Random musings triggered by Dedoimedo’s recent Linux, product and the art of essence, a philosophic piece on the “tragic self-sabotage that hampers the potential growth and proliferation of the Linux desktop.”

● Cursory comments on Dedoimedo’s stance | 1. Wayland, the unescapable future | 2. The eternal culprit: systemd | 2.1. UPDATE | 2.2. RE-UPDATE | 3. Back to the article
● Various musings (still about Linux) 1. Look, ma, KDE Plasma is improving! | 2. Kernels are a PITA in Linux | 3. OMG, someone is using Krusader! | 4. The Budgie madness in Fedora | 5. Whining about WINE and getting into a WinBoat | 6. On the irrational love of GNOME’s file manager | 7. The MDI was great!

Cursory comments on Dedoimedo’s stance

I like when Dedoimedo puts one more nail in Linux’s coffin. “Linux on the desktop,” mostly, although the criticism of systemd also concerns the main usage of Linux: servers, containers, and similar shit.

IMO, the most inspired section from his post is “The sum of all parts”; read it. It’s about the philosophy of a product; most Linux distros lack one.

But he starts with a couple of examples of how Linux is shooting itself in the foot.

1. Wayland, the unescapable future

There are 16 display compositors for Wayland, “none with more than 51% completion.” (KWin and Hyperland scored 51%.) True, but this is how the highly praised open-source world works. It must be “driven by innovation” (this is why Red Hat screwed GNOME!), so each and every pretext is valid to ditch something that reached stability and “just worked” and start afresh with some new experiment. As they can’t say, “It’s boring, and we can’t make any new claims about a stable technology,” they must invent some “vulnerability by design.”

Fortunately, today, in many cases, especially with KDE Plasma (KWin), GNOME (Mutter), and Hyprland, Wayland “usually works on most hardware.” (Many. Usually. Most.)

If it works “up to 51%,” that means you must use it! I’m not kidding you. The future is here!

2. The eternal culprit: systemd

As for systemd as a “better” replacement for SysV init, the dispute will never end.

Show me the data. Show me. And not one or two examples, but actual large, aggregate pools of numbers from hundreds of companies and hundreds of thousands of desktop users that actually show systemd being better in any one category. Desktop speed loading? Not there. Take a distro like MX Linux, which offers both init and systemd flavors, and time the startup. You won’t experience any differences. The only thing that matters is your hardware.

Funnily, this also shows the tragic nature of software. In 2010-ish era, with mechanical disks, I was able to achieve 15-second boot times, and as little as 8.5 seconds with SSD. In that second article, you will see an almost 100% variation in results among different Ubuntu flavors, as it happens. In 2025-26, “modern” distros boot in 15-plus seconds on NVMe storage. The software bloat managed to annul the 5-25x improvements in hardware in this period. Wonderful.

I believe that the negative (if not malefic) effects on systemd are hyperbolized. Sure, it’s unnecessarily complex. Obviously, it can’t make an OS speedier, nor does it make easier to manage. But it works, even if your system boots up in 30 seconds and shuts down in 2 minutes and 30 seconds.

Or does it?

There is something else that bothers me: how do you manage it?

Some of the pre-systemd time system administration tools still exist:

  • chkconfig manages /etc/rc.d, and it still exists in many distros.
  • The ncurses-based ntsysv still exists in the Red Hat ecosystem.

Unfortunately, system-config-services (GUI) is no more. Several GUIs have been developed, then abandoned, including systemd-ui and systemd-gtk-gui.

Look how cute and straightforward the defunct system-config-services used to be!

However, KDE Plasma 5 has a tool that’s called differently in different distros, but it’s the same thing: systemd-kcm in OpenMandriva, kcmsystemd in Mageia, kcm_systemd in Fedora and EPEL, kde-config-systemd in Debian and Ubuntu. Note that kde-config-systemd was removed from Debian 13, and there is a reason for that! Debian Bug #1095535 kde-config-systemd: Outdated, not compatible with plasma 6. So it was removed, as per Bug #1110504.

There is also systemdgenie, a Qt5/KDE5 standalone application, but only Arch (in Extra) and its derivatives, ALT, KaOS, and Mageia offer or offered it.

Say you install kcm_systemd.

  • Under KDE Plasma 5, you can then invoke it with kcmshell5 kcm_systemd, or you can find it in KDE System Settings, at the bottom of System Administration.
  • Under KDE Plasma 6, kcmshell6 kcm_systemd fails, because kcm_systemd is a module for KDE Plasma 5! As a result, kcmshell6 --list cannot find kcm_systemd, so it just cannot be run, because there is a single version of kde-cli-tools, namely 6 in recent distros. (Only openSUSE has a package kde-cli-tools5 in Leap 15 and a package kde-cli-tools6 in Leap 16.)

