SJVN and his dogs got soft-headed
Steven J. Vaughan-Nichols is possibly the most famous business and technology journalist ever. I used to read him in BYTE and PC Magazine in the heyday of the PC. Lately, I am more and more disappointed by his opinions, even more than I am disappointed by Liam Proven’s from the other side of the pond.
I do not mean any disrespect, but his stubbornness in stating that immutability is the future of Linux (and of computing) is decidedly harmful and pisses me off! Yes, we will eventually reach that point of no freedom, with everything containerized in all OSes (the Android model of sorts), but does this glim future need to be brought closer by SJVN?
I will discuss his latest piece, What the Linux desktop really needs to challenge Windows, in which he’s terribly wrong in my opinion.
My stance on Linux on the desktop
Nothing new for my very few constant readers, but I have to repeat it for the search engines.
3 months ago, I said I should abandon Linux forever. I even installed Win11 IoT on my oldest laptop, just to see. But I complained very strongly (sic) about Linux in May 2024 (abstract), in September 2021, and even as early as in 2007. For the time being, though, after having traded Ubuntu MATE 24.04 LTS for Lubuntu 25.10 for a while, I am stranded (for lack of a better word) in MX 25 XFCE, but I could as well have installed debian-live-13.2.0-amd64-xfce.iso and further configure my system. That is, on my newest laptop, which is a €400 cheap Acer. I am frugal.
Let’s separate my complaints about Linux and the reasons “Linux on the desktop” is never going to be a thing.
Linux is broken by design because…
❶ 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.
In Linux, even a statically compiled binary (which is atypical anyway) would not run on a Linux distro much newer than the one on which it had been compiled (we assume it’s a commercial, closed-source software, or that we just can’t have a toolchain able to rebuild it from sources on contemporary distros). You see, the Linux kernel lacks a stable ABI, because Linus Torvalds doesn’t want it. And if it’s a dynamically-linked one, those old libraries couldn’t run or couldn’t even be installed on modern distros. This is Linux, baby! You can only run software built for the specific versions of the libraries existent in your Linux distro. (Flatpaks, snaps, and AppImages are attempts to work around this situation.)
As a side note, Apple has broken several backward compatibility (including via emulation) several times in the past (Apple sheeple never complain, they just pay) and Google, when it doesn’t steal your money, actively tries to prevent you from installing apps designed for older versions of Android.
❷ In Linux, drivers cannot be installed separately from the kernel, the way you can install Win10 and even some Win7 drivers in Win11, and they would still work. Even DKMS drivers need to be built for the exact version of the active kernel. So what do you do when a new kernel stops supporting a device, or it introduces a regression? You suck it, that’s what you do?
But surely regressions are eventually fixed in Linux the same way regressions are eventually fixed in Windows, right? Right? Wrong!
EXHIBIT ONE: I experienced a regression in an audio jack, if that’s even possible! In Dec. 16, 2020, they fixed the support for the headset jack in Acer Travel Mate laptops P648/P658, and they broke it forever in Acer TravelMate laptops P645, which is my old laptop from 2016. This regression is going to last until the end of life on Earth.
EXHIBIT TWO: A completely unacceptable regression in a MediaTek BT chip! The BT support for MT7663 was removed from the kernels 6.12 and above, and also from 6.11 builds newer than a certain minor update (in Ubuntu 24.04 LTS, it worked with 6.11.0-17 but not with 6.11.0-26 and newer), declaratively because MediaTek decided to abandon the Linux driver. But how can it still be included in the 6.8 kernels, which are still updated in Ubuntu 24.04 LTS? This regression is also going to last until the end of the Universe.
❸ The update model in Linux is “nothing or everything”! This means, if you’re running a rolling-release distro, you cannot opt out of any specific upgrade of an app or library: everything will always be in the latest version (with a certain delay). But if you’re running a point-release distro, you will only get security fixes and some other patches, but no version upgrades of anything! This is even more annoying for EL or any other kind of LTS distro. Sure thing, for such cases, backports, extra repos, PPAs, COPRs, Flatpaks, snaps, even AppImages might provide partial fixes. (And Fedora is inconsistent because, during a point release, it upgrades the kernel and the desktop environment, but not other software.)
Funny thing, Android and iOS work as a rolling-release distro, but only regarding user software, not the OS itself. You are supposed to keep all apps updated, without exception! Sure thing, you can configure them on “manual update” and update them one by one, but this only delays their update. Some apps would even refuse to run unless they’re updated to the last version! And, once you update an app, there’s no way to roll back to a previous version. That’s the centralized “one app store for all apps” model, similar to a distro’s repositories, where a general upgrade affects all the installed packages. In contrast, in Windows, it’s your choice whether to update any given 3rd-party program, except for apps from the Microsoft Store. Linux will never adopt Microsoft’s model! And there’s always a “sound” reason: “oh, but the vulnerabilities present in the software that is not up-to-date!”
