It’s because SteamOS identifies itself as Arch. Omitting this information is either dishonest or uninformed.
Because you hear “Arch” and it gives the impression that they’re being played on a Linux desktop, not a Steam Deck
While that may be true, I still use my Steam Deck in desktop mode for a bunch of stuff besides gaming. Writing, job applications and interviews, using reddit because it’s the only device I have that isn’t detected for ban evasion, watching shows/Youtube. Maybe I’m atypical, but I don’t see why the Deck would offer a desktop mode if it wasn’t meant to be used.
Steamos identifies itself as “SteamOS Holo”.
Also, that article isn’t measuring SteamOS in the first place. When you look at the steam survey with the default filters it won’t list SteamOS. If you switch it to Linux only it will show SteamOS as 36.47% of Linux installs (0.84% of all steam installs) so it’s clearly not feeding into the Arch percentages.
This is very obviously false. With the default filters with all OSs shown, Arch has 0.20% marketshare and Linux has a total of 2.29%. That means Arch is about 8.73% of all Linux systems in the survey. If you select the Linux only results, then SteamOS appears as its own entry, alongside a few others like Flatpak. We can see two things here:
- SteamOS Holo is 36.47%. This was very clearly not counted as a part of Arch Linux in the all OSs tab.
- Under these filters, Arch is even higher at 9.7%.
What’s impressive here is not just the confidence with which you called the article dishonest and uninformed while not spending half a minute to check your false assumption, but also how many people upvoted you. This was trivial to prove wrong and in fact people have already done that below. Why are people so eager to believe the article is wrong that they will jump to agree with a blatantly wrong comment while having no knowledge of the situation themselves?
I’ll take the L on this one. It’s a combination of the article only using the screenshot of the first view as evidence and me late night posting on Lemmy while falling asleep via NyQuil.
Am I missing something or is 36.47% not greater than 9.7%? Why is SteamOS not shown as the most popular Linux distro without the Linux only filters?
This contradicts the article claiming Arch dominates the Linux gaming scene and not SteamOS.
SteamOS seems to not be counted at all in the first page. Apparently, it’s not just “All OSs combined” vs “Linux only” but there are additional filters applied. Perhaps the first page is desktop-only. The article either also cares about desktop gaming specifically or is uncritically parroting the survey page. I think both Valve and the article writer should be clearer about what they’re talking about.
btw
Isn’t it because SteamOS is based on Arch?
Based
I feel underrepresented as a Void user.
Although the absurd number of hours I’ve played a certain popular gacha under Lutris might not trigger the Steam metrics, I demand credit for dumping 45 hours into a poorly translated RPG Maker looking project!
I’ve thought about Void. And LFS. And I submitted some packages for Alpine, although I’m not running it anywhere except as container bases.
Last time I really strayed from the Arch ranch was Artix, and that was TBH pretty painful on a day-to-day basis.
I’d like something like Arch but with less systemd. ChimeraOS looks promising, once it stabilizes. But how’s Void treating you? How’s xbps? I’m pretty in love with pacman; rolling release is a must, but IME you really only realize how good or bad a package manager is after it’s too late, and you’ve been using it long enough to hit your first dependency hell/upgrade issue. After years of hell with RPM and deb, pacman was a godsend.
runit isn’t my favorite initd alternative (dinit ftw, at the moment), but it beats systemd and I don’t have a huge amount of experience with it. Do you like it?
Critical to me is being able to easily toss together package manager recipes for stuff that isn’t in the official repo; I really believe in keeping systems clean by only installing through the package manager. Pacman packages are stupid simple to write and easy to work with, and yay makes things even better. How’s xbps in this area?
EFS boot is easy? Stuff like btrfs boot partitions and snapper support easily available? No idiocy like trying to force users onto Wayland prematurely?
Runit works well enough for me; I’ve only added one nonstandard service (launch a custom tool to drive an external stats display) and it works fine. My ,xsession has to load some polkit and pulseaudio stuff but that could be because I’m not using a full desktop like KDE/GNOME/XFCE that do those things for you.‘’
I don’t really try to do custom package recipes because I tend to ./configure;make;make install stuff I want at random.
EFI boot is no problem. I think my root is btrfs, but the /boot/efi is vfat. Refind is pretty first-class, but sometimes it has stupid conditions where it tries to default to the wrong kernel version if you have multiples installed (I think it sorts by timestamps or filenames in a way that sometimes work counterintuitively; discarding old kernels largely fixes it)
Haven’t really had too many showstopper problems with xbps. I probably sledgehammer it a bit-- occasionally when it says a repo certificate is out of date, I usually end up doing a full update rather than selectively upgrading packages.
I’ve been enjoying void on an old Thinkpad just to mess with. How’s the gaming experience been on it? Any issues with Steam/Proton running well?
Steam runs fine. I think I had to install some Vulkan packages manually because I was getting some hallucinogenic colours in Genshin Impact (installs fine via Lutris). I have a few minor issues with games not loving losing the mouse cursor if you move it onto another display, but I think you can tame most of them by running in Gamescope so it doesn’t realize there’s a second monitor the mouse can leave to.