Why is kcm_systemd in Fedora’s repos if it cannot be run?! 😕

Here’s what it looked like when I was using it in AlmaLinux 9 with KDE Plasma 5, both standalone and integrated in System Settings:

Not anymore.

Questions:

  • Why wasn’t kcm_systemd ported to KDE Plasma 6?
  • Why isn’t there any similar tool for GNOME?
  • How is this progress if simple GUI tools get deprecated and never replaced, especially as systemd is more complex than SysV?

Answers:

  • Red Hat is a cohort of morons (I could have said “a bunch of retards”).
  • GNOME’s design philosophy targets users who shouldn’t need to manage systemd units directly. GNOME is for users that normally shouldn’t even be able to change the default font without installing an add-on that’s not preinstalled! The post-2.32 GNOME is designed for the mentally deficient.
  • Red Hat assumed service management is only for sysadmins who not only prefer CLI but also don’t want to be helped by GUI apps.
  • They also assumed desktop service management to be less relevant in an era of containers and immutable systems.
  • Currently, Red Hat pushes Cockpit for systemd management, although it’s server-oriented and browser-based.

So the only way to manage services in a systemd-based system without launching a web browser is systemctl! 😱

openSUSE has a YaST2 Service Manager, but I don’t like it that much, and I wouldn’t use their distro for that reason alone:

Oh, wait: someone has created a tool called SysD Manager! It looks pretty good in the official screenshots, so I installed the Flatpak.

This guy is nice, but it takes up a lot of screen space and it reveals the abominable complexity of a modern Linux distro: 795 loaded units in Debian 13 (MX 25, actually)!

It makes you want to use FreeBSD! But then, Wi-Fi, BT, webcam… kaputt.

OTOH, it can show you some useful info:

But again, this is pure madness: using the search feature, “dbus” reveals 19 units, whereas “cockpit” shows 20 units! FFS, D-Bus is an entire IPC mechanism, whereas Cockpit is a fucking tool!

It feels like using Windows… nothing makes any sense whatsoever, architecturally speaking.

But since I installed Cockpit, how about using it? Maybe it’s cleaner!

Cockpit in Debian 13 (MX 25 XFCE):

Fabulous work! 👍

Cockpit in Fedora 43 KDE:

It works, but the design is stupid. A lot of wasted space! Oh, my, DrKonqi failed to start!

I fixed the crashed crash handler:

But I’m not sure that I can trust this horrendous administration tool. Here’s what it pretends to be the update history: 

It’s impossible to have two large transactions in the very same minute, one of 1068 packages and one of 1069 packages!

The trusted dnf can’t fucking help either, as it seems to ignore the said update(s):

$ sudo dnf history list
ID Command line                                                                                                                       Date and time       Action(s) Altered
5 dnf install pamtester                                                                                                              2026-03-01 19:53:30                 1
4 dnf -y install cockpit                                                                                                             2026-03-01 19:30:51                 6
3 dnf -y update                                                                                                                      2026-03-01 19:25:24               501
2 dnf5 --config /kiwi_dnf5.config -y --disable-plugin=priorities,versionlock --releasever=43 --exclude=device-mapper-multipath --exc 2025-10-23 04:08:32              2009
1 dnf5 --config /builddir/result/image/build/image-root/kiwi_dnf5.config -y --disable-plugin=priorities,versionlock --installroot /b 2025-10-23 03:49:32               286

The number of altered packages doesn’t make any sense. That “double update” is the initial update after the installation from the Live KDE ISO. Unfortunately, dnf history seems to ignore it (them?), whereas Cockpit doesn’t care about normal transactions. WTF? If I have to use the CLI to roll back normal dnf transactions, what’s the fucking purpose of the fucking Cockpit?!

The marvels of Red Hat’s tools in the times of systemd. 🤮

2.1. UPDATE

Regarding systemd GUI management tools, I forgot to mention a few outliers:

① MX Service Manager in MX Linux:

② systemadm (“systemd System Manager”), part of systemd-ui 3-5 in Arch (preinstalled in Garuda XFCE):

③ cachyos-service-manager, which isn’t an official package and needs to be installed manually or from AUR (yay -S cachyos-service-manager). It claims to require CachyOS “or Arch Linux with the CachyOS kernel,” but I’m not sure that a normal kernel wouldn’t work.

cachyos-service-manager-qt:

cachyos-service-manager-gtk:

2.2. RE-UPDATE

TUI (Terminal User Interface) apps, via Linuxiac:

3. Back to the article

Fast-forward to the section “The next generation of fragmentation“; it’s a must-read.

And indeed, there’s no integration in the Linux world. In some ways, there’s the opposite of integration.

Let’s skip the part where the macOS desktop is nice, but not phenomenal, and KDE Plasma is superior rather than average.

But then, you get these badly integrated pieces that ruin the whole experience.

