It peaked at 4.05% in March. The last 2 months it went just below 4% as the Unknown category increased. For June the reverse happened, so 4.04% seems to be the real current share of Linux on Desktop as desktop clients were read properly/werent spoofed.
You must log in or # to comment.
Man i hope linux becomes more popular
I see multiple posts on reddit everyday asking for advice for migrating to linux. I think linux userbase is increasing a lot since Window’s questionable recall announcement.
And the valve steamdeck But some people install windows on it which defeats it’s linux purpose
I feel like both “people who install windows on the steam deck” and “people asking for advice for migrating to linux on reddit” are just vocal minorities which you encounter on the internet but don’t really influence the Statcounter’s results in a meaningful way. Generally (from my view) it’s the kids who got a steamdeck for xmass and the coders who use ubuntu for work influencing the numbers.
Nevertheless, Valve’s work with proton has pretty much crushed the argument that Windows is needed for games. That use to be a major sticking point, preventing people from leaving Windows - but now not so much.
If you play games that requires anti cheat It’s gonna be harder to switch
Also vr support
Halflife alyx?
There’s some other good ones too, but fair point
A lot of mods are also windows locked too.
Nexus mods is working on a Linux client which is really exciting! Also Steam Workshop works on Linux. This covers a ton of use cases.
Not saying everything is 100% perfection, but it’s easier than ever to switch, and only getting easier.
I imagine “Windows locked mods” would probably also benefit from just disconnecting the internet and keeping it set up just the way one likes it, since MS is gonna drop Win10 soon.
That’s the case with WMR VR headsets. Sadly don’t see those getting cracked to work on Linux any time soon. :(
I just saw the news for Nexus mods like 20 minutes after I posted that. Hopefully it can be integrated well soon.
But yes, over time, things will continue to get better. Even Nvidia finally started working on open drivers for their GPUs.
Almost all anti-cheats work on linux or offer linux integration or builds. It’s the scummy unethical publishers who run the typical games that uses anti-cheat who refuse to pay engineers to make the minimum effort to support linux. Because it would undermine some of their bullshit claims used to manipulate their players. Fortunately for some people like myself, the typical game that requires anti-cheat is not a game they would want to play anyways.
Sorry, I stopped playing factorio on my work Linux computer. I will play next month to get us back up.
Don’t worry, I’ve been playing plenty of Linux factorio for the last few weeks to make up for it.
Thank you for your service
Question: Why is BSD so low? (And why/what is unknown?)
Less hardware support than Linux without enough substantially better about it to make it anyone’s clear preference. Which isn’t to say it doesn’t have advantages over Linux. Just that the average BSD user is going to be able to easily swallow their pride and run Linux if things went wrong with a BSD install (trust me, I’ve literally done this, these people do exist)
Because it’s not overly popular as a desktop os, you are far more likely to see it in certain appliances and server applications etc, none of which will show up in a pagevisits based statistic.
There really isn’t a compelling argument for BSD other than interest and hobby. It doesn’t have the industrial use case Linux has.
The BSDs got screwed over by a lawsuit in the 90s that made a lot of people hesitant to use them (coincidentally leading to the creation of the Linux kernel). Inertia carried it from there and Linux ended up getting more hardware and software support, which is the primary reason that people pick Linux over the BSDs now.
Statcounter numbers are to be taker with boulders of salt. you can look at many metrics and especially when you filter by country. you will see a lot of erratic unexplained changes. jumping down and then a few months lather up by sometimes up to double digits.
I switched to librewolf so I will now show up as windows
Same, there must be a percent or so of Windows actually being Linux instead.
I unspoofed Librewolf back to Firefox + Linux. That way I’m not contributing to Chrome and Windows market share and perceived dominance. Plus the more people don’t spoof, the less of a need there will be to actually spoof at all as the Linux market share increases.
The problem is that websites sometimes change there behavior.
After seeing Garuda Linux set my user agent to Windows, I set my Windows install user agent to Linux.
Seeing Twitch.tv login break after changing my user agent was hilarious
What did it say? Why does it break?
Yeah me too, safety in numbers. Maybe if Linux desktop gets bigger than Windows they’ll swap it around 👨💻
Can we commit to only posting about round number percent changes?