“Linux on the desktop” will never happen because…
❶ There is absolutely no major hardware vendor committed to produce laptops certified to be compatible with Linux. Not with a Red Hat or an LTS Ubuntu from 3 years ago, but with all present and future kernels! This would require components whose drivers are maintained!
RHEL-certified or Ubuntu LTS-certified systems do exist, but they’re few and far between. They cannot be found in the normal places people use to purchase laptops, being them online or brick-and-mortar stores. When something like 0.1% of the available laptops are certified to run Linux, with every component supported (Wi-Fi, BT, webcam), you can’t possibly have a “year of Linux on the desktop”!
❷ The major commercial distro makers, and by this I mean Red Hat Inc., Canonical Ltd, and SUSE S.A., couldn’t care less about the public! They only care about the enterprise users!
RHEL is not targeting the public at all. And Fedora is not reliable enough for general use.
Ubuntu LTS does target the public, but the rare laptops that come with it preinstalled almost always have the previous LTS version (with severely outdated software) and the main (GNOME) flavor. I’ll talk about the choice of a DE later.
Debian doesn’t have a commercial entity behind, but only Software in the Public Interest, Inc., a US 501(c)(3) non-profit organization.
openSUSE Leap is not actively marketed for OEM inclusion. In version 16, they even stopped releasing Live ISOs with KDE, XFCE, and GNOME (for some reason, they’re called “appliances”).
Arch is for guinea pigs, and Manjaro GmbH & Co. KG is too small to be of relevance.
Or, some big hardware name (Acer, ASUS, Dell, HP, Huawei, Lenovo, MSI) could have adopted Clem’s Linux Mint, however butt-ugly it is!
❸ GNOME killed the Linux desktop. While this is Red Hat’s fault, Canonical is actively supporting Red Hat’s nonsense once they adopted GNOME as the official desktop environment.
I very strongly believe that GNOME is designed by retards, for retards. (Yes, Linus Torvalds is a highly functional, extremely IT-competent retard.)
GNOME is designed for tablets and touchscreens. GNOME is breaking the desktop metaphor familiar for decades to people who weren’t born yesterday.
GNOME perverted the “less is more” principle to mean “you don’t need what we say you shouldn’t need, and configuring things that are easily configurable in all other desktop environments require extension, hacks, and shit.”
GNOME stripped Nautilus from features that every other GUI file manager has, especially a view mode that would display the files in the current directory as text on several columns, without details. It’s called “Compact List” or similarly in Dolphin, Thunar, Caja, Nemo, PCManFM, PCManFM-Qt. I cannot manage a folder with 3,000 files in GNOME’s Files if I can only see 30 files at a time!
GNOME tries to mimic some aspects of macOS, but inconsistently. For instance, Canonical should have opted to stick to Unity, as it has a global menu. And Files (the dumbified Nautilus) is dumber than Apple’s Finder (which lacks a “Compact List” view, but has Miller columns).
But the worst offender (after the file manager issue) is everyone’s fascination with Dock. A panel with icons that cannot have text alongside the icons brings computers to the pre-1995 era! Not only does one give up the opportunity to have relevant information about all the open windows without moving a finger, but it’s even difficult to tell apart a launcher from a running app! A dot, a dash, or a color?! Who the fuck designed such shit? This fascination also contaminated Windows itself, KDE, Cinnamon, and XFCE, but at least the “Win95-like” taskbar behavior can be restored in those environments (and, in Win11, via 3rd-party software). GNOME cannot have it back.
Red Hat is a cohort of severely mentally deficient people. Canonical, too. Options exist, but they lack the common sense and even the IQ to consider them!
- For its full configurability, KDE is the best desktop environment for Windows people. It’s the desktop of choice of Igor Ljubuncic aka Dedoimedo 🙂
- For its choice of several predefined layouts via MATE Tweak, MATE should have been adopted at least by Canonical. MATE Tweak can make MATE look like anything you want, even like macOS! It used to be my desktop of choice. But I have reasons to believe that MATE is moribund. 🙁
- Finally, XFCE is less demanding and very configurable but “incomplete” in the sense that it requires external accessories from MATE or GNOME to offer a full desktop experience: a PDF viewer, for instance. It could have benefited from some love, eh? It’s the desktop of choice of Liam Proven 🙂
SJVN’s idée fixe
SJVN is a long-time adept of immutability. Take a look at what he wrote almost two years ago, on Jan. 4, 2024: What is immutable Linux? Here’s why you’d run an immutable Linux distro. Subtitle: “Safety and security are immutable Linux’s calling cards.” Yeah, what else?