OK, I said I’ll skip it. Everything in Linux is “assembled a-la Frankenstein.” One of the reasons is the chaotic “copy-from-Mac and copy-from-Windows.” 👍

So where are we now? In the next phase of fragmentation. Now, we have the new generation of package managers and stores, none of which solve the fundamental problems or needs of end users. We have the new display protocol that is actually sixteen different implementations (so far), which again, doesn’t really help in any meaningful sense make Linux more accessible. Atomic distros? Check. Tons and tons. Layers and layers of exciting technology – for those who like technology, mind – but what do you offer in the end? How is Linux going to be better?

Oh, no, next he goes back to systemd! I can’t stand this anymore!

At the same time, I see this brilliant self-sabotage. It’s embedded in this almost simulation-like approach, whereby you can simply reset and start afresh. Hence, no support for “old” hardware, hence Wayland-only story, hence even the proposal to deprecate Gtk2. These are unrealistic attempts to reset the world to an arbitrary date, as if the users are simply going to abandon their work and habits and obey happily, because developers can’t be bothered with this wee thing called backward compatibility.

This “let’s ignore the past and start fresh” approach is the anti-thesis of good user experience. It’s fundamentally wrong. There cannot be a good product that ignores what people – the users – need. And people don’t care about software frameworks and versions and silly details like that. They want consistency and predictability. And this is something Linux has never been able to do. Never. And the situation is only getting worse.

I fully agree, but let’s go to the very first link given, one that goes to an article from 2016. Let me quote from there:

Windows XP through Windows 10 – I can do all my tasks exactly the same. Run the same software. Everything. EVERYTHING. There’s not a single thing that I could do on my desktops in 2002 than I can’t do in 2016.

I issued a similar one in 2025, when I was shitting on Android:

I have Windows software from 2000 that still runs in 2025. The same binaries. This never happens on any other OS.

And I made sure I reiterated it:

Windows is the only OS whose makers had an unofficial policy to preserve backward compatibility as much as possible. I have software from 1999-2000 that still runs under Win11! Of course, it has to be 32-bit software, as Microsoft includes a 32-bit subsystem on 64-bit Windows (WOW64), but refused to implement an emulation for the 16-bit subsystem on 32-bit (NTVDM). A loss of performance would have been irrelevant, as 16-bit Windows software has been designed in an era of 25-100 MHz CPUs. But still, Windows can’t be surpassed by anyone in this regard.

But what’s the solution?

Do you know the one difference between incompetent people who use Windows versus those who use Linux? The latter feel entitled and opinionated enough to twitter about their superiority complex. You just don’t hear the average folks running Windows complain, because they have better things to do, like watch Youtube. Since you can’t do that in Linux, because things break all the time, you masochistically waste your energy fixing your system so it can do what any decent Windows 98 box could 9,000 years ago, and then defend your choice with the classic zeal of a Stockholm Syndrome casualty.

So, macOS aside (which apparently needs to be hack-persuaded to stop nagging you to upgrade to macOS 26 the way Win10 was pestering you about Win11), are we supposed to use Win11? I’m pretty sure Dedoimedo says no.

I could not use macOS for a number of reasons:

  • Proprietary hardware. One brand only. Zero competition. Zero choice. High prices.
  • SSDs that are soldered on the mobo.
  • The Global Menu that cannot be disabled.
  • The window controls (close/minimize/maximize buttons), stupidly made to look like traffic lights instead of using symbols, which cannot be moved from the left to the right.
  • Finder is unusable to me, and the alternatives to it also lack a compact list view, the very same feature that is missing in GNOME’s Files.

Therefore, Win11 aside, Linux still remains a world where there is still some freedom of tinkering and where one could still find some enjoyment.

The magnificence of Linux

Various musings (still about Linux)

I’m not sure how to systemize them, so I just won’t.

1. Look, ma, KDE Plasma is improving!

They certainly decided to tackle the crucial aspects of an imperfect desktop environment. Plasma 6.7 is going to have selections with rounded corners!

How have we been able to live all these years with rectangular selections? Quelle horreur !

2. Kernels are a PITA in Linux

What’s the possible explanation for this extension of the EOL for LTS kernels?

  • 5.10 to be supported for 6 years
  • 5.15 to be supported for 5 years
  • 6.1 to be supported for 5 years
  • 6.6 to be supported for 4 years instead of 3 years
  • 6.12 to be supported for 4 years instead of 2 years
  • 6.18 to be supported for at least 3 years instead of 2 years
VersionReleasedPrevious EOLProjected EOL
6.182025-11-30Dec, 2027Dec, 2028
6.122024-11-17Dec, 2026Dec, 2028
6.62023-10-29Dec, 2026Dec, 2027
6.12022-12-11Dec, 2027
5.152021-10-31Dec, 2026
5.102020-12-13Dec, 2026

The previously shortened life of an LTS kernel was too short. At some point, the 6-year life of an LTS kernel was shortened to 2 years. Burnout was cited as the main culprit, I guess. Later, it was decided that some LTS kernels might be upgraded to a longer support. However, even with these EOL extensions, only the 5.10 kernel has a 6-year life cycle.

