I was a bit disappointed with desktop Linux during the late 90s and early 00s because many applications within the typical ecosystem weren’t great or mature yet (I still shudder when thinking about the amount of time I spent trying to get audio to work nicely across OSS, alsa, artsd, esound, … this was before pulseaudio or pipewire, obviously), and then I temporarily switched back to Windows as a primary OS just because of gaming compatibility. But since the mid-10s or so, Linux grew to become very capable for at least casual gaming as well. So there was no need for Windows anymore and I just wiped it off of every device, leaving Linux as the only OS. This was also during a time where Windows 8, 8.1, 10 and later 11 showed up and this really showed to me that Windows was doomed to die a long-winded death. The last decent version of Windows was 7.

It was funny (and untrue) when desktop Linux was constantly at ~1% market share. But now with desktop Linux slowly but constantly rising, it’s actually true!

Same here. Built a new PC like 6 months ago when it was all barely starting (good timing) and my notebook and phone are also fairly recent, like 1.5 years old. Tablet as well fairly recent, 2y old, but that’s really only used for home media consumption so it doesn’t even matter in that case. So I should be good for the next 5+ years or so. Hopefully the AI bubble has burst until then.

Arch on Desktops&Notebooks, Debian on Servers.

KDE Plasma is one of the easiest desktops to use for a user coming from Windows, that’s for sure.

Gnome also has its place. I’ve used it for a few years (up to the Plasma 6 release, where I fully switched to Plasma on Wayland) and I like Gnome in general, its UI is a “modern” mix of the MacOS desktop and a mobile phone UI, whereas Plasma goes the traditional (one might also say boring) Windows desktop look&feel route (although you can also reconfigure Plasma heavily to look and feel however you want, but by default it’s very Windows-like).

Main difference, if you set aside the UI, is that KDE Plasma offers a ton of settings and features (available easily via the GUI), while Gnome intentionally doesn’t - that means to change some things you need to put in more effort in Gnome (use extensions or change settings via dconf editor, or custom CSS files. and so on).

Then there’s the factor of stability - Gnome has always been rock-solid stable for me, while KDE Plasma since 6.x release has been very good but still not rock-solid. I’ve had maybe around 5-7 plasmashell crashes so far during the whole time of using Plasma. Plasma has a recovery feature for this integrated now, so it isn’t as bad, but it’s still not rock-solid. The small price you pay for all the great and many features Plasma has is that it’s (still) slightly less stable, but due to it being able to self-recover most users probably don’t need to worry about it. I’ve had it on Windows as well - whenever the explorer shell crashes, it instantly restarts. You might have noticed this when suddenly the taskbar and everything is shortly gone but then re-appears. It’s kind of the same in Plasma, just that you also get a notification that plasmashell has crashed. I’ve never had any of this happen in several years of Gnome usage.

However, the Gnome devs are doing some weird decisions… they have very strong opinions about things and tend to not agree to things which they should agree to in order to make interoperability with other software easier. There have been a lot of unnecessary discussions about how to handle client-side window decorations, desktop portals, how to implement cerrtain Wayland extensions and things like that and often there are 2 factions at the end: the way Gnome wants to handle it vs. the way everyone else wants to handle it. I don’t like this attitude or selfishness at all from the Gnome project.

But in terms of software and UI quality, I think Gnome definitely has its place. Whether you like its UI or not, is something that every user will see differently. I’d assume that users coming straight from Windows and who maybe also want to continue to use the Windows UI paradigm on Linux will tend to find Gnome a bit alien. That’s understandable. Gnome has its own UI paradigm that doesn’t try to cater to Windows users. So you need to adapt to it, not the other way around. I also never really care that much about what desktop environment I’m using, because I’m doing so much with Emacs, the terminal and the browser, it almost doesn’t matter much what’s “around” that. So take my opinion with a grain of salt - I can probably use any UI paradigm and be fine with it. So maybe I’m not the best person to judge how well a UI is done. But I still wanted to give credit and criticisms where they’re due.

Yes. It’s already grown from ~1% to ~6% within the last couple of years. There are several major external factors at play: Valve helping to push gaming on Linux, the continued and increasingly big enshittification of Windows, and the current deranged US regime (resulting in less trust and less users of US-company-produced proprietary operating systems). Remember that Linux or the open source BSD variants are the only (usable/practical) operating systems you can use if you want to achieve digital sovereignty. Plus, it’s also getting even better over time by itself of course (that’s the internal factor).

There was or is a specific bug with earlier Ryzen CPU generations that causes system freezes when the CPU enters the powersaving c-state #6. If that applies to your machine, try the kernel parameter max_cstate=5 for a while. I had one PC build where this 100% resolved the system freezing after several minutes. Note though that this is ~6 year old info. Issue might have been resolved in the meantime. But worth a try probably if you have an older Ryzen CPU.