The definitions are correct:
Immutable Linux distributions come with a read-only core system. This means the base operating system, once installed, cannot be modified during regular use.
Patches, including system updates, are done, during a reboot. This is called an atomic upgrade, which means that the update of everything is handled as a single transaction. If something goes wrong, you can easily revert to its previous state.
This architecture approach significantly enhances system security and stability, as it prevents unauthorized changes and reduces the risk of system corruption. Instead of updating things piecemeal the way most major Linux distributions do, everything is updated at once.
Why? Because since the core system is read-only, it’s far less vulnerable to malware and tampering. If you can’t add anything to the operating system, you can’t corrupt it. In addition, immutable distributions use containerization for applications. This further isolates programs from the core system and each other. So, even if you do have a bad application, it has far less access to the underlying system than, say, a Windows application does to Windows.
So far, this is very much like the Android model, which I deeply hate! The OS is updated all at once, and I don’t have any control over it! Is Linux supposed to “evolve” in a system that cannot be rooted and tinkered with?
This comes at a price:
To install applications, you use containerized applications and universal package formats such as AppImage, Flatpak, or Snap. While many old-school Linux users hate this approach, this containerized method for application installation has several advantages.
First, they’re distribution agnostic. That means if you have a Flatpak of Discord, you can install and run it on any Linux distribution that supports Flatpak.
These containerized package systems also avoid the dependency issues that come with traditional package managers. The old-school package managers often must update not just the applications, but all their software dependencies as well. Sometimes that isn’t possible. Meanwhile, the container packages contain all the software you need to install and run its program.
Sorry, but I am not buying this shit. I’m an old-school user of everything, not just of Linux! Starting with CP/M (Felix M-216 with 8086), if not SIRIS (on Felix C-256, a clone of the French IRIS-50) and a clone of RSX-11M (running on CORAL 4011). I cannot code in assembler anymore, but stop removing freedom from me! I accept using Flatpak as a choice and as a last resort, but not as a mandate!
One interesting difference between an immutable Linux distro and Android is that the immutable Android OS image contains a number of GUI apps that get updated from the Google Play Store upon the first connection, so they take twice the space. On the other hand, Flatpaks and Snaps might require several versions of the GNOME libs and KDE libs, so the result is even worse.

Back to the piece from December 22, 2025, What the Linux desktop really needs to challenge Windows.
I’ve run Linux desktops since the big interface question was whether to use Korn or Bash for your shell. Before that, I’d used Unix desktops such as Visix Looking Glass, Sun OpenWindows, and SCO’s infamous Open
DeathtrapDesktop.Unless you’re a fellow gray-haired computer or Unix geek, chances are you’ve never heard of, never mind used, any of these. Fast-forward to 2025, there are more than a dozen significant Linux desktop interfaces. These include GNOME, KDE Plasma, Cinnamon, MATE, and on and on. They’re all too likely to be as forgotten as the first three Unix interfaces I named. Why? The same reasons you don’t know a thing about the Unix desktops.
Bullshit!
Primo, OpenWindows and Open Desktop weren’t that bad as visual metaphors! Quite the contrary.
Secundo, by writing “more than a dozen significant Linux desktop interfaces,” SJVN mixes the true desktop environments with the so many tiling window managers invented by and for brainwashed hipsters. Whoever wants to pretend they’re honest shouldn’t write “GNOME, KDE Plasma, Cinnamon, MATE, and on and on” but at least “GNOME, KDE Plasma, XFCE, MATE, Cinnamon and so on,” because XFCE was there before this failed Cinnamon that only members of the Mint sect care about! Also, LXQt is very much alive.
SJVN managed to piss me off very badly from the beginning.
I’m old fashioned about my desktops. I want the power in my PC, not in the cloud.
Indeed. Except for the case when I’d genuinely want to run an LLM. Running it locally has absurd hardware requirements unless you don’t care to get garbage for output.
I also like my old, but not yet ancient, PCs to keep working.
OK, now he will get to “Win11 is bad, use Linux Mint instead,” right? Oh, but there are regressions in Linux, too. Whatever.
These reasons alone have given the Linux desktop a boost. By my count, as much as 11 percent of the desktop market is now running Linux one way or another.