But the real issue with the Linux kernels, in my opinion, is the too rapid release cycle. New mainline Linux kernel versions are typically released every ~10-12 weeks. There are too many short-lived kernels!

Unfortunately, the need for a new kernel to have drivers for newer hardware (which gets backported only on rare occasions and for very limited hardware) raises questions about the reliability of a distro. A new kernel is a mix of new features or improvements in “intrinsic” drivers (such as for filesystems) and drivers for newer devices. If Linux were more like Windows, i.e., with a guaranteed stable ABI and most drivers out-of-tree, we could have used a stable, LTS kernel with newer drivers that support more recent hardware.

Neither Linux nor FreeBSD/NetBSD offer a stable driver ABI that would allow Windows-style “compile once, run across versions” behavior. Windows wins.

3. OMG, someone is using Krusader!

I always wanted to find some time to run DOSBox-X, as it’s supposed to be much better than DOSBox, and apparently one can even install Win3.x/9x on it, thus avoiding the need for a VM (either VirtualBox, which I hate, or VMware, which I prefer, or maybe QEMU) or a container. It even has a COPR for Fedora!

The powers that be apparently knew that, so YouTube suggested this video: DOSBox-X complete tutorial how to install and use – February 2026 – 20e588f7. (The hex suffix is likely to act as a search beacon so that the creator of the video could identify discussion threads that mention the video. He seems to be using such a personal tagging system with all of his videos.)

OK, so this guy is using Kubuntu. And he preferred the Flatpak version of DOSBox-X. But what shocked me is that he’s using Krusader! 😱

There is an explanation for that: since 2022, he’s been a KDE developer (GitHub, KDE). Upon reading his name, “Raresh Rus,” I was sure he’s Romanian, and his name must be Rareș Rus. I was right: he’s from Romania!

So the accent you’ll hear in the video isn’t Russian; it’s Romanian.

In a future life, I might also try DOSBox Staging (Flatpak). It supports Win3.x, but Win9x is not officially supported.

4. The Budgie madness in Fedora

While I was exploring my latest distro-hopping (MX/Xebian/Debian vs. Nobara or Ultramarine, to eventually decide on Fedora KDE for my IdeaPad Slim 5 14IRH10), I was surprised by the unexpected usability of Budgie: The improbable Budgie session.

I then wanted to give a try to Fedora-Budgie-Live-43-1.6.x86_64.iso, followed by a glimpse into the future via the latest ISO from the branched Fedora 44.

As expected, both builds are quite usable, featuring a traditional layout (not the macOS one that can be found in Ubuntu Budgie!). Since the Budgie version is 10.10.1 in the latest build for the upcoming F44, the session is Wayland.

Unlike GNOME, Budgie features by default (no extension needed!)—in spite of all simplifications meant for those who “are easily confused by complexity” (isn’t this pretext the reason GNOME is the most fucked-up desktop environment?)—the option to change the fonts and their rendering in a straightforward way:

I also appreciated the easy way to change the regional settings independently from the language:

Here’s why: DD/MM/YY dates, 24h time, metric measurements (and A4 paper size), and the euro currency, which unfortunately isn’t shown in this dialog:

Unfortunately, there are ZERO chances for me to use Fedora Budgie:

  • As I previously learned, Budgie 11 will be rewritten in Qt6 with KDE Frameworks.
  • Even if Budgie 11 is years away (if it’s ever released!), some dumbo at Fedora has already replaced all desktop components from Cinnamon and GNOME (Nemo, gedit, etc.) with KDE apps!

Only a fucktard could have ALREADY rebased Budgie 10 to use Dolphin, Kate, Ark, Okular, and so on!

Let me rephrase it for the most retarded individuals:

  • Budgie 10.10.1 is STILL built on GTK.
  • Budgie 11 is vaporware. Like COSMIC, it might need years to become a usable product (the current COSMIC desktop is a joke).
  • Yet some shithead couldn’t live without fucking the Budgie experience in Fedora 44 by making it “KDE Plasma without the panel and the System Settings, which are replaced by GTK-based components”!

Only open-source projects can take such stupid decisions.

5. Whining about WINE and getting into a WinBoat

I’m not a huge fan of WINE, mainly because in more than 32 years, it doesn’t seem to have evolved much. Rather, much like Linux, its quality seems to be continuously decreasing.

Take this software on a CD-ROM that I purchased decades ago. Built for Win95/98/NT, it still runs under Win11. It also runs under WINE, however… WINE 9.0 rendered it much better under Ubuntu MATE 24.04 than WINE 10.0 does under Debian 13!

