gnome paying attention to security:

>[Gnome] implements permission control for privileged Wayland protocols like screencopy. There are other desktop environments and window managers with Wayland support, but I am not aware of any permission control implemented by them.

https://privsec.dev/posts/linux/choosing-your-desktop-linux-...

I.e., Gnome is the only DE the author quoted above knows of that actually takes the trouble to make use of a design feature that X lacks, but Wayland has: namely, an architecture that allows a DE to prevent the deplorable situation of most Linux installs in which any app can read the contents of the window of any other app.