SteamOS, the Valve-made Linux distribution that was originally for the failed Steam Machine initiative has gradually vanished into the sidelines but it seems it won't be forever.
A while ago, we did see indications that Valve would work on SteamOS 3.0 "Clockwerk" back in 2018 but they've still been very quiet on it since apart from a few minor package updates to SteamOS 2 "Brewmaster".
Valve have been extremely active on other fronts though of course. As a quick bit of history: for Linux they put out Steam Play Proton, the ACO shader compiler for AMD, this new Steam Linux Runtime container system, the micro-compositor Gamescope and there's more with people working on all sorts under contract for Valve to improve Linux.
Still, SteamOS though, what are Valve going to do with it? Sounds like when they go back to it eventually, it might not be Debian-based. In a GitHub issue on the SteamOS page about it "languishing", another user replied with an email from Valve developer Pierre-Loup A. Griffais:
Yes, definitely lots of work still going on. Right now the focus is on core technology itself rather than distributing it, but we intend to get back to that in the future. I wouldn't expect much more movement on Debian-based Brewmaster at this point, however.
"Debian-based Brewmaster"—huh? Speculation here, but that sounds like they might be looking at a different base for whatever SteamOS 3.0 turns into.
One day then, we can clearly expect to see some movement on SteamOS once Valve get all the pieces of the Linux gaming puzzle into a state where they're truly happy with properly pushing it again. Perhaps, this will be after we finally find out what the heck Steam Cloud Gaming(#1, #2) turns out to be? Whenever we find out, we will of course let you know.
Thanks for the tag, mdeguzis.
Quoting: WorMzyWhich in turn doesn't sound too practical for something I've always assumed was supposed to be a pretty turnkey affair--hit a button and be ready to play games, kind of thing.Quoting: rustybroomhandleI think Arch-based would be the sensible option here. Arch-derived distros just seem way more stable and upgrade-fubar-proof than debian-derived ones.Really? Debian is the gold standard for stability. The problem is that people want a stable OS with a "bleeding edge" graphics stack, and that's where the house of cards falls over.
Arch can provide the bleeding edge, but it requires that YOU (the end user) know what you're doing.
Quoting: HoriQuoting: AllocHaving a full-featured open source "console" desktop from Steam would be cool imo. So not only Steam's BPM but something that resembles consoles more. Supporting streaming services in the interface (Amazon, Netflix, Spotify etc ...), playing videos from the local network, maybe support for Plex. So that you could really use that thing as a TV box and not have to hop around different interfaces for everything but games. Maybe they'll get there ...
Either way I wouldn't care about the underlying distribution ... the SteamOS thing (for me) should just work as-is, I wouldn't use that for desktops anyway.
I never understand why people want Netflix and other apps like that in Steam / consoles... don't their TVs already have those apps? Why would you fire up your pc/console just to watch a movie when you can do it straight from the TV, using the TV remote instead of a gamepad?
The thing is, many Smart-TVs are highly insecure with far and few between security updates, if ever. For this reason, I avoid using them, as much as possible, and replace the "smart" part by a TV-Box that gets maintained regularly.
Last edited by Mohandevir on 25 March 2020 at 7:03 pm UTC
Quoting: WorMzyReally? Debian is the gold standard for stability. The problem is that people want a stable OS with a "bleeding edge" graphics stack, and that's where the house of cards falls over.
Arch can provide the bleeding edge, but it requires that YOU (the end user) know what you're doing.
Considering "bleeding edge" is a 16xx series nVidia which is around the $350AUD price range or needing to compile mesa (for newer AMD gfx) to have working on debian based distros is a pain in the you know where compared with Arch there is already an AUR ready to use.
Remember in order to appeal to current gamers you need to support newer hardware to those that are getting sick to death dealing with Win10 issues. Debian based distros unfortunately fall too short of that mark.
Look don't get me wrong, I have a laptop that has a GTX 960m and it more or less works out of the box with any deb based distro. But anything new in the last ~ two years debian considers "bleeding edge".
Quoting: dubigrasuThey also said about future SteamOS being possibly image-based (not apt/etc) for better control and consistency, so whatever distro will be based on won't likely be a plain Arch/Debian/whatever with an "SteamOS" sticker slapped on it.
Good. With all the world to homed, there is no reason that the entire OS can't or shouldn't be mounted as "read-only".
As for the Debian gang being butthurt about their Valve previously noting Arch as an investigative option -- lol. That's what happens when Canonical/Ubuntu decides to kill 32-bit support. I mean you guys don't EVERY game developer on Steam still has the source code or funding to port to 64-bit like the FOSS community which doesn't run on funds but passion.
Checkout which OS is #2
https://www.gamingonlinux.com/users/statistics
Quoting: Spanner_ManBeing Debian based by itself wasn't a problem for SteamOS (well, before they gave up on it), they periodically provided new kernels and drivers.Quoting: WorMzyReally? Debian is the gold standard for stability. The problem is that people want a stable OS with a "bleeding edge" graphics stack, and that's where the house of cards falls over.
Arch can provide the bleeding edge, but it requires that YOU (the end user) know what you're doing.
Considering "bleeding edge" is a 16xx series nVidia which is around the $350AUD price range or needing to compile mesa (for newer AMD gfx) to have working on debian based distros is a pain in the you know where compared with Arch there is already an AUR ready to use.