Wrong count. Don’t fucking include ChromeOS, which is for retards! You made me feel very badly about you. That was the second stroke.
Unix died because of endless incompatibilities between versions. Linux succeeded on servers and everywhere else because it provided a single open operating system that everyone could use. With the desktop, though, we saw, and still see, endless incompatibilities.
This is exactly my take. There is no backwards compatibility at all!
Linus Torvalds also saw this. He’s long thought that we have way too many desktops. He’s right. If someone goes to DistroWatch, they’ll find upwards of a hundred desktops. Who has time to figure out what’s best? I don’t, and I cover this stuff for a living, and once ran a site called Desktop Linux.
No, there are NOT “upwards of a hundred desktops,” you old moron! Third strike, 25 to life, no parole!
There are exactly 14 desktop environments if we make sure we don’t count the window managers: GNOME, KDE Plasma, XFCE, MATE, Cinnamon, LXQt, Budgie, Unity, COSMIC, Deepin, UKUI (from Ubuntu Kylin), Pantheon (from elementary OS), Trinity (if this is a thing), LXDE (frozen). Then, there are 8 surviving stacking window managers (Openbox, Fluxbox, JWM, FVWM, IceWM, Window Maker, AfterStep, Blackbox) and at least 14 tiling window managers of significance (i3, Sway, AwesomeWM, Xmonad, bspwm, Herbstluftwm, Qtile, Dwm, Hyprland, River, dwl, Hikari, Labwc, Wayfire). There are certainly more “entities” than the above 36, but:
- Most of them are utterly irrelevant, and especially irrelevant regarding the adoption of Linux on the desktop.
- “Linux on the Desktop” could have happened by focusing on a few of the aforementioned 14 desktop environments. Nobody cares about your “upwards of a hundred desktops”!
That’s just the surface of the problem. Under that, you’ll find arguments over how to manage software packages and the library incompatibilities they must deal with. Distro builders constantly have fits building and rebuilding programs to run on their Linux distros. The traditional ways of delivering Linux desktop apps, such as DEB and RPM package management systems for Debian and Red Hat Linux, respectively, simply don’t scale for the desktop.
They fucking do scale! Don’t mistake poor distro management, poor project management, and poor management as a rule with “they don’t scale”! Distros cannot be consistent when:
- Retards push unfinished technologies such as Wayland and then decide to stop supporting X11!
- Tried and tested frameworks, libraries, or technologies are abandoned and replaced with brand-new experiments because the developers were bored with the old ones that worked!
- Despicable people force snaps instead of packages for no good reason other than preparing the path to a future immutable-only desktop!
But SJVN has one God and one God only: containerization, which, in this context, leads to immutable distros!
We have the answer: A containerized software package delivery program that bundles all required dependencies into a single, useful package. Today, we should all be using Flatpaks, Snaps, and AppImages to install programs instead of worrying about library incompatibilities and the like. This also saves vendors a lot of time since they can deliver a universal version of their program that will install and run on anyone’s Linux desktop without the hassle of porting it to each and every distro.
SJVN is Satan’s tool! I’m an atheist, of course, but he’s still Satan’s tool.
The problem? It’s fragmentation once again. Some people hate containerized packages because they use more storage space and RAM than old-style programs. Others dislike one or the other packaging system for other reasons. For example, my favorite desktop operating system is Linux Mint. However, Mint’s leaders don’t like Snap because its parent company, Canonical, has too much control over the Snap store and has used Snap to replace some of the apt package installation program’s functionality.
To have Linux Mint as one’s preferred distro is another proof of cognitive decline.
On the other hand, of course it’s fragmentation if you have them all at once on the same system! Tell you what: why don’t you use a Docker container for each and every major GUI app?
- One Docker container for Firefox.
- One Docker container for LibreOffice.
- One Docker container for VLC.
- …and so on.
What’s better than full containerization, you moron?
Another problem, as Torvalds pointed out in 2019, is that while some major hardware vendors do sell Linux PCs – Dell, for example, with Ubuntu – none of them make it easy. There are also great specialist Linux PC vendors, such as System76, Germany’s TUXEDO Computers, and the UK-based Star Labs, but they tend to market to people who are already into Linux, not disgruntled Windows users. No, one big reason why Linux hasn’t taken off is that there are no major PC OEMs strongly backing it. To Torvalds, Chromebooks “are the path toward the desktop.”
That no major hardware vendors are committed to Linux is an open secret. But no, Chromebooks are the path to mental retardation!
Just look at Android, he argued. Linux won on smartphones because, while there are different Android front ends, under their interfaces, there’s a single, unified platform with a unified way to install programs. He’s right.
