Can Steam finally force developers to make Linux versions?

images
yeah, that would be nice... might have changed since I last checked, flatpacks or whatever are a relatively new thing, I suppose, and replace deb and rpm... but probably just another "standard" because Linux people love to fork everything instead of actually nailing the first thing ... but some apps just require /configure /make /make install, and won't uninstall clean or even at all when you click on uninstall or just want to upgrade to the current version some months later.
Linux is nice, for a free thing, with a pointless number of desktop environment choices, but it's so easy to run into weird problems, just because no obvious reason, just because a bigger upgrade failed... whatever. Valve would be wise to not do anything stupid. Some gamers might switch, being fed up with forced Win11, and it is finally a time were gaming is working on much stuff, but it needs to evolve a lot more to be actually noob proof and then with bigger market share trigger more native support.
 
Until your average PC user can navigate Linux without issue, Linux will always drag behind Windows. I have said this time and time again. Linux is too fragmented to reliably handle most PC users. The reason MacOS does well is because a user can navigate it with relative ease.

Arch Linux, Bazzite, CachyOS, and all of the other Linux distributions suffer because Linux isn't universal. Want to get an app? Time to figure out how exactly to do that because there are different apps and "app stores" for different versions of Linux. Different desktops, different conventions, different settings panels, different theming that breaks in different ways. A user learns muscle memory in one place and it does not transfer. Right click here does one thing, right click there does another. Settings live under three different menus depending on the OS.

Package managers are another wall. You've got apt, dnf, pacman, zypper, then Flatpak, Snap, AppImage. Then distro repos versus third-party repos versus vendor scripts. This is too much for a new user to try to navigate without making it a serious time commitment. If the path to install an app depends on a flowchart, the OS is not ready for the average person. Software availability is also a major factor. Missing Adobe is not a small thing. Not being able to easily use Microsoft apps, especially for people like myself who HAVE to use them for work, makes it a no-go. You can argue alternatives, but the industry standards are the industry standards.

The Steam Deck works because Valve hid the fragmentation. One store. One runtime. One set of defaults. One recovery path. That is the model that helps non-technical users. Sure, people can go to the desktop and do more than that. But the Steam Deck, much like traditional game consoles, benefits from the simplicity of "install the game and go". That's what the average user wants. The more the experience looks like that, the less Linux feels like a chore.

This is why I want Valve to seriously work on SteamOS to make it an actual OS that can be reliably used by the masses, not just by people who are fine with learning command-line and all the different ways that Linux operates. I hate to say, "Be like Apple," since I think Apple is a sucky company, but I want Valve to be like Apple was with the start of MacOS. Understand what people need from the OS, and work towards making it something that can accommodate the end-user rather than telling the end-user to change in order to accommodate the OS.
 
Last edited:
I think a lot of people don't realize what they are asking for is an oem device that ships with linux. Steam deck is one of the very few pc's that doesn't ship with windows as an option. When it came out and people tried to install windows, stuff was broken.

I think the way forward with steamos is to get it on more and more oem devices. Some general distro of steamos is not going to magically make everything work like a steam deck.
 
Arch Linux, Bazzite, CachyOS, and all of the other Linux distributions suffer because Linux isn't universal. Want to get an app? Time to figure out how exactly to do that because there are different apps and "app stores" for different versions of Linux.

that is the main issue.
Snap and FlatPack exist, but FlatPack versions of apps can have limitations that a dedicated distro binary doesn't have. now try explaining that to the average PC user, that the FlatPack version of the file-search browser they have in their preinstalled app store doesn't have access to the list of your default apps for certain file types, so you get constant pop-ups and the context menu can't list certain apps or app based context options...

so literally how you install an app can change the functionality of it... and installing it without just downloading the flatpack can mean also looking through a list of dependencies to make sure you're not missing something
 
Last edited:
Baldur's Gate 3 already run native linux on steamdeck, put over hundred hours in it. I would like to see native becoming forethought insted of after for developers, valve is doing more good on that front then anyone else.
 
The Steam Deck has sold 3-4 million in 3 years. Valve may even be working on updating the Steam Box.

The next step for Valve is to pay AMD to make a custom SOC. Eventually Valve may get numbers rivalling Playstation and Nintendo in their "consoles". Many devs will start making native Linux versions of their games.