“MS Office with its integrated spyware and other anti-features doesn’t meet basic security and data protection requirements so I use LibreOffice”

I still like Arch the most. Since ~22 years now.

Mint or Fedora(KDE) are great choices. Kubuntu, PopOS or OpenSuSE might also be suitable for beginners. Stay away from Arch-based distributions until you are at least a bit more experienced.

Intel (anything) works without problems on Linux (in fact, Intel is among the most Linux-supportive companies out there and most or all of their drivers are open source and part of the kernel, as it should be in the Linux world).

Nvidia GPUs used to be problematic in the past, it’s better now, still not as great as AMD GPUs are on Linux (they’re literally plug and play these days) but I think when going with the distro mentioned above it’s going to be just as easy Just make sure to enable support for NVidia drivers or “enable 3rd party drivers/repositories” (you’ll be asked during setup) so that those distributions will also install those slightly non-standard Nvidia packages which they might not do otherwise for “purity” reasons.

Linux Mint is often recommended to users coming from Windows, so… Kubuntu, Pop!OS and OpenSuSE are maybe also decent for that use case.

Manjaro isn’t recommended. They made lots of weird decisions and mistakes in the past, maybe still do. Wouldn’t trust them. Endeavour or Cachy are the current recommendations for “easy Arch”. If you’re able to install and maintain vanilla Arch, I’d recommend Arch though. Cut the middleman.

I’m only a little bit familar with the TUI browsers. I’m also not sure about gemini and gopher support so you have to look that up on each project page, but I can give some general directions:

• Lynx is basically the oldest TUI browser, so probably not the best and no modern choice, but still maintained I think
• ELinks started as a fork of Links (and Links started as an alternative to Lynx, so both ELinks and Links are newer than Lynx). It has a lot of features and is actively maintained, so it’s decent I think. Probably better than Links (and Links is probably better than Lynx)
• Links2: no idea, just know that it exists. If it’s still actively maintained I would suggest comparing it to ELinks because they’re both probably similar (both related to but newer than Links))
• W3m is the one I’d recommend, it’s powerful and can be integrated more easily into other applications. For the classic TUI browsers, it probably comes down to the choice between w3m and elinks
• There’s also a modern project called Carbonyl which is essentially Chromium running in a terminal, so this one might be “better” than all of the above in terms of features and modern website compatibility. But again, it depends on what you want out of a TUI browser - if you only need something basic this is probably overkill. But I didn’t try it out.

It’s for window management related hotkeys. Obviously. All about windows. With a lowercase “w”.

Tips for coping:

• Call it the Super key (actually the correct label I think)
• Bind window management related hotkeys to it

Technically, nothing you use in tech is ever really “simple”, there’s tons of complexity hidden from the common user. And whenever parts of that complexity fail or don’t work like the user expects it to, then the superficially simple stuff becomes hard.

Docker and containers are a fairly advanced topic. Don’t think that it’s easy getting into this stuff. Everyone has to learn quite a bit in advance to utilize that.

To play games, you went into the wrong direction when fiddling with wine directly, or even just indirectly by using bottles You COULD do that, but you’ve literally chosen the hardest path to do so. You should use something like HeroicGamesLauncher, Lutris or Steam in order to manage your games, install and launch them fairly easily. These will take care of all the complex stuff behind the scenes for you.

When I was new to the Linux desktop world (late 90s to 200x) I tried lots of different distros and (X11) window managers and tools and whatnot. Changed themes a lot. And so on. And I think there’s value in all that, because it expands your horizon of what’s possible on the desktop, how different UI/UX paradigms work out in practice for you, and you learn how to use different environments.

On the other hand, there’s also value in having a consistent, well-integrated desktop environment. It can mean less “pain points” in various circumstances, and it’s also efficient when multiple programs share the same libraries or code base instead of having separate tools all around.

In the end, it comes down to what works best for you. But this might also change over time. For example I’m really considering switching to Cosmic once it’s mature. I’m also considering taking a look at Niri because it seems well thought-out. But currently I feel cozy using Plasma at home and Gnome at work because Plasma is currently the least-annoying and at work I still use Gnome because it’s been historically more stable than Plasma for me. I’ve tweaked Plasma’s hotkeys so they work more like Gnome’s and since I also need to use a couple of Windows-based systems at work I’ve also configured common Windows shortcuts like Super+L, Super+E, Super+R so that they all behave the same everywhere.

Oh, and my distro is Arch everywhere because I’ve used it for ages now and I like its technical simplicity, stability and modularity. It’s the one distro that gets in my way the least.