This is exactly the problem! Android is the worst possible model for everyone who values freedom! Rooting is strongly discouraged, bootloaders are locked and sometimes cannot be unlocked, banking apps refuse to work on rooted or unlocked systems, the immutable system cannot be touched on unrooted systems, and the apps are containerized and also subject to absurd restrictions!
With such a mindset, there’s no wonder that Linux will keep being a niche OS on the desktop (read “laptops”)!

Let me summarize the main reasons Linux failed and will keep failing on the desktop:
❶ Linus Torvalds, for his stupid architectural decisions (no stable ABI, hence no separately installable drivers and no backwards compatibility), for the huge mess that is the Linux kernel, for the decreasing quality in a kernel that moves like quicksand.
❷ Acer, ASUS, Dell, HP, Huawei, Lenovo, MSI, etc.
❸ Red Hat, because of the path taken by GNOME and because they literally don’t care about Linux on the desktop.
❹ Canonical, because they went the GNOME way.
❺ Steven J. Vaughan-Nichols.
Vox populi
I took a look at the commments section and I noticed a few things.
① A sarcastic comment with which I agree:
Oh, wait, I’ve got a better one: VMs! One VM for each and every program. Just the bare OS and the program. Want to browse the web? Start your FF/Chromium VM. Want to edit a file? Another VM and app.
② A retard:
This is something I’d LIKE to see on smartphones. That would provide the ultimate level of security. Apple already runs what is effectively a hypervisor on iOS – it manages memory protection and the iOS XNU/Mach kernel and the Secure Element’s seL4 OS run underneath it at a lower privilege level. Having apps run in their own VMs would be the next logical step.
③ A lucid guy:
Careful what you wish for. You may end up with secure system but only useful to the gatekeeper that forces you into managing all your data through their walled garden (and controlling what you can and can’t do). Akin to iOS. Even Android has become hostile to idea of users’ freedom with both Google and handset makers taking another cut of what’s left of the original concept.
④ An indication that not only the huge SUVs are obscene, but perverts are everywhere:
One VM for each and every program.
Sadly, I know people who seriously do this. Unsurprisingly, all three of them are software architects at their companies.
One has a desktop with 192GB of RAM, 32TB of NVME SSD disk space, four Ryzen processors, no less than six NICs. Everything is a Proxmox instance. Email is a 4GB Linux VM with Thunderbird. Browsing is a 32GB VM with Firefox, running an average of <cough> 300 </cough> tabs. And he considers that to be a perfectly normal desktop, not excessive at all.
I lost my faith in humanity. Or, rather, in humankind.
⑤ Some other guy noticed what happens when you install your very first Flatpak, and lots of libraries are pulled:
The Linux Mint Software Manager includes both the regular “System Package” version and the Flatpak version of the Gnome Calculator app.
The system package version is 7MB.
The Flatpak version is 1.1 GB to download, 3.6 GB of disk space required. For a 7MB calculator app.
No.
Fuck, no.
⑥ One more point:
In addition to the size issue, snap etc mean you are reliant on someone else keeping the snap etc up to date with the security fixes for all its dependancies.
Do all the snaps get updated when common libraries get security updates? No, they do not.
⑦ A funny one:
Remember when Windows 3.11 programs used to come bundled with their own versions of system files to ensure they’d run correctly?
Who knew it was the future of Linux.
The reply:
A lot of that was down to the fact that Microsoft decided not to implement shared library management in their Operating System.
This rapidly led to DLL hell, then Microsoft overlaid ActiveX and COM over the top of this DLL hell. Later Microsoft decided that this wasn’t hell enough and .NET came along where we have hundreds of copies of exactly the same DLL files (that’s what they are under it all) splattered all through the WinSXS disaster zone alongside so many file and directory links that the disk operating system is unable to work out how many files and directories there really are, let alone how much space is taken by the mess. In the end, in effect, it is very similar to a flatpack installation where the smallest of applications has oodles of duplicated component files in WinSXS and these just accumulate over time, they are never removed.
Well, these can be cleaned, but your average John Doe would be clueless as to how to do it.
⑧ Some realism:
Literally the only limiting factor for me is a small group of 3rd party software. Solidworks, Autodesk, Ableton, NativeInstruments.
If Valve’s gaming efforts can also handle these, I have zero use for windows at all.
But it’s a vicious circle here: why would anyone develop Linux versions of expensive commercial software if no major hardware manufacturer cares about Linux on the desktop?
Let us remember the reasons behind the death of Windows Phone, which, after all, did have a dedicated manufacturer! Apps, apps, apps!