As Xbox exits hardware, Steam enters the space as an alternative. Eventually Linux will get the same native 3rd party support as Xbox. This will incentivize many PC users to go full Linux. Eventually Windows may become the less attractive choice and some devs will start skipping Windows entirely.
No one is going to switch to Linux anytime soon. It's like 4% market share and trying to force developers to make games for an extremely niche OS is a really dumb idea.
 
The only influence Valve has is if they create a sizable install base of SteamOS or Steamdeck users but that install base is smaller than 9M I believe. Worse than an xbox or VR. They can try and pull a "series s" move and mandate a linux version if a publisher wants to publish to steam but I don't think that would be wise.
 
Last edited:
Valve could in theory offer some sort of discount to developers that make their games Deck verified, but that's about it and every reason why has already been said.
 
Don't the game engines need Linux compatibility also?
Are the popular ones (unity, unreal) available for Linux?
 
Doesn't Valve make 99.9% of its money through Windows?

If Windows (game development) collapsed where does that leave Proton? Proton needs Windows as a base to "convert" the games I assume. Or is it much more than that?
 
Last edited:
What Valve needs to do is create some kind of protected mode for SteamOS with robust anti-cheat built in that devs can hook in to.
 
Even Easy Anti Cheat has a Linux version.
So it's just up to the game developer to have support for it.
I wonder if that is actually possible since "Linux" is barely meaning something specific. Developers might decide one day that they support SteamOS and then need to have playtesting and customer support on exactly that but supporting Linux with all its hundreds (?) versions that may or may not be compatible would probably be a nightmare. In the end it might be SteamOS, original Arch and Debian, maybe Ubuntu, maybe Fedora and everything else is not officially supported, 5 is probably already too much.
 
I wonder if that is actually possible since "Linux" is barely meaning something specific. Developers might decide one day that they support SteamOS and then need to have playtesting and customer support on exactly that but supporting Linux with all its hundreds (?) versions that may or may not be compatible would probably be a nightmare. In the end it might be SteamOS, original Arch and Debian, maybe Ubuntu, maybe Fedora and everything else is not officially supported, 5 is probably already too much.

I doubt devs would target a distro. Rather they would target a kernel, API and driver version.
Maybe someone that knows more about Linux can confirm this.
 
I doubt devs would target a distro. Rather they would target a kernel, API and driver version.
Maybe someone that knows more about Linux can confirm this.
It just has to be something specific. However that is reached. Unlikely but otherwise some idiot might think his Server Edition Linux which just comes with Terminal might suddenly run games...
I know the app I needed some years ago only mentioned Ubuntu, maybe Debian too, I think two platforms were supported, but that was all they mentioned on the website. Kubuntu, Xubuntu, Lubuntu and whatnot was already "your gamble", iirc they wrote it might work on Ubuntu forks or something like that but don't ask for support. That was before they dropped Linux support alltogether.
Same with my main work app. Had Linux and MacOS support 10 or 15 years ago but discontinued everything but Windows. I think there is only one alternative left that kept Linux and MacOS, while all others dropped their Unix and Solaris origins.
 
Not political if your anti-cheat is kernel level as Windows is a fixed closed kernel whilst Linux you can build your own.

This could easily be circumvented as a separate Linux kernel module. Several proprietary software programs, such as Nvidia and Broadcom drivers, work this way.

So it's more a matter of having the willingness to provide support. Some developers have denied support for Linux, and none of them did so for technical reasons. They all said they would only support one platform. When I talk about support, I imagine it includes patches, help desk, moderation, etc.
 
that is the main issue.
Snap and FlatPack exist, but FlatPack versions of apps can have limitations that a dedicated distro binary doesn't have. now try explaining that to the average PC user, that the FlatPack version of the file-search browser they have in their preinstalled app store doesn't have access to the list of your default apps for certain file types, so you get constant pop-ups and the context menu can't list certain apps or app based context options...

so literally how you install an app can change the functionality of it... and installing it without just downloading the flatpack can mean also looking through a list of dependencies to make sure you're not missing something

I don't think that's a problem. It seems more like laziness to me.
I find the excuses curious, it's as if everyone was born knowing how to use Windows and didn't spend years learning various methods.

