Steam will no longer support Ubuntu 19.10 and later

From Ubuntu 19.10 Steam will not support or recommend Ubuntu anymore.
Is Ubuntu no longer going to be a desktop distro?
This choice will drive away many users, including me, because I have many games on Steam.


This is concerning for me, as Ubuntu is the only distro that I dont have to fight with to get my ThinkPad nvidia graphics to work correctly. I was hoping to completely remove WIndows with the great work Valve has done with Proton. Now I fear I will have to go back to Windows. I really with Canonical would re-consider this timeline and work with developers to transition to 64-bit.


1 Like


My opinion only:

The Ubuntu developers have talked about taking this step for years, and have repeatedly explained why Canonical does not want the burden and cost of maintaining these 32-bit libs anymore.

So what? Lots of Ubuntu is already NOT maintained by Canonical engineers.

If you want 32-bit libs maintained and in Ubuntu, then form a community project and take over the maintenance. Specialized repos (like Medibuntu and Backports and Partners) have a long history in Ubuntu, and all the infrastructure is already in place. Just need the game-loving volunteers to step up, get organized, and make it happen.


You’re misunderstanding the main problem.

Now that Ubuntu is taking this step, developers of commercial software for Linux will certainly release 64-bit binaries in the future. However, nobody will touch legacy software (like drivers for hardware released before 19.10 or old games) and new software written for Windows (which we might want to get to work via Wine) will probably still frequently contain 32-bit binaries for at least another decade, since Microsoft will support 32-bit applications forever.

I’m not going to comment on the political aspect of this, since Canonical has already made their decision and since I’m in no position to tell them what to do, but that’s the technical side of it. 32-bit apps that people might want to get to work on Linux will not go away because of Ubuntu changing to 64-bit only, no matter how much Canonical lobbies for it.

Containers might be a reasonable solution for some of that software, but someone still has to build and maintain a working container for each legacy application, which means it’ll only happen for the most popular software packages, if at all. And that solution really doesn’t help the average end user who has no idea how to containerize legacy software.


Correct me if I am still misunderstanding this, but Valve wont make a steam.deb for future versions of Ubuntu. Even though 64-bit apps will work fine, they will not be able to support legacy 32-bit, therefore they wont make a simple deb installer. Would that be a correct statement?

Yes. Even if Valve ported the Steam client to 64 bit (AFAIK, it’s currently a 32-bit application), you wouldn’t be able to install most existing Linux games on 19.10 and most likely not even a single game supported by Proton / Steam Play.

After all they’ve already invested in Linux gaming, Valve would now have to maintain the entire 32-bit Ubuntu stack themselves if they wanted to continue supporting Ubuntu, so I definitely cannot blame them for dropping support instead. From their perspective, Ubuntu just decided to stop supporting Steam and Linux gaming, not the other way around.

1 Like

Take this recent post as you may…

I’m not a gamer so I’m not personally concerned with the change. I think this is not necessarily on detrimental of Ubuntu being “desktop oriented”. Gaming is just one aspect of the desktop. If 32 bit libs are an important maintenance burden and the released resources become focused on other aspects of the desktop, that’s a win. After all, Linux in the desktop is a niche product that is much better described as developer-focused than as gaming-focused. Linux is popular and loved in StackOverflow surveys and yet is struggling in the desktop after a traumatic decade full of not yet polished significative changes: Gnome 3 experiments and forks, Wayland, Snap/Flatpak and others. It’s better to stabilize and polish the desktop experience, provide hi-dpi and multi-dpi first-class support, a good package manager, bullet-proof updates, etc. I realize this is a dangerous line of reasoning, because the freed resources might become unemployed or mis-employed instead, but I prefer to believe in the diagnosis made by the developers themselves. I see movement and convergence in the desktop front, Canonical employees are actively contributing to Gnome, GUIs and APIs are stabilizing, we now have some kind of “default desktop” (whether you love it or not). But they aren’t a lot of developers and they are undertaking a formidable task that most of the time is not even fun, but yet is much more important than the next fancy desktop or distro. So I want to believe they are looking for the smartest way of allocating scarce resources. For the little I know about that, I think it’s realistic to suppose that Linux will never be on par Windows for gaming. If the burden of maintaining a set of libraries necessary for gaming will relocate critical resources in order to set them working on non-critical goals, that won’t be following the Pareto principle. So let gaming support for a community of enthusiasts. If they have better things to do and prefer, instead, to run Windows in a VM or dual booting their gaming sessions, well, they will be optimizing their time too.

Linux not being the mainstream OS for gaming is not a reason for making it harder for game devs and users. The amount of games natively ported to Linux is constantly increasing, sure it is not a fast growth but it is indeed growing. Valve recommended Ubuntu because it is the easiest distro for newcomers from Windows and Mac. In my personal case I don’t want to use Windows 10 due to the spynet it is and the horrible update mechanism it has, I also know someone who is really tired of Windows for the same reasons and wants to migrate to Linux, but if that person finds out that most games do not work out of the box (and other kinds of 32-bit software) then that means we don’t have a real alternative.

Ubuntu always had a reason for existing on the Linux world, to be an easy to use distro for people that want to escape from Windows and Mac but still want a convenient way of using their computers. I don’t want to see Ubuntu get less user-friendly at a time more and more users are migrating to a platform that respects user privacy and free software. And Ubuntu is still the easiest to use Linux distro that maintains a nice balance of both stability and new software, a sweet spot that really hasn’t been reached by any other distro that I know of.

Yeah, a desktop supporting all kind of games is -ceteris paribus- surely a better desktop, but my point is that things like moving forward Wayland to get decent hi/multi-dpi support, improving packaging and distribution, etc. are more important than gaming for the desktop, they are BASIC. These deficiencies probably alienate more users than the inability to game. Then, it’s up to you to buy the resource allocation argument or not. I closely follow development on the Gnome front and I decided to buy it because I realize how scarce paid resources are (by Gnome Foundation, Red Had, Canonical, etc). Maybe Canonical is “killing” the desktop or whatever, but it doesn’t feel like that to me. Also, try convince people to abandon the “new day new distro” program and organize around useful stuff like a repo of community-supported gaming libraries, for example.

Although games are probably more widely used by the desktop community at large, I know from experience that there are a lot of small companies that rely on software to run their businesses whose vendor long ago went out of business themselves. Those software programs are NEVER going to be translated to 64-bit and many small businesses just don’t have the resources or the time to convert to newer, supported programs (if it works, don’t fix it). Linux itself had a similar issue some years back when they dropped 16-bit support from the kernel. That issue hit me as well as I still have a lot of Win16 programs I run for which I can find no equivalents. There was enough backlash on that decision that 16-bit support was put back into the kernel. I hope Canonical pays attention to this history.