⑨ From a long comment about having switched to Mint:
Of course, Mint is not Windows. I am well aware of this and the consequences that things are different. After all, if you switch from one car to another car from a different manufacturer, it will be similar, yes, but there will be differences. As an experienced “driver”, I can cope with that and as a “hobbyist mechanic”, I don’t mind getting my hands dirty and changing aspects of the car. I mean OS.
Can Mint be used as a desktop for the average Joe in the street? Honestly? I don’t see why not. What does Joe use? A browser. That’s probably about it these days. And Waterfox (thanks again, Liam!) and Firefox are more or less identical in operation under both Windows and Mint – which is how they should be.
But will Joe trust Mint (or any of the alternatives)? That’s the bigger problem, I don’t think he will. He knows Windows, so is happy with it. He doesn’t care about privacy – heck, I bet he keeps his wireless active on his phone in the shopping mall, subscribes to all sorts of marketing lists, etc. – and this AI thing? It does things for you so you don’t need to – result!
I tried Mint via USB on three completely different devices and in every single case, it found all the drivers and just worked. I was so surprised at this, despite what Liam et al. advocated. But Average Joe? He’s the man who has bought Ford (or GM/Vauxhall/Renault/VW/etc.) all his life. Windows just works, too, and he doesn’t need to do anything to get the latest Ford, it comes direct. And if he has to get a new PC, then so what? His existing one is only four years old, that’s about right, isn’t it? PCs are only designed to last for 3-4 years, like cars, so it’s an investment in the future.
Yes, the Linux desktop will succeed, but only for people like us who know. For Average Joe (and Josephine), stick with what you know. Take the easy option, even if it does end up costing you money and privacy.
Oh, so he bets on Mint and against Red Hat and Canonical’s bet on GNOME! Let’s not forget that there would be no Mint without Ubuntu and no Ubuntu without Debian. (So not even LMDE could survive without Debian.) Linux Mint is a fragile parasite, so to speak.
⑩ A testimony witnessing a completely opposite situation:
To offer an alternative opinion, I’ve just moved to Linux Mint. I’m not in IT, but consider myself at least a bit tech savvy (I’m an engineer). Moving to Mint was a s*$show. It didn’t recognise either my graphics card or the WiFi drivers. It took me about 4 hours to get to the point where I could just log into Mint.
Since then, there have been other issues (playing DVDs still doesn’t work half the time). But I’ve stuck with it. I am relatively happy, but mainly because the alternative of going back to Windows is worse. I still get unexplained crashes (I suspect Firefox is to blame, but haven’t tracked down the issue yet), and other issues.
I can stick with it (and hopefully work out the issues in time), but there would be no chance that anyone else in my family could or would deal with these issues. And Mint Cinnamon is supposed to be the simplest and most windows-like of all.
If Linux wants to become the replacement for Windows then it needs to make the install process and the first steps easier, more consistent, and frankly more Windows like, so people can get their feet under the table, without having to get into the minutiae of the Terminal window and the various specialities of Linux.
People can always learn that stuff afterwards, but the start is where you will gain and keep people…
That needs one of the three major names (Red Hat, Canonical, SUSE) to target the desktop and seriously work towards this goal! But this is never going to happen.
⑪ Again:
“If Linux wants to become the replacement for Windows…” Therein lies one problem. There is no single “Linux” to want to do this, instead there are over 100 distros all of which do things their own way and there’s minus-infinity chance of homologation.
⑫ And again:
You are basically alluding to the real problem: how does Linux get on to the high st. So that Joe Public visiting PC World get to see and thus buy a system with “Linux” (i.e. a specific distribution pre-installed).
Which was the problem Canonical was targeting with Ubuntu 20+ years back…
You see, people (yours truly included) are much more aware than SJVN of the real problems of “Linux on the desktop”!
⑬ Oh, my, some people seem to hate Google!
Re : Chromebooks “are the path toward the desktop.”
Sorry… but no they are not.
IMHO, Chromebooks are just vehicle for Google to slurp your data and you are paying them for the privilege. IMHO, they are so stripped back like the Netbook of old that they are a technological dead end.
All but one of my Linux systems run a GUI. The desktop is either XFCE or Cinnamon. They do the job that I want without fuss.
And another free man:
I have recently installed Debian XFCE on a Dell 3180 Chromebook. It wasn’t easy, but I figured it out, and it is now a vehicle for GW-BASIC and UCBLogo running on DOSBox. Going to introduce my granddaughter to programming. Total cost: $30 for the Chromebook and another $40 for a new battery.