What’s interesting is that multiple trackers are now saying it’s above 4 percent. Last time something was posted, people questioned the data and where they got their data (which they should). Now there’s multiple sites showing a real increase.
No, we will not stop
4.040000000000000000000001!
We could contact that person as a community
Based on a world population of 8 billion, that would be roughly 0.00000000000000125% of a person. And is also not even representable in a 64 bit float so I had to do this math in my head (Calculator just says 0)
No, we have to post these when it’s the year of the Linux desktop
Fwiw, my blog’s statistics say Linux is around 10% and I know a lot of browsers identify themselves as running on Windows when they’re not, so I wonder how it’s measured.
I would wager thats your audiences bias showing up. If you did that measurement in lemmy users, you would get likely 90% Linux users.
lol no you wouldn’t. I’d be shocked if Lemmy users were greater than 20% Linux.
You can tell from the neofetch screenshots of their VMs showing weird configurations like 3 cores and 6 GB RAM with 15 minutes uptime.
Weird configs aren’t always a tell, my daily driver is a desktop with a tigerlake mobile engineering sample cpu
Are you implying that the host OS in those situations is not Linux?
What’s is the main topic of your blog?
Travel, nothing tech related
We should do a poll, who uses Linux?
Seems like you will get a biased sample here…
Why?
(Joking)
I think the venn diagram overlap of linux users, and users of adblockers or noscript users that block such tracking is quite large. Must impact the statistic significantly.
I use ad blocker but do not spoof user agent. If uBlpck does not block statcounter, I should be counted.
why do we negate chromeos we are 10% at least so there we really are
For starters doesn’t it only run apps from the Google play store?
and what kernel would you say it and android runs? relevant?
This is the format collision discussion that has no solution so far. A tablet that runs windows is counted as Windows. A laptop that runs android does not. Neither does an android cellphone. It all boils down to web browser user agent fuckery. This is why steam’s numbers are more reliable than other sources, they’re direct hardware surveys.
But the point is that a steam deck is not (but in a way it is basically just) a PC. There are tablets than run desktop interfaces and now there are laptops that can be used as tablet. Eventually the artificial mobile vs. PC/desktop/laptop schism will stop making sense.
404: Not found
In all seriousness, I think government bodies switching to Linux (UK’s, China’s, some Indian states’) attributes the most to this.
Even if that’s the case, it’s telling of Linux’ maturity.
Oh absolutely!
No I think it’s the Steam Deck. It’s like half of all actively used Linux machines.
Source? Last I checked, the Steam Deck was very much in the minority even when narrowed down to just desktop Linux.
I confused it with Steam statistics sorry
Source is the Steam hardware survey set to show Linux data only. He forgot to mention the statement is only true for Steam Gamers, not for all of Linux desktops outside of Steam.
I believe the Steam Deck would be a significant portion of the Linux desktops, but Steam’s survey might be a biased source, still
What does significant portion mean to you?
Half of the Linux machines on Steam, not the entirety of Linux.
Yes sorry you’re right
A very important distinction
But that’s not really a Desktop is it? If we’d count mobile device we’d also have to include Android and then the situation would look completely different.
It is. I use it as such regularly. Keyboard+mouse+screen = browsing firefox as usual. Works quite well. Libreoffice, okular, signal desktop… I’ve used worse computers in recent years, steamdeck desktop experience is better than many 4 to 5 year old cheap laptops with win10 or win11.
Steam Deck is a desktop. It is exactly the same PC hardware and software you are using on your desktop PC. It runs the same games and is software compatible. Steam Deck is a desktop PC.
Android has a different hardware (not x86 compatible), is focused on phones, its eco system of software is not compatible with PC and in reverse does not run your PC software. Android based smartphones are not a PC.
So your definition for “desktop” is if it’s an x86 compatible architecture? Seems pretty random to me. Btw, there are x86 android device. IMO a desktop is something on the top of a desk to do typical “office work”. PCs, Macs, Laptops, etc. but calling a SteamDeck game console “Desktop” is pretty dishonest I think.