I think one should learn enough to be flexible and be able to use everything, while also not being too narrow-minded and just focus on one solution too much. What works best for you now might not be the best choice for you in a couple of years.

While this is stupid, it’s also minor and doesn’t really deserve such clickbaity titles. You’ll only be locked out for 24 hours, which is not long, and you’d have to switch Proton versions 5 times within 24 hours for this to happen, which should be very rare to begin with.

I play exclusively on Linux/Proton for years now and at the MOST I’ve switched Proton versions like 3 times within 24 hours, 99% of the time I switch them never (because 99% of the time it just works) or just once (e.g. switching from regular Proton to Experimental or GE-Proton, which might help wth very new games). That means I would probably never be affected by this.

I"m not trying to defend this, this limitation should not exist, but I’m trying to position this correctly because the titles are misleading clickbait, they suggest you’ll be locked out permanently.

• To learn Arch, install it from scratch (without archinstall), it’ll force you to read the Wiki and learn a lot of necessary commands in the process. After the installation, just keep using it. Using a Linux distro full-time as the only installed OS is the best way to keep at it and truly learn it over time. There’s no magic bullet here. Just keep using it and solving problems or issues as you go, learning more and more stuff as you go. If you need other OSses as well, run those in a VM. I don’t recommend dual-boot setups.

• Don’t blindly copy/paste commands you don’t understand. Always try to understand them first. Some commands can be very disruptive or even destroy your configuration. If you don’t understand it or are able to adapt it so that it fits to your particular configuration or system, you can EASILY damage a configuration, or even make your system unusable. Also, some people like trolling other people and deliberately share harmful commands. Generally, test potentially destructive commands or complex commandlines before actually running them.

• Document major config changes that you do. This is useful because you’ll be able to undo certain changes or even replicate your current system configuration fast when you change distros or have to reinstall in the future. For example my current Arch-based setup is fully documented in form of an almost-directly executable shell script. It does require some interactions but very little. If I ever have to reinstall this system, or upgrade my hardware, it can be done insanely fast and it’ll have the exact same configuration. This goes from basic partitioning and encryption all the way up to dotfiles and individual program configurations.

• Don’t feel the need to learn hard/advanced tools like Vim or Emacs unless you really think you’re getting an advantage from that and aren’t hesitant to put in the time and effort to learn them. Most people don’t need to use them. They’re amazing tools but you need to be prepared to lose quite a lot of time to learn them before you can become productive with them, and this might not be a tradeoff that’s useful for every single user. You can also get away with much simpler tools, like nano (as a console-based editor) or whatever programmer’s text editor you want.
• Similarly, whether a pure WM or compositor plus assorted tools compared to a full desktop environment is worth it for you or not, is up to you. There’s no wrong or right answer here. I’ve tried out pretty much everything and these days use KDE Plasma because I like the consistency and integrations and dislike having different, inconsistent stand-alone tools for panels, menus, notifications, wallpaper, file manager and so on. But again, there’s no wrong or right answer here. Just what makes more sense for you. It’s worth learning how to be able to configure and use a minimalistic setup, for sure. So trying it out doesn’t hurt and increases your knowledge overall. In general, in the Linux world it’s good to always know enough to not be screwed once some component suddenly doesn’t work anymore. For example, a competent Linux user should be able to deal with (temporarily) not having a GUI and fixing his system via commandline.

• A minimalistic, DIY distro like Arch can be amazing to learn everything, if you want to do that at least. If you just want a working desktop system with as little effort as possible, then don’t do that. But if you intend to learn every detail, then a distro like Arch is better suited for that goal than a “bloated”, fully pre-configured distro like Mint or Ubuntu is. Because Arch is much simpler on a technical level than those are. It’s much easier to understand e.g. the relatively simple package building process on Arch than it is on Debian/Ubuntu-based distros. But this “simple” explicitly refers to technical simplicity or minimalism. Most users expect something else when they hear “something is simple”. Arch is not simple as in beginner-friendly, but it is simple in terms of technical complexity, which is why many advanced users and tinkerers like it because it doesn’t stand in their way. It also means though that you HAVE to learn many things, e.g. how to configure a firewall, because it doesn’t come with any preinstalled by default. With Arch, the admin is supposed to know about everything and configure every component himself, at least on a somewhat basic level.

• If you want to go to even more details, you could also try out a source-based distro like Gentoo or Crux, which can also be a great learning experience, but it’s even more details regarding compilations of each package, dependencies, compile-time options, etc. you have to deal with than with a minimalistic binary-package-based distro like Arch, so whether that’s useful for you or not is up to you of course.
• While we’re at it: LFS (Linux from Scratch) is not a distro per se, it’s a guide on how to build your own distribution from scratch. It’s VERY time intensive and not recommended unless you truly want to learn how to build a complete distribution from scratch, or maybe start your own distro some day which isn’t based on another existing distro.