This was an experiment on my part to determine whether there was life after Chrome for this hardware.
⑭ Brainwashing America:
One year I had to deal with about 200,000 Chromebooks for a school. No, that is not a typo. 200,000. (no it was not me by myself. Of course we had a team. I probably laid hands on just several thousand)
It’s an OK system, but it’s meant for a closed garden experience or personal light use.
But yes, by far the biggest problem is that it’s Big Google Brother.
⑮ Throatwarbler Mangrove, whom I have met in the past (see here and here):
I disagree almost entirely
Windows has not succeeded on technical merits, and it will not be defeated on technical merits. Windows succeeded because Microsoft pulled out the stops on getting it onto mainstream personal computers, collaborating (some might say colluding) with major PC manufacturers and, of course, Intel to ensure that Windows was ubiquitous. For Linux on the desktop, there’s no one who has the market clout to implement a similar strategy, which means that Windows will remain the default PC operating system for the foreseeable future.
Addressing the point about the enshittification of Windows, etc., very few people actually care, from my observation. Although Windows 11 certainly has more than its fair share of annoyances, it continues to be sufficiently functional for home and business users, and it has desirable features and capabilities that Linux lacks, such as OneDrive, Active Directory support, and the ability to run familiar software. It’s mainly the furry-toothed geek community, who are a tiny minority of the population, who find Windows 11’s behavior problematic enough to go through the hassle of installing a new OS.
For my part, I continue to use Windows as my primary desktop OS because I can turn off the bits I don’t like and because Linux has enough of its own foibles that it’s not better enough to be worth the pain of migrating.
Sorry, nerds, it’s not that your baby is ugly, it’s just not as pretty as you think.
I won’t comment on the above statements.
⑯ Santa Claus is here!
We need to stop with all this replacement software in GNU/Linux and just settle into stability. Wayland isn’t ready, and will never be ready. X11 has been here forever and all we need is just some acceptance to say enough is enough, use what is here, and stop messing around with fadware and trendware. We don’t need X11 replaced.
Desktop: Mate or Xfce
Graphics: X11
Audio: Pipewire or PulseAudio
Input: Libinput
Networking: NetworkManager
Sessions: sddm
What is so hard to just follow a simple roadmap for making a cohesive system for FreeBSD or GNU/Linux we can drop Firefox, Chrome, Steam, Gimp, Libreoffice, OBS, Mplayer, etc. software onto and finally have a Year Of The UNIX Desktop?
What is so hard is we have to many wishy washy developers who think stability isn’t good enough and too many puppets who like to go along with anything and everything because a trend or fad is cooler than something founded.
Enough is enough. Stabilize the UNIX userland finally…
Otherwise, people are going to going back to Windows time and time again. Nobody wants broken software, half working software, or software that is just never going to last and be replaced every year.
The YOTUD is said to be every year, then it gets pushed back another year, and then never the next year.
You want to beat Microsoft, then follow a formula, and stop being wishy-washy.
YES, YES, YES!
⑰ One of my points, one more time:
What Linux needs to be pre-installed on PCs
People keep pointing out that installing Linux is easier than it’s ever been, so users won’t have any problem installing it.
They’re right that it’s easier than installing Windows 95/98/2000/XP or OS/2 was, back in the day. They’re wrong that it means users won’t have a problem with it.
The problem isn’t how difficult Linux installations are, it’s that it requires installation at all.
The modal Windows user today didn’t install Windows, it came pre-installed on his machine. Oh, he configured it when it first booted up, certainly. He gave his name, his location, his (now mandatory) Outlook account, and the like, but he didn’t partition and format the hard disk/SSD, he doesn’t know how much memory he has, and he certainly didn’t look at, let alone set, jumper settings on the motherboard like “the good old days”.
Linux in 2025 is unquestionably easier to install than Windows XP was in 2001. But it’s not 2001 any more. Comparing ease of installation from media isn’t the point, we’re comparing installing Linux from media against pre-installed machines.
Sure, there’s System76 and Tuxedo, but if Linux is really going to compete on the desktop, it’s going to need vendors who offer PCs with Linux preinstalled, and with support.
Microsoft is continually shooting itself in the foot and driving many users away, but even the thought of installing an operating system from scratch nowadays is intimidating to many people. Sure, many will give it a shot, and many of them are surprised at how easy it was. But many more have no more interest in installing a computer operating system than they have of doing their own oil change on their car.
Indeed. See, you old Steven fart, what reality looks like?
Reinforcement:
Finally.