PC in the first place means x86 architecture. Desktop is the operating system, here a Linux desktop or Windows OS. Nothing random here. Steam Deck is a Desktop PC with form factor of a handheld. Just like any other Laptop is a Desktop PC in a notebook form factor. It runs, plays and uses the exact same hard and software as a Desktop PC. Calling Steam Deck PC not to be a Desktop is pretty dishonest. Also saying desktop PC must be on a desk for typical office work is random too.
Also this is all about stats. Don’t forget, this is not your opinion or my opinion. I am speaking from the stats perspective, as the Firefox browser (or any other browser) is just a Desktop PC from the perspective of the stats.
If you say I am pretty random and being dishonest, then first, you are discounting others opinions, secondly did not even understand the purpose of my reply. Maybe you should watch out your language next time before responding if you want further discussion. I hope my reply here made it crystal clear.
The names are pretty clear and are about form factor. Desktop is something on top of a desk. Laptop is something on top of your lap. Hand-held is something you hold in your hand.
The steam deck is a hand-held game console - doesn’t matter what OS is it uses. It’s true that most stat tracking sites count it as “desktop” but not because it’s a desktop computer but because the user agent looks similar to desktop user agents.
If I install Android on a tower PC it doesn’t randomly become a smartphone even though all browser trackers would register it as a smartphone.
And Valve using a “typical desktop OS” on their handheld console doesn’t magically turn it into a desktop PC.
The example with Android makes no sense as a argument in our discussion.
The steam deck is a hand-held game console - doesn’t matter what OS is it uses
Wrong. It does matter. Plus, it matters what software and what configuration it uses the moment it gets counted in the stats.
The discussion is not about what you try to tell. You can say what you want, it does not matter that the Steam Deck uses hardware that is PC and software that is a hybrid of Desktop and Gaming operating system. And when going to the web with a web browser, then the Steam Deck is in Desktop mode, and therefore seen as a Desktop PC. Not only act it like that, the functionality is a Desktop PC.
The Steam Deck has two modes if you forgot that, and we are talking about the Desktop PC mode. Which is what the stats are all about.
But how many use it for browsing, which I imagine this data is from?
We don’t include Android here. What I meant is that the Steam Deck does count in that statistics.
Steam deck has a full fat kde desktop on the stock os
I could install a full fat kde on the entertainment system of a car - still wouldn’t call it a desktop PC.
You can plug in a keyboard, mouse, and monitor
Connect the Steam Deck to a compatible dock and you can quite easily use it as a desktop. At the end of the day, it’s still an x64 based PC that’s just handheld.
I’m not sure that’s really a good argument. I can connect an android smartphone to a monitor, keyboard and mouse and call it Desktop. It’s also just an arm64 or x64 based PC just handheld.
A Desktop PC IMHO is a device that is used for everyday “office” work and neither android smartphones nor steamdecks are that - but laptops for example are (IMHO)
no, the statistics are based on browser agents, very few steam deck users browse the Internet on their devices. it’s also only half the Linux devices on steam, not of all Linux desktops
In Steam maybe. But this is StatCounter which is website visits. I doubt many Deck users are browsing the web.
For some reason I think a lot of them (probably even more than half) have tried browsing the web or at least using the desktop mode at least once.
Oh that is a good point, why didn’t I think of that!
I’m fairly sure it’s deficiencies in StatCounter’s measurement that’s accounting for it. Statistical noise, basically.
It’s probably even higher than that. These stats are mostly based on website visits I believe. And many Linux users are also privacy-minded and might spoof their OS in the browser. I bet a large portion of the Unknown is actually Linux too.
It’s hard to tell, as there are so many things that influence it. A huge factor is selection bias, as only a small number of website embed StatCounter, and that’s very likely to not be a representative sample. I’d bet that the influence of that is magnitudes larger than of user agent spoofing.
And a portion of the “Windows” as well. Hiding in plain sight and all.
The youtuber matt from thelinuxcast sucks.
I am regular user, i don’t code for living and my job is not tech related. I wanted to try linux and many of you guys supported and now I’m using Linux since 2 weeks its linux mint. That matt guy was so against linux mint that i thought it was shit too. But when i installed and started using it. It has been a smooth journey. Many people in linux community were helpful. But people like matt really make it for us regular guys scared to use linux. I really hope many good linux user help regular people switch to linux and increase this number.