On Windows: You can download software from a website, risking ending up on dubious sites and getting "contaminated" software. When you download, you'll find .exe, .msi, .msix, .bat, .vbs, .dll, compressed files (rar, zip, 7zip, tar). The installation window varies considerably, and if you're not careful, you might install unwanted software. There's also the Windows Store, with its strange software. And there's Winget and Chocolatey, which allow installation via command line. Speaking of which, you can also install software using PowerShell.

On a Mac it's the same, you have the Apple Store, you can download .dmg files from a website, some install by clicking on them, others you need to move the file to Applications, others you install via command line using brew.

Linux is usually the easiest of all: they copied the Android model. Almost all distros have a central hub where all the software is located. Distros like Arch are DIY, made for customization. Even so, even using the command line, it's practically foolproof. If you can read, you can use it easily.

Not to mention the various configurations people need to make, such as using debloaters to clean Windows or when they need to troubleshoot software that won't work or is malfunctioning. Overall, it's been years of learning.

Linux is a different system, you really need to learn new things. It's no different from migrating from Windows to Mac, Android to iOS, Xbox to Playstation, etc. There will always be a learning curve.
 
Honestly, if I were Sony right now, I'd be contributing to Bazzite development and releasing Steam Deck versions just to help undercut Windows.
 
Steam does have a Linux compatibility layer so as a dev you can target that with a native Linux version and be reasonably confident that it is going to just work going forward -- I will sometimes resort to it to run non-Steam native Linux games that otherwise fall over. That said, when I think about developing a wee game, even if I set up the dev environment on my Linux box, I really want to be targeting Win32/Proton as the platform to build for. Developing Win32 (well, technically 64-bit but you know what I mean, the Win32 API) exes on Linux and testing them directly under Proton means you can be confident you're serving Windows & Linux Users -- as soon as you veer off into supporting a second distinct platform the maintenance cost goes up exponentially. Don't even mention OSX.
 
Better to persuade than force.
This.
Release desktop SteamOS first, insure it meets all security requirements for native anti cheat (devs will want it) and give devs good tools for porting from Windows to SteamOS and for updating games. Also get Nvidia to create better drivers for Linux and keep them updated, so the share of SteamOS users slowly and steadily starts to grow.
 
I doubt devs would target a distro. Rather they would target a kernel, API and driver version.
Maybe someone that knows more about Linux can confirm this.

Kernel versions matter mostly for performance features, not baseline compatibility. The kernel is essentially responsible for keeping userspace ABI stable. The real target is APIs, driver versions, and a controlled runtime. Devs will still do QA testing on specific distros. If SteamOS gets released as a desktop OS, you can bet that devs are going to make that one of their Linux distros to test on.

Because different distros can have different performance impacts, it is possible that a dev could say a game is Linux-native, but that they will only support issues on SteamOS. It wouldn't mean the game would fail to run, or even that it would have a negative performance impact on another OS. It would just mean any issues would have to be replicated on SteamOS, and support for a different OS would be limited. It's what devs already do with Windows. "The only OS officially supported is Windows 11." That doesn't mean the software doesn't work on Windows 10 or Windows 7. It just means that wasn't the targeted and tested OS. And if I'm being honest, with something like Linux which has an insane amount of variation in the different OS options, that's probably the wisest course of action if Linux-native games become more mainstream. Pick maybe 3 major distros and do your testing on those. Everything else could work, but guarantees can't be made.
 
Last edited:
Anti-cheat isn't exactly a problem, it works on Linux. It's more of a political issue than a technical impossibility.
There's a website that lists which games work with anti-cheat:

Oh, I know. Which is why I said I'm waiting for that "Happy medium" between the Users and the Devs. I play the AAA shooters like Battlefield, Call of Casual, Apex, etc. And with GTA6 coming, I ain't got time for no bullshit. But once all the parties involved act right, I'll be on Linux immediately. I'm sick of Windows. Only reason I don't dual-boot is because I'm lazy. Sigh
 
That hippy OS runs 90% of the internet.

BRmihYqsE2U4nhD3.png
Doesn't matter what % Linux is used on the backend if all you get out of it is a measly 4% userbase. Noone is going to switch to Linux anytime soon and trying to force people to is a good way of alienating them from your platform
 
Top Bottom