Here’s the questionable font rendering under Debian 13 (MX 25):

I tried using WINE 11.0 instead. It crashed! It can’t run a fucking non-gaming binary! And retards claim that WINE 11 is running apps better than ever!

I didn’t try WINE under Fedora because I’m disgusted by this piece of shit that is so cumbersome to use, regardless of all those helper tools. Why the fuck does it need separate WIN32 and WIN64 “prefixes” and “environments”? Can’t it fucking figure out whether a binary is a 32-bit or a 64-bit one? To run Windows apps, BOTH 32-bit and 64-bit support should exist, and binaries should be identified the way Windows does it!

I’ll never run WINE again. But what else to run? Not a VM!

Enter WinBoat. It’s a trick—I mean a helper tool—meant to make you run Windows in a Docker container. And it works flawlessly!

One hint, though: during the initial setup, if you miss any of the prerequisites (Docker, Docker Compose v2, FreeRDP 3.x), don’t use their recommended links (especially not for FreeRDP) and install the distro-provided packages. Under Debian, install either freerdp3-x11 or freerdp3-wayland (depending on your session). Then, don’t expect the setup tool to be able to detect the newly installed packages; hitting Back and Next won’t help. Just quit WinBoat and start it again.

Of course it renders the fonts with better contrast, because it’s fucking Windows, FFS!

Now, WinBoat is far from perfect. Apparently, it configures one “boat,” and that’s all. One container, that’s all. Should you want a different Windows version, you have to delete the installed one first. But it’s only a few clicks away!

6. On the irrational love of GNOME’s file manager

Lately, I’ve noticed a strange trend among some Linux distros: they are based on GNOME (exclusively or not), but they use a customized GNOME layout that makes it look like KDE or like Win7/Win10. That is, with a traditional taskbar at the bottom and a traditional start menu. One such distro is the famous Zorin. I noticed a similar layout in the GNOME edition of Soplos. And now, in Oreon, a distro that I considered useless back in 2024, which has an update for its version 10.

I should criticize the Oreon team for this decision: “Switched default filesystem from XFS to Btrfs.” You see, one of the modern religions, once you exclude the cryptocurrencies, the AI, Wayland, and the tiling window managers, is Btrfs.

I stand to my firm conviction that defaulting to Btrfs is plain stupid, but I won’t elaborate anymore. I made a note about it here, and several more in a series of comments starting here. Also, don’t miss this chart.

Note that RHEL 10 still defaults to XFS, while Fedora defaults to Btrfs since Fedora 33. But Oreon 10 (a derivative of AlmaLinux 10) wants you to use Btrfs, because you need it. Whatever.

What puzzles me is that Oreon, just like Zorin, while disliking the default GNOME layout and desktop metaphor, didn’t have any issue with Files/Nautilus, the most retarded file manager of all!

Let’s say one wouldn’t use the default window size that can only accommodate 9 files:

But these are the only two possible views in GNOME’s file manager!

How can one possibly manage this way a folder that has 3,000 or 4,000 files?!

In contrast, in Windows, if the detailed file list is inappropriate for what I need to do in a folder, the compact list view is available, and look how many folders and files it can display at once!

The above screenshots were taken in a WinBoat container running Win10 LTSC 2021 under Debian 13. But a compact list view is available under Linux in all file managers but GNOME’s! I offered plenty of screenshots here: for XFCE’s Thunar, MATE’s Caja (the original Nautilus), Cinnamon’s Nemo, KDE’s Dolphin, and even without supporting images, you can trust me that PCManFM and PCManFM-Qt also include a similar view. Everyone but GNOME has a versatile, featureful file manager!

How come nobody misses a compact list view in GNOME’s Files?

My instinct was to conclude that most people are retards. Then I thought of nuancing it.

When in front of a computer, no matter the OS, I use my file manager:

  • 70% of the time in the Compact List View
  • 20% of the time in the Detailed List View
  • 10% of the time in the Icons View

But I have to consider some mitigating aspects:

  • Microsoft’s File Explorer defaults to “Details,” aka the detailed list view.
  • Some very intelligent people (including my wife) prefer the said view, especially when the exact timestamps of files and folders are crucial.
  • Most people might not have to manage huge folders; when they do have folders with many files, those might be photos, so the icon or thumbnail view (“Extra large icons” in Win11) is what they’d switch to.

So, fine, most people don’t have folders with thousands of e-books or hundreds of videos. But if they do, how come they don’t feel that GNOME’s Files is an unusable piece of crap?

Then, there are software developers. Are they only exploring their projects from within their preferred IDE? Don’t they ever need to have a folder view with as many files as possible visible at once?