I assume the problem is hardware. Matt’s hardware didn’t work well with LM, therefore Matt thinks LM sucks… I do wish there was better hardware support but it’s the reason apple went with 1 product = 1 OS = 1 general set of hardware. Sure not every iPhone has the same hardware, but that’s why they have the model numbers, and it’s so much easier to test 200 model mixes than 2,000,000 (Android). Windows gets all the debug info sent directly to them like the others but they also have a huge stack of hardware they can use or they can buy it to test.
Did Matt try putting the regular build on a newish machine? That’s what I did with my current and was struggling until I put the latest kernel on it, should have gone with Edge, but had little trouble after)
This is an interesting post. I know the YouTuber you are talking about. I do not really like him but I would have considered him harmless. It is interesting to get your take.
Mint is great and quite popular. So, if you are going to make a YouTube video saying otherwise, I guess you have to go out of your way to come up with reasons why. I am sure that makes it sound worse than it is. Something to think about.
I have distro preferences too but most of the differences really only make a difference to those already using Linux. It is a bit like arguing about Ford vs Chevy in front of people that have never seen a car ( or truck ). In the grand scheme, they are both amazing and mostly the same thing. To listen to a fan though, one is God’s gift and the other is trash. Same with Linux distros.
I get what you mean. I see a decent chunk of often more tech-proficient Linux users putting down Linux Mint, and it saddens me because even though I don’t use Mint anymore, it was still the first distro I properly daily-drove and I still consider it an amazing system for people who are new to Linux.
I’m very glad you’ve been having a good experience with Mint!
I don’t get them putting Mint down either, and I’ve built multiple Gentoo systems… I don’t need an easy distro but still use Mint and like it for what it brings (basically, it’s Green Ubuntu, what Ubuntu was supposed to be before they lost their way)
Mint is great and is absolutely enough for most people using computers, still as of now. It comes with its limitations though:
- By default it runs pretty old kernel. This is fine if your hardware is at least 3 years old. It allows to easily switch to newer kernel with just few clicks, but I expect newbies to not be aware of this at all. Oh, and I don’t know if it offers some custom kernels like tkg etc, which some might want to squeeze best gaming perf etc.
- Cinnamon is still limited to X11. If you have multi-screen setup, VRR, mixed refresh, mixed DPI etc, it’s better to switch to Wayland. Plus, Xorg server gets less and less maintenance and development. All the innovation moved to Wayland, so the experience on X will remain pretty stale.
- The Ubuntu base makes it so that for 3rd party software you either need deb packages or PPAs. Some will argue (me included) that it’s not the best solution
All of the above can easily be irrelevant to you and Mint is just perfect for what you need. It’s important to point out limitations of that choice, but crapping on it because you don’t like it is just pointless fuss
On your last point, there’s also Flatpak which is available right from the baked in software center… That’s not without its issues too, but they’ve been an overall smooth experience for me so far
Yes, Flatpak fixed a lot of the old shenanigans we used to have when everything was either native package, or a binary to hope for the best and install libraries manually, or source code to collect everything that’s needed for building and again, hope for the best. It is however designed to provide a way to install graphical apps, but can’t handle everything native package does (like out-of-tree kernel modules, CLI utils, system services)
I believe it can do CLI, but that’s not always been the case and not a lot of CLI apps adopted it as a result
But for most of what the typical user, or even a lot of what a technical user, needs, it does a good job
Do you have an recommendation for a distro? I wanted to use mint, but i probably need wayland for a multiscreen setup with different scalings.
Well, if your GPU is NVIDIA, you will also need a bleeding edge rolling release distro for now. Other than that, anything that ships recent version of KDE Plasma or GNOME (the first one handles Xwayland with DPI scaling a bit better imo and is generally more functional)
Garuda (Arch-based distro) is worth looking at, but has a ton of UI customization that I personally don’t like and, same as with Manjaro, I’d spend some time on cleaning it up to more vanilla state.
Endeavour is cool option as well, but other than GUI installer, it provides you with Arch base and is purely terminal-centric.