I’ll do a (simplified) Windows analogy, if you’re already familiar with Windows.

Microsoft Windows is closed-source/proprietary, which means only Microsoft has the source code for it, and only Microsoft is legally allowed to create or distribute copies of Windows. “Windows 11” for example is a “distribution” of Windows containing the “Windows NT kernel” (core of the OS) alongside other important software to make the OS usable, like a boot loader, service layer, graphical interface, desktop environment, and lots of included “system” applications like a file explorer, a web browser, apps to adjust settings, apps to display menus and task bars, and so on.

“Linux” by itself is just the kernel, the core of the OS. Which is by itself not a “usable” operating system yet, just like holding a CPU in your hand doesn’t allow you to use it yet. More components are needed for that. Since Linux is open source and under a permissive license, anyone (even you) can go ahead and create an operating system made with the Linux kernel. If you do that, this is called a distribution or “distro” of Linux. Since there’s not just one company allowed to do that, many distributions exist. They all made their own operating system on top of the Linux kernel. Even though hundreds of distros exist, only a handful of them are actually popular, stable, secure and recommended for general use. They all use similar, but sometimes different software to include in the distribution. Like the Linux kernel, most of that software is open source so it can also be modified or extended.

Since “Linux distribution” is rather long to write, people often just write “Linux” but mean the whole distribution, not just the kernel. These are just common inaccuracies in communication, but what the person meant should be obvious from the context.

Common and recommendable Linux distributions (= full, usable operating systems) include: Linux Mint, Ubuntu, Fedora, OpenSuSE, Arch, Debian. These are full operating systems and they all include the Linux kernel at their core. Of course, the similarities go further than that. Most distros are similar enough that if you’ve learned one, you can also use any other with little additional things to learn. However, some distros are deliberately a bit more different or tailored to more specific users or use-cases, for example Arch targets more experienced Linux users because it’s a very minimalistic distro, it expects the user to know which packages he wants to install. It pre-installs almost nothing. You can think of this like “Windows Server Core” where it just boots into a minimalistic terminal by default, no usable GUI yet, but you can of course install the desktop environment and everything if you need it and make a full-featured desktop out of it. The distro just doesn’t want to preinstall anything which you later might not like, which is why it gives you the choice, but that makes it a minimalistic distro and it’s harder for beginners to use that way. Other distros like Mint are much more similar to the client editions of MS Windows in that they preinstall everything the user needs for a desktop OS and more, so that the user can boot into and use the desktop as quickly and easily as possible.

And then there are even more special-purpose distributions like Kali Linux which includes things like penetration testing tools (i.e. “hacker tools”), which makes it a distribution for IT security people, so they can boot into it and have access to most needed tools right away without installing much else (also good on a bootable USB stick). But usually, in general threads like this one, people don’t talk about specific-use distros, but about generalist distros which you can install and use as a regular desktop OS.

Desktop environments also exist on Windows but there’s basically only one, made by Microsoft. In the Linux world there are several to choose from. The most common ones are: KDE Plasma, Gnome, Cinnamon, XFCE. These desktop environments contain window managers or compositors, task bars or panels, menus, various tools like file managers, process viewers and text editors, and various background programs. This is all needed for the user to have what is commonly known as “a desktop environment”, because if you didn’t have one, you’d be basically staring at a screen containing at most a cursor and a wallpaper, with no way for you to interact with anything. Of course, these can look and feel different from each other (just like Windows looks and feels different than MacOS), and they have different features and strengths and weaknesses, but their goal is always the same. And as usual in the open source world, there’s not just one project but multiple, and out of those multiple a couple are popular, viable and stable enough so that they are usually included in most Linux distributions. Which is why most distros also give the user the choice to have a specific variant of the distribution with a specific desktop preinstalled. For example, Ubuntu also has Kubuntu (= Ubuntu with preinstalled KDE Plasma) or Xubuntu (= Ubuntu with preinstalled XFCE). These can have various names but in the end it’s just the base distribution (“Ubuntu”) with a different preinstalled “face” so to say (and you can change those faces or desktops from within the same distro, of course). Most other things are exactly the same between those distribution variants.

As a new user, you don’t need to learn about everything. Just pick an easy to use generalist desktop distro like Linux Mint and use the default desktop environment or variant which they provide or recommend by default. You can start experimenting with more choices later on if you want, but you also don’t need to. If you have something you’re comfortable using, then you can just stick with that.