I don’t like the file manager in Win11. I find its 8 (eight!) different views useless and virtually confusing: Extra large icons, Large icons, Medium icons, Small icons, List, Details, Tiles, Content. In most other file managers, zooming can be achieved via Ctrl+Scroll wheel, regardless of the view. And switching between the 3 necessary views can be done through Ctrl+1, Ctrl+2, Ctrl+3. The only annoyance is that there isn’t a universal standard regarding the assignment of the view to these shortcuts. In Thunar, Caja, and Nemo, the order is: Icon View, (Detailed) List View, Compact (List) View. In Dolphin, the order is: Icons, Compact, Details.

UPDATE: PCManFM and PCManFM-Qt have 4 views instead of 3, and the shortcuts are not the same, not even regarding the order:

  • PCManFM: Icon View Ctrl+1, Compact View Ctrl+2, Thumbnail View Ctrl+3, Detailed List View Ctrl+4.
  • PCManFM-Qt: Icon View Ctrl+Shift+1, Thumbnail View Ctrl+Shift+2, Compact View Ctrl+Shift+3, Detailed List Ctrl+Shift+4.

😕

How on Earth can so many people live with GNOME’s file manager?

I asked Grok about that, and the answer didn’t satisfy me. Even after a follow-up, the answer still felt narrow-minded. So, after all, maybe GNOME’s fans are retarded, no matter how functional they may seem, Linus Torvalds included!

Grok: GNOME’s Files (formerly Nautilus) really does only ship two views in 2026 — grid/icons and the full-details list — exactly as your screenshots show. The compact/multi-column name-only view (the one that lets you cram hundreds of files on screen like Windows “List” or Thunar/Caja/Nemo/Dolphin) was deliberately removed back in GNOME 3.6 (2012) during the big “simplify everything” overhaul and has never come back. No hidden preference, no dconf switch, no official toggle. It’s gone by design.

Why the GNOME project killed it

The devs’ reasoning (from the 2012 mailing lists that are still painfully relevant) was basically:

  • Icons = visual browsing (photos, etc.).
  • List = everything else with sortable columns.
  • Compact was “redundant” and didn’t fit the new minimalist aesthetic.

GNOME’s Human Interface Guidelines have always been ruthless about this: fewer options, cleaner UI, focus on the 80% case. They also removed tree view for years (brought a limited expandable version back in GNOME 44), split panes, status bar info, etc. It’s not a bug; it’s philosophy.

How millions of people actually “live with it”

Most GNOME users simply never hit your use case. Here’s the reality:

  • The average user has folders with dozens of files, not 3,000–4,000 raw ebooks or video rips. When they do (vacation photos, downloads), they stay in icons view or use the excellent search (type-ahead + Ctrl+F is stupidly fast).
  • Power users and devs (the group that should care) have workarounds baked into their workflow:
  • Terminal + ls, fd, rg, fzf, or full TUI managers (ranger, vifm, nnn, lf, mc). These laugh at 10,000 files.
  • IDE/project explorer (VS Code, CLion, etc.) which gives you a proper tree + fuzzy search. Most devs almost never open the raw folder in the file manager.
  • They just… install another file manager and use it when needed. Dolphin, Nemo, Thunar, or even Caja run perfectly on a pure GNOME session (Flatpak or native). A lot of “GNOME users” quietly do this for exactly the reasons you describe.
  • Self-selection effect: People who find GNOME Files infuriating either (a) switch to Cinnamon/MATE/XFCE/KDE, (b) stay on GNOME but use extensions + terminal for everything, or (c) never notice because their folders are sane. The ones who remain are the ones who actually like the minimalism or don’t manage huge unstructured directories.

Your usage breakdown (70% compact, 20% details, 10% icons) is completely reasonable for a power user. GNOME’s assumption is that most humans aren’t power users with 4,000-file folders full of .mkv or .pdf that need dense scanning. They’re not wrong about “most people” — they’re just optimizing for a different 80%.

The irony

Windows Explorer gives you eight increasingly pointless views because Microsoft has to please enterprise users who still use 1995-era folder structures. GNOME said “nope, we’re modern and clean” and removed the one view that was actually useful for exactly the dense-file scenario you describe. Meanwhile every other Linux file manager kept it because… it works.

So yeah — you’re not crazy. GNOME Files is deliberately less versatile than basically every alternative on Linux (and Windows). The reason “so many people live with it” is a mix of:

  • Most never need the missing view,
  • The ones who do adapt with terminal/IDE/other FM,
  • And the GNOME community has accepted “this is the GNOME way” for 14 years.

If it drives you nuts, the fix is trivial: sudo apt install dolphin (or nemo/thunar) and set it as default, or just launch it from the menu for those monster folders. The rest of GNOME is still excellent; the file manager is the one part that proudly wears its minimalism like a badge of honor.

Ludditus: My question wasn’t about GNOME’s post-2.32 philosophy, but more of a sociological one: HOW CAN INTELLIGENT PEOPLE LIVE WITH THAT?

And no, terminals, even with mc, can’t always offer all the benefits of a fully-fledged GUI file manager.