OpenSuse Tumbleweed - personally never liked RPM based distros, and Zypper (the package manager) is pretty slow, but the distro itself is beloved for being great balance between bleeding edge and stability and set of GUI tools for system management.
Fedora (defaults to GNOME but has decent KDE spin) is great option for a workstation, but it has couple of advanced configurations that not everyone needs (like SELinux getting in your way sometimes) and because it’s very concerned about licensing, you might need to manually add extra video encoding/decoding drivers and media codecs with royalties.
Nobara is modification of Fedora that comes pre-packed with gaming related stuff. It’s pretty cool, but in my experience it tends to ship experimental packages with regressions.
Last but not least, if you’re patient enough to read through some documentation and follow guides step by step, can invest some time and want to learn, go with Arch Linux. The ISO ships now with text-based
archinstallinstaller that guides you through basic install and brings fairly functional Linux system out-of-box EASY, but still pretty bare bones without any assumptions on what you’re going to use it for and how. Everything that you need (like gaming, bluetooth, printing, file sharing…) has its Wiki pages with thorough explanation in form of step-by-step guide. It’s all about installing packages and changing config files, it’s not doing anything automatically for you.Well, if your GPU is NVIDIA, you will also need a bleeding edge rolling release distro for now. Other than that, anything that ships recent version of KDE Plasma or GNOME (the first one handles Xwayland with DPI scaling a bit better imo and is generally more functional)
I keep hearing people say this. But I’ve got an nvidia card, and I just went with the default Mint Cinnamon install and I’ve had no problems whatsoever. I guess maybe my card isn’t new enough to run into whatever problems other people are talking about.
… Actually, there is one minor annoyance. I get lots of nvidia flatpak updates; and they are large downloads. I’d prefer not to be downloading gigabytes of graphics card updates every week. But other flatpaks demand that I have the latest nvidia stuff, so … I guess that is an nvidia annoyance that I experience. I don’t expect that to be fixed by a bleeding edge distro though!
Yeah, Flatpak installing user-space driver for itself is unfortunately not solvable until there’s open source driver that is part of the Mesa project. Every time you update the driver in your system, Flatpak must update its nvidia-utils too, because their versions must match exactly. For Mesa drivers, Flatpak also installs the drivers as Flatpak, but they’re compatible back and forth and it only updates when it ships new version.
The cleanup should be more automatic, but try
flatpak uninstall --unusedBecause it always worked on X11 and Mint Cinnamon is just that. I used NVIDIA graphics on X11 in 2007 and, apart from the extra dkms driver that could break at times, it was fine, and much better anyway than ATI/AMD with proprietary fglrx driver. Rest in piss son of a bitch.
The question was what would I recommend for Wayland. Only the brand new 555 driver combined with most recent compositors (and other packages like mesa, xwayland,…) offers decent NVIDIA experience. It’s a matter of new distro releases around this fall.
Mint should have Waylant support if you don’t use Cinnamon; I know Xfce has Wayland support (though I don’t use it, they can pry X11 from my cold dead hands)
Neither Cinnamon, XFCE or Mate have stable Wayland support. You need GNOME or Plasma for that if you want a desktop (or wait for the new Cosmic desktop and new PopOS)
Cinnamon can run Wayland in experimental mode. It’s just an extra click during login. Mint also has direct support for flatpaks repositories, with flathub by default directly on the software center.
Any experience you can share on how complete and stable is that experimental session? I probably wouldn’t throw newbie on that
There’s a bug where flatpaks seemingly disappear from the system the first time you run Wayland. But it resolves with a reboot. It happens too if you change back from Wayland to X11. Other than some minor glitches from very old software that hasn’t seen an update in decades, it runs perfectly fine.
I am still hoping it will hit 10% market share within my life time. I remember when it was predicted to hit that in 2010, obviously it didn’t happen*. Of course for me personally, the year of the Linux Desktop was 2007 when I was finally able to use it as my main OS at home, I tried it before many times since 2003.
* not counting systems that use the Linux kernel but aren’t considered a traditional GNU+Linux desktop.
I am still hoping it will hit 10% market share within my life time.
Do we really want that?
