Want to see the dirty innards of more Valve code? Well you're in luck as they now have a lot of work involved in the Steam Runtime on GitLab including the Pressure Vessel container.
Valve has for some time now had their own GitHub account, which is where they listed many different open source projects like GameNetworkingSockets, Proton and more. However, they've now added a bunch of other projects to their own hosted GitLab.
You can now find the steamrt group on their GitLab, which contains projects for various parts of the Steam Linux Runtime, including the source for the much newer Pressure Vessel container system which according to Valve contractor Timothee Besset on Twitter was previously "only available as a tarball release" from their download servers.
What is Pressure Vessel? It's kinda of like a simple version of Flatpak made for Steam games. Within the Linux Steam client, you can select "Steam Linux Runtime" under the right click -> properties menu of games at the bottom like so:
This then puts those Linux game builds into a game-specific container. There's many reasons for it, like allowing developers to test against a contained environment, and have it run across any Linux distribution and allow old games to continue working long into the future. Learn more about it here, where Collabora engineer Simon McVittie gave a run-down of their work. Valve are also now using the latest generation of the Linux Steam Runtime for the Proton 5.13-1 compatibility layer too.
When we queried on why Valve are now putting more up in the open on GitLab, instead of the GitHub that was being used originally, Besset mentioned to us "The projects you see on GitHub are often mirrored from an internal repo. It's awkward and creates friction for community contributions. The projects on gitlab is where we will do our work in the open.", which is awesome.
Quoting: minidouI still don't understand what a user is supposed to make of "Steam Linux Runtime".Not so much right now. Although it is already useful for getting games like Dying Light to work properly.
The whole idea is to give a stable runtime environment to the game developers where you don't need to care about what distribution is the game run on, or that upgrades to the system would break something. It is essentially a Linux compatibility layer for Linux to address the main argument a lot of people have against Linux - fragmentation.
Hardware + SteamOS + new version of SteamController + HL:Alyx free.
It would be great. Valve has the big enough resources to Steam to be able to do this. And now, with Proton would be a dream for us see that.
Who knows.
Normally it's bundled with Steam and just kind of works from there, which is really great, but it would be better if it became a general thing that distros can incorporate. If I understand it rightly and that can be done of course.
Quoting: PalancaI think Valve want to launch with Microsoft and Sony their own console.HL:Alyx doesn't make sense to be a free bundled game for a console. Most people who would buy a console would not use it for VR.
Hardware + SteamOS + new version of SteamController + HL:Alyx free.
It would be great. Valve has the big enough resources to Steam to be able to do this. And now, with Proton would be a dream for us see that.
Who knows.
Unless of course they want to release a VR-only console but I don't think that's a good idea. Standalone headsets that don't depend on external hardware are a much better option and they are proved to be working very well by the Oculus Quest.
Sure, PCVR is still much better in most ways (not all ofc) than standalone headsets, mainly because of a PC's power, but that won't be the case for very long. And investing into a whole new console and console category when it is only going to be relevant for e.g. 5 or so years, doesn't make any sense. And even then I think people would still rather choose standalone headsets for convenience, or traditional PCVR headsets if they want a PC anyway or more power than what would basically be a prebuild PC in a small case with many compromises.
If Valve will release a console, and I hope they will, it is going to be probably a Steam Machine v2. If they get a handle on the library size (which they seem to be doing very well with lately) and the user experience (the Big Picture UI, while cool at first, is severely lacking), and the "plug-and-play"/"it just works" kind of convenience that a normal console offers, then yes a new SM could prove successful.
We've seen how exclusive games influence the sale of a console greatly, and PC has a lot of "exclusives" of its own. IMO the idea of a "console" which could run any (well not quite any, but still) PC game is very intriguing, even if you can't use it as a normal desktop PC.
Obviously a problem would still remain: price. Powerful PCs are expensive. The consoles are so cheap because they cover the costs through other means, and because they are such closed environments. But the question is... could Steam do that? Or rather... is Valve willing to sell hardware at a loss in order to increase their sales and influence, and would it even be worth it in the end?
Last edited by Hori on 30 October 2020 at 12:24 pm UTC
Quoting: LinasThe whole idea is to give a stable runtime environment to the game developers where you don't need to care about what distribution is the game run on, or that upgrades to the system would break something. It is essentially a Linux compatibility layer for Linux to address the main argument a lot of people have against Linux - fragmentation.
That's how I read it too. Which is awesome, as usual for anything Valve is working on related to Linux. Fragmentation is one of those hard problems with Linux, and this goes right to the heart of addressing it.
I don't think Valve is planning another console. Even Steam Machines weren't really consoles as such, more like prepackaged PCs with a living room tailored experience, like a media center PC.
To me the way I read everything they're doing, it feels like they've taken a step back, looked at the Linux gaming ecosystem, identified every pain point they can find that's preventing adoption and they're just systematically going through each one and trying to either fix it or at least reduce the pain as much as possible. I think they're hoping if they fix enough pain points and wait long enough, eventually organic growth will take off.
The question is why of course. Why do they care. It is possible that perhaps they're just passionately and ideologically supportive of Linux. Or the popular theories of it being a preemptive defensive strategy against any future moves Microsoft might make. Either way, at this point they're putting so much effort into this, it's clearly one of their main goals now, not just a side interest, they're 100% committed to this.
Will it work? Honestly I think it will. But it's definitely going to take a long time.
The analogy I'd use is ... imagine that Linux gaming is like a 100m diameter ball of lead, on a flat plateau of land on top of a hill.
We're been trying to get the ball rolling. Once it does build up speed and start going down the hill, it will become an unstoppable force. We've been trying to get it rolling for years, but it's a 100m diameter ball of lead and we're just a dozen people pushing at the side, having no impact. The stubborn bastard wouldn't shift.
Valve's efforts with Steam, Proton, DXVK, ACO, Pressure Vessel, etc, is like bringing 4 giant trucks up to the top of the hill to help us out, they've strapped them up to the ball and are gunning the engines with the tires screeching. The ball is shifting a little, starting to move, a few inches at least, it's working but it's slow. If they keep at it for the long haul and the tires don't pop, eventually it will build up momentum and it'll work.
If it works, Valve is best positioned to benefit from the success. A new mainstream gaming platform will emerge and Valve will be the centre of it's universe.
Last edited by gradyvuckovic on 30 October 2020 at 1:01 pm UTC
Quoting: gradyvuckovicThe analogy I'd use is ... imagine that Linux gaming is like a 100m diameter ball of lead, on a flat plateau of land on top of a hill.
I dig this analogy alot!

Quoting: Perkeleen_VittupääQuoting: gradyvuckovicThe analogy I'd use is ... imagine that Linux gaming is like a 100m diameter ball of lead, on a flat plateau of land on top of a hill.
I dig this analogy alot!
Yeah. HUGE potential energy.. but just really difficult to get moving.
See more from me