What we need is Acer, or Dell, or Lenovo (or better, all three) offering a budget range, all with the same distro preinstalled and a bunch of popular apps.
What that distro would be, I have no idea (I have my own preferences but that’s not important to my point). But I guess that’s part of the problem. Until a single distro is targetted by multiple big PC vendors as being suitable for preinstalling and marketing, this whole discussion is moot (or possibly moo if you’re a Friends fan).
BINGO! Oh, no, but what if that distro is GNOME-based?
⑱ Another long story, which I copy in its entirety:
Snatching Defeat from the Jaws of Victory
My hair is fully gray, and I do recall my introduction to shell on the Vax, and Unix on Sun SpacStations, and I do hold out the best hopes to see the “Linux Desktop” becoming more prevalent on all computers, but history teaches us that the developers always find a way to snatch defeat from the jaws of victory. 2007/8 were particularly painful as both the major desktops abandoned well developed code-bases and well liked desktop environments to follow the lead of the new group of “kids with crayons” chasing desktop “widgets” and “effects” that added nothing to the ability to use a computer, but which “threw the baby out with the bathwater” and set the Linux Desktop back at decade or more.
KDE 4.0.4a (alpha) was pushed by openSUSE 11.0 as the default desktop and blackscreened continually for years. The desktop was a hodge-podge of apps that had been ported to KDE4 and those that languished well into Plasma before finally being ported (and broken, and taking years before, e.g. konqueror –profile filemanagement could open up looking the same way it did when it was shut down). Gnome fared no better with gtk+3 breaking backwards compatibility with every new point-release for a decade or more. Icon and widget factories, etc.
This effectively relegated the Linux Desktop to hobbyists, of which I’ve always been one. If it takes hours to collect information, author bug reports, and then spend weeks if not months of back and forth, that part of giving back to the community which we gladly do. That is also the complete deal-killer for the Linux Desktop in a business setting. There is no way any company can retrain workers on a monthly basis to deal with issues created by fights between toolkit makers or sloppy porting of apps from one toolkit to the next.
Which brings us to the present, after a long road to finally get Plasma frameworks and gtk4 settled. And what to the good stewards of the Linux Desktop do to ensure another decade of defeat snatched from the jaws of victory? Oh, they decide to tear down X11 and follow the kids with crayons pushing Wayland. They make it tribal, not based on merit or features or stability or any other metric you would look at in choosing how your computer works. That will certainly give business the warm fuzzy it needs that things have stabilized to the point of the Linux Desktop being a reliable alternative.
We may as well just go build out a bizillion data centers and burn the planet down… Why we can’t learn from history is bewildering and frustrating. But I’m a hobbyist, my Linux Desktop will work fine, but I’m not sure I have another decade to give. At some point sanity must prevail.
Well, my only chance is to remain a hobbyist, then?
⑲ A bit of nice bashing:
Today, we should all be using Flatpaks, Snaps, and AppImages to install programs instead of worrying about library incompatibilities and the like.
We absolutely should not. I avoid Snaps like the absolute plague because every single Snap install is broken in some manner. Have a non-standard “/home” layout due to reasons, no Snaps for you! Want to clearly see disc usage? No Snaps for you! Want to be sure a vulnerable library has been replaced? No Snaps for you! And on and on and on. They are horrific, and I am going to assume the same horror extends into Flatpacks and AppImages.
Basic DEBs are the way to go.
⑳ The last one is depressing:
It’s dead, Jim
I love Unix and Linux (as a dev)… I eat my own dog food as do many here. We (Linux Users) are and always will be in the 0.1%.
If and when the 99.9% of other users get sick of Windows they will buy a Mac or use a web-based app.
All this talk of the Linux desktop becoming a major player is crazy talk.
We can all dream nostalgic about how we got a PC in 1977 and ran CP/M on it but the same dynamic existed then as it does today… only edge cases like us had the drive (no pun intended) to do geeky stuff.
Even if every Windows app anyone ever uses becomes web based/thin client (i.e. the OS is irrelevant), is Suse or Canonical etc going to invest the money to market Linux, somehow break the Wintel alliance or outspend Apple on marketing to get Linux in front of the general public… NO.
Of course the irony is iOS and Android are Unix/Linux but the 99.9% don’t know that or care. They just buy a device, turn it on, go to the app store, install an app and move on. Nor do they know/care that Linux runs the backend of all the Cloud services they use (outside of Azure).
But I am happy with this because the lack of large commercial ownership/influence means Linux won’t get hijacked.
Leave Linux for its rightful owners… us geeks.
Oh, but where are the defenders of SJVN’s immutable system?

Leave a Reply