We have it pretty good right now. I would actually say we’re living in a golden age of desktop Linux: there’s constant innovation, good support, you get to do pretty much everything you need, while flying under the radar.
Linux has won the majority of the industry (servers, mobile etc.) so it’s not like it has anything left to prove.
If it starts getting noticeable on the desktop I fear we’re just gonna get negative attention. Users who take and not contribute, because Windows had taught them to be entitled. Unwanted attention from Microsoft, who I bet are not going to be doing nice things once they start getting paranoid about it.
I really don’t think that large companies like Adobe will care about Linux even at 10% and even if they did, they are a super toxic company nowadays, the least we get to interact with them the better.
There are many games & software with no Linux support, not to mention AC blocked games. Increased marketshare could change a thing or two, at least.
Do we really want that?
As long as competition and choice continues to be the mantra of the Linux desktop, then yes, I’d love to see more and more people using it.
We have it pretty good right now. I would actually say we’re living in a golden age of desktop Linux: there’s constant innovation, good support, you get to do pretty much everything you need, while flying under the radar.
Very true.
Unwanted attention from Microsoft, who I bet are not going to be doing nice things once they start getting paranoid about it.
I mean, Ballmer called Linux a cancer pretty early on, so that ship sailed a long time ago.
I really don’t think that large companies like Adobe will care about Linux
Once they start losing large sums of money due to people switching and finding viable alternatives, they certainly will care. Right now Adobe has one main thing going for them – apathy and muscle memory of the aging demographic of their users. That will eventually change.
the least we get to interact with them the better.
Absolutely. I used to be an Adobe fan, back when Kevin Lynch was a part of it, and I was a Flex developer. Then Jobs wrote his thing about Flash, and a year later, not a month after Jobs’s death, Adobe dumps Flex – and literally overnight my position changed from Flex to HTML5 and Java.
not counting systems that use the Linux kernel but aren’t considered a traditional GNU+Linux desktop.
Does that mean you don’t count Alpine towards Linux market share? It mostly doesn’t use any GNU stuff.
You can also compile the kernel with LLVM instead of gcc, use musl instead of glibc, and use BSD coreutils instead of GNU coreutils.
Do they use the BSD userland instead? Interesting…
Perhaps the definition isn’t good enough or accurate. What would you call a system that perhaps uses Darwin kernel or Hurd plus GNU user land, or any combo of.
Do they use the BSD userland instead? Interesting…
I think Alpine uses Busybox, but it’s feasible for a Linux distro to use BSD coreutils. Not sure if any do that, though.
OK, who did you guys bully over basic tech support questions?
uMatrix prevents me from loading this statcounter website. :-( Can’t lookup how they measure things. In the comments people assume the stats would be counted by just looking up the user agent, which is a naive approach. I don’t think agencies dedicated to stats are doing it this simple. They have way more possibilities to track and to look what browser you are using. The stats are more accurate than you probably think. They do not need to know the exact version of browser you are using, just which type and maybe what operating system you are on.
If a script for Windows does not work on Linux and fails, then they know you are not on Windows in example.
Why are you still using uMatrix in 2024? Wasn’t it discontinued 3 years ago or smth?
It’s still useful with all its functionality. Even if it does not get updates, it still blocks most stuff and I can enable or block stuff too. Plus the blocklists it uses are still updated, as they act independently from the main addon. There is no better alternative in my opinion. I use uMatrix + uBlock Origin.
I wonder what’s on the ‘unknown’
Amiga.
Bare Metal, they are injection Ethernet cable directly into their bloodstream.
In essence Netrunner then?
That 6% attributed to “unknown” is the one true OS, the only one ordained by the Almighty… Temple OS!
True Temple OS has no networking
Based security approach
It does have networking if you truly believe!
Heretical, you will burn in hell
likely content blockers preventing the trackers from working properly and invalid user agents. So i would expect about the same ratio of usage on there as well. Maybe very slightly more Linux since maybe the users are more likely to tinker with their browser configs and install content blockers, but even there Id say its an extremely slim minority of even linux users who do that
StatCounter also sometimes miscounts when new versions of windows or macos come out. At one point (I think at windows 11 release) there was a huge dip in windows 10 users and a huge gain in “unknown” and it was quickly fixed.