Remember in order to appeal to current gamers you need to support newer hardware to those that are getting sick to death dealing with Win10 issues. Debian based distros unfortunately fall too short of that mark.
Look don't get me wrong, I have a laptop that has a GTX 960m and it more or less works out of the box with any deb based distro. But anything new in the last ~ two years debian considers "bleeding edge".
Quoting: Guestbut it is better than freezing the whole software base for 2 years just to make sure lazy devs can write their code and work without them having to modify it in the future.
i think you didnt understand the issue here.
its not about being an lazy developer, but having to chose between add an new feature and rewrite your program to fix issues caused by thirdy party code.
i dont give a fuck what python version my distro is using, if none of the programs that i run are running the lastest version of python, they dont need the new features to do what they do, they should not upgrade.
but i do give a fuck when i try to run an effect on gimp, and it simply dont work, breaking my workflow from:
"now, i need to aply the effect x, then y, they z, to make the photo montage i need to do for my work"
to: now i need to waste 3 hours figuring out why the effect X dont work anymore, 2 hours fixing it, 3 hours fixing the z effect that broke after i fixed the x effect, and... what are the steps i was going to follow to make my work anyway? i cant even remember it anymore!
that is why we need to keep things working for 2 years, imagine rewrite the entire blender every 2 years, do you REALLY want to put that work upon yourselft?
i'm not saying you gonna need rewrite everything, but good luck figuring out what broke (testing everything to make sure nothing broke) and why it broke (when its not even your fault, but the fault of some thirdy partie)
not to mention games, windows has at least 10 years of backward compatibility and 90% of the marketshare.
imagine convincing developers to port every library that an game rely on to linux, then port the game itself, only to break stuff 2 years later, with our current marketshare?
or imagine convincing players to game on linux, increassing our marketshare, only to then they figure out that they cant play anymore any game that launched 2 years ago, because those game developers didnt bother to update it?
no one will spend 200 millions of dolars to make an game, than rewrite it every 2 years to make sure its still working, they dont do that on windows, years later most of the profit was already made and any unexpected maintaince may not pay itself.
if developers had to rewrite the games every few years to make sure its still working, they wouldnt make games as amibitous as they do, they would have to scale down their games budget to deal with maintaince cost, the entire industry would evolve slower.
and why? so we can play on linux? if windows were like this, or linux where the most used on desktop and worked like this, those games would be console exclusives already, we would lost completely the most open platform to play games (pc) just because we want an open operating system to run on that platform.
Quoting: elmapulQuoting: Guestbut it is better than freezing the whole software base for 2 years just to make sure lazy devs can write their code and work without them having to modify it in the future.
i think you didnt understand the issue here.
its not about being an lazy developer, but having to chose between add an new feature and rewrite your program to fix issues caused by thirdy party code.
i dont give a fuck what python version my distro is using, if none of the programs that i run are running the lastest version of python, they dont need the new features to do what they do, they should not upgrade.
but i do give a fuck when i try to run an effect on gimp, and it simply dont work, breaking my workflow from:
"now, i need to aply the effect x, then y, they z, to make the photo montage i need to do for my work"
to: now i need to waste 3 hours figuring out why the effect X dont work anymore, 2 hours fixing it, 3 hours fixing the z effect that broke after i fixed the x effect, and... what are the steps i was going to follow to make my work anyway? i cant even remember it anymore!
that is why we need to keep things working for 2 years, imagine rewrite the entire blender every 2 years, do you REALLY want to put that work upon yourselft?
i'm not saying you gonna need rewrite everything, but good luck figuring out what broke (testing everything to make sure nothing broke) and why it broke (when its not even your fault, but the fault of some thirdy partie)
not to mention games, windows has at least 10 years of backward compatibility and 90% of the marketshare.
imagine convincing developers to port every library that an game rely on to linux, then port the game itself, only to break stuff 2 years later, with our current marketshare?
or imagine convincing players to game on linux, increassing our marketshare, only to then they figure out that they cant play anymore any game that launched 2 years ago, because those game developers didnt bother to update it?
no one will spend 200 millions of dolars to make an game, than rewrite it every 2 years to make sure its still working, they dont do that on windows, years later most of the profit was already made and any unexpected maintaince may not pay itself.
if developers had to rewrite the games every few years to make sure its still working, they wouldnt make games as amibitous as they do, they would have to scale down their games budget to deal with maintaince cost, the entire industry would evolve slower.
and why? so we can play on linux? if windows were like this, or linux where the most used on desktop and worked like this, those games would be console exclusives already, we would lost completely the most open platform to play games (pc) just because we want an open operating system to run on that platform.
Exactly! And even on Windows with their "10 years of backward compatibility" enterprises still spend months of testing their applications on new releases of Windows 10 before they decide to support that new release or not (which e.g is why so many enterprises where still on Windows 7 for so many years).
On the side of the enterprises, it's understandable that they want something that just works (since these programs are not open, they can't get fixes "for free" even for the trivial bugs), and for that goal i think debian or ubuntu works most similarly to windows (updates). That's why most games target these distros.
I fully agree with the commenters that say the linux compatibility is up to the companies too. I too have many games from 2010-2015 that run just fine almost 10 years later, needing no library downgrades (or manual compilation). I think most people here know that which devs or porters make those games by now.
More to the topic of SteamOS i think this just means they will roll out a new release sometime. The work on the VR stuff and their usual work on the drivers probably merit that too.
Last edited by dvd on 26 March 2020 at 8:41 am UTC
See more from me