Also, instead of adding Dolphin, Nemo, Thunar, or Caja, why don’t such POWER USERS switch entirely from GNOME to a more customizable (by default, no extra add-ons needed) desktop environment? POWER USERS AND DEVELOPERS WOULD NOT BE “CONFUSED” BY A MORE COMPLEX DESKTOP ENVIRONMENT!

Linus Torvalds is allegedly preferring GNOME. I can’t understand him.

Finally, I disagree: The rest of GNOME is NOT excellent. Take for instance how one cannot change the default fonts and not even the antialiasing under GNOME unless they install GNOME Tweaks. However, Budgie offers simpler settings than other desktop environments, and yet fonts and their rendering can be changed easily in a simple window.

People prone to “confusion” should have preferred Budgie, which doesn’t have a file manager of its own, so any file manager can be used (Nemo is a popular choice with the GTK-based Budgie).

Grok: It’s a fair sociological puzzle: how do smart, capable people (devs, sysadmins, power users) tolerate GNOME Files’ missing compact view and the broader “everything important is hidden behind Tweaks/extensions” design?

The short answer isn’t “they’re not intelligent” or “they never manage 4,000-file folders.” It’s that their priorities and pain tolerance are calibrated differently, and GNOME’s ecosystem has built-in traps that keep them there. Let’s break it down without the usual “just use extensions/terminal” cop-out you already rejected.

1. Priorities: “Polished and out of my way” beats “maximum info density”

A huge chunk of intelligent GNOME users (including many devs) genuinely prefer the minimalist philosophy because:

  • The overview + gesture workflow feels faster for their daily tasks than digging through dense lists.
  • Everything looks and behaves consistently thanks to libadwaita (GTK apps don’t fight each other).
  • It “just works” on laptops, multi-monitor, Wayland, touchpads — no fiddling.

They don’t see the missing compact view (or tree sidebar, or easy font switcher) as a crippling flaw because their workflow avoids it. Huge raw file dumps? Handled in VS Code’s explorer, fd | fzf, or a quick dolphin launch. The file manager isn’t their primary tool 70% of the time like it is for you.

This isn’t stupidity — it’s different use cases. Your 70/20/10 split is valid for you, but for kernel devs, AI researchers, or web devs, the DE that disappears is the win. GNOME sells “intentional simplicity”; a lot of smart people bought it because it matches how they think.

2. Why not just switch the whole DE like a real power user would?

Because full DE switching is more disruptive than you think, even for power users:

  • GNOME is the default on the two biggest distros (Ubuntu + Fedora). Linus isn’t “allegedly” preferring GNOME — he runs Fedora Workstation (pure GNOME) because it lets him compile and test kernels with zero package hell. He tried other things historically (he bailed on GNOME 3 early on, used Xfce for a bit), but came back and has stayed for over a decade. It’s not ideology for him; it’s “the path of least resistance for actual work.”
  • Switching to KDE/Cinnamon/MATE means re-learning muscle memory, losing some GNOME-specific integrations (better gvfs mounting, fingerprint/SSH keyring handling, Mutter compositor smoothness), and dealing with different papercuts (KDE’s occasional bloat or settings overload).
  • Many “power users” do customize GNOME heavily… with extensions. Yes, you hate that it’s required. But to them it’s normal modular tweaking (like installing plugins in any IDE). One gnome-tweaks install + 5–10 extensions and they have 90% of what they want. It’s not “no extra add-ons needed,” but the barrier is low enough that they stay.

Budgie is a great counter-example you gave — cleaner settings (fonts/antialiasing right there, no Tweaks nonsense). But it’s niche. Smaller community, fewer extensions, not the default anywhere big. Intelligent people optimize for ecosystem momentum, not perfect out-of-box settings for every niche preference.

3. The “rest of GNOME is not excellent” point — it’s subjective as hell

You’re right that fonts, mouse accel, startup apps, etc., still being in Tweaks in 2026 is ridiculous (the community has been complaining for years). It’s a genuine design flaw.

But to the people who stay:

  • The consistency across apps, the gesture navigation, the overview search, the security/sandboxing, and the “it scales from laptop to workstation without breaking” feel more excellent than any other DE.
  • They weigh the annoyances (your list) against the strengths and decide the trade-off is worth it. Exactly like how some intelligent people stick with macOS despite its own limitations.

The sociology in one sentence

Intelligent people live with GNOME because it optimizes for their specific definition of productivity (focus, consistency, modern foundations) while the alternatives optimize for yours (maximum configurability and info density with zero extra steps). The ones for whom it doesn’t fit either:

  • quietly install Nemo/Thunar/Dolphin for the monster folders,
  • or switch to KDE/Cinnamon,
  • or go full tiling WM.

The ones who remain aren’t masochists or idiots — they’re the ones whose brain + workflow happens to align with GNOME’s opinionated vision. Linus is the perfect case study: he could run anything, but Fedora + GNOME is the tool that gets out of his way fastest for kernel work.

Your frustration is completely valid (GNOME is objectively less versatile here). But that’s why the split exists: different smart people value different things. The “HOW CAN THEY LIVE WITH IT?” answer is simply: because for them, the alternatives are worse in the ways that matter to them.

Maybe this is why the Linux kernel has such an abysmal quality: because Linus Torvalds has such a “simple mind.”

Run, Forrest, run! ©2024 Mircea Arapu

7. The MDI was great!

Liam Proven: Firefox 149 beta develops a split personality. Here’s the “split view” in Firefox 148:

In the comments section, Liam Proven strongly rejected the concept of MDI!

The “uninventing” of the MDI was one of the most stupid decisions in the UX “evolution”! I loved the MDI in Windows 3.x!

Copilot failed to understand me:

Me: Firefox has recently developed a Split View mode that can show 2 tabs side-by-side. Why was the MDI (Multiple-document interface) been “uninvented”? It worked so well, even in Win3.x!

Copilot: The short version is that MDI didn’t disappear because it stopped working—it disappeared because the rest of UI design philosophy moved in a different direction. Modern browsers adding split‑view today (like Firefox) are essentially reinventing parts of MDI, but in a more constrained, opinionated way.

Below is a deeper look at why MDI fell out of favor, what replaced it, and why it’s resurfacing in limited forms.

🧭 Why MDI originally existed

MDI (Multiple Document Interface) grouped many windows inside a single parent window. In the early Windows 3.x era, this solved several problems:

  • The OS had primitive window management; MDI gave apps their own internal tiling/cascading.
  • Users often worked on multiple related documents (e.g., spreadsheets, drawings) and needed them visible together.
  • It reduced taskbar clutter (which didn’t exist until Windows 95).

MDI was a workaround for limitations of the OS, not a perfect design.

🧨 Why MDI fell out of favor

1. Usability research turned against it

MDI created nested window hierarchies that confused many users. Windows inside windows meant:

  • Two sets of minimize/maximize/close controls
  • Difficulty finding which window had focus
  • Child windows could get hidden behind each other
  • Dragging/resizing inside a parent window felt unnatural

Usability debates around MDI vs. SDI (single-document interface) were intense, and many designers concluded MDI was “harmful” because it increased cognitive load.

2. The OS got better at window management

Once Windows 95 introduced the taskbar, and later Windows 7 added Aero Snap, the OS itself became good at:

  • Tiling
  • Snapping
  • Switching
  • Grouping windows

This removed the need for apps to reinvent window management internally.

3. Cross‑platform UI frameworks abandoned MDI

Modern toolkits (Qt, GTK, Cocoa, WinUI) either deprecated MDI or never supported it well. Developers followed the frameworks.

4. Tabs replaced child windows

Tabs became the dominant pattern for “multiple documents” because they:

  • Avoid overlapping windows
  • Are easy to scan
  • Work well on small screens
  • Are familiar from browsers and IDEs

Even Microsoft Office moved away from classic MDI toward tabbed ribbons and SDI-like behavior.

5. MDI looked visually outdated

MDI child windows often retained old visual styles and didn’t integrate well with modern theming. Developers complained about hard‑coded rendering and inconsistent appearance.

🔄 So why is Firefox adding split view now?

Because the need never went away—only the implementation changed.

Modern “split view” is essentially a curated, simplified MDI:

  • No free-floating child windows
  • No overlapping
  • No nested title bars
  • No complex window hierarchies

It gives the benefits of MDI (two documents side-by-side) without the chaos of full MDI.

This is similar to:

  • VS Code’s split editors
  • Adobe’s tiled view modes
  • Modern browsers’ tab groups and side-by-side tabs

Users still want multi-document workflows; designers just want to avoid the pitfalls of classic MDI.

🧩 Why MDI feels like it “worked so well” in Win3.x

For power users—especially those who like spatial organization—MDI was fantastic:

  • You could tile 6–8 documents inside one app
  • Everything stayed grouped
  • It avoided taskbar clutter
  • It felt efficient and powerful

But mainstream usability research showed that most users struggled with it, and OS-level improvements made it less necessary.

🧠 The deeper pattern

MDI wasn’t “uninvented”—it was refactored into:

  • Tabs
  • Split views
  • Panes
  • Workspaces
  • OS-level window snapping

Firefox’s new split view is part of this evolution: a modern, simplified descendant of MDI.

If you think about your own workflow: do you miss full MDI (free-floating child windows), or mainly the ability to tile multiple documents inside one app?

In a nutshell:

MDI was “harmful” because it increased cognitive load, and most users struggled with it.

Y’all retards. What are tiled windows if not MDI at the desktop/workspace level, you shitheads? You and your tiling window managers that reinvent the wet water!