Valve announces SteamOS a Linux based open source gaming OS

[Mr Bean]

Steam

Thousands of games, millions of users. Everything you love about Steam.
Available soon as a free operating system designed for the TV and the living room.
Steam is coming to a new operating system

As we’ve been working on bringing Steam to the living room, we’ve come to the conclusion that the
environment best suited to delivering value to customers is an operating system built around Steam itself.
SteamOS combines the rock-solid architecture of Linux with a gaming experience built for the big screen.
It will be available soon as a free stand-alone operating system for living room machines.
Living room & Steam

Finally, you don’t have to give up your favorite games, your online friends, and all the Steam features you love just to play on the big screen. SteamOS, running on any living room machine, will provide access to the best games and user-generated content available.
Fast forward

In SteamOS, we have achieved significant performance increases in graphics processing, and we’re now targeting audio performance and reductions in input latency at the operating system level. Game developers are already taking advantage of these gains as they target SteamOS for their new releases.
Cooperating system

Steam is not a one-way content broadcast channel, it’s a collaborative many-to-many entertainment platform, in which each participant is a multiplier of the experience for everyone else. With SteamOS, “openness” means that the hardware industry can iterate in the living room at a much faster pace than they’ve been able to. Content creators can connect directly to their customers. Users can alter or replace any part of the software or hardware they want. Gamers are empowered to join in the creation of the games they love. SteamOS will continue to evolve, but will remain an environment designed to foster these kinds of innovation.
Four new Steam features focused on the living room.

Available soon in both SteamOS and the Steam client.
In-home Streaming

You can play all your Windows and Mac games on your SteamOS machine, too. Just turn on your existing computer and run Steam as you always have - then your SteamOS machine can stream those games over your home network straight to your TV!
Music, TV, Movies

We’re working with many of the media services you know and love. Soon we will begin bringing them online, allowing you to access your favorite music and video with Steam and SteamOS.
Family Sharing

In the past, sharing Steam games with your family members was hard. Now you can share the games you love with the people you love. Family Sharing allows you to take turns playing one another’s games while earning your own Steam achievements and saving your individual game progress to the Steam cloud.
Family Options

The living-room is family territory. That’s great, but you don’t want to see your parents’ games in your library. Soon, families will have more control over what titles get seen by whom, and more features to allow everyone in the house to get the most out of their Steam libraries.

Well color me interested to see what they do with this.

[Zaune]

Huh. So they really are moving into the console market, in a way. Going for a Linux-based OS should be interesting as well; might take a bite out of Microsoft's market share if enough big names get onboard with it.

[Arthur_Tuxedo]

The problem with the SteamBox concept is lack of standardization. The bundled controller (assuming there is one) will need to be configured in many games, which kills the easy "plug and play" that MS and Sony's consoles enjoy. Many gamepad-enabled PC games assume that you also have a keyboard / mouse when it comes to joining multiplayer matches and similar functions, which means users will need those inputs handy. Windows games can only be supported through emulation, which introduces potential headaches. Add all of these together and most Steam-users will probably decide it's better to just stick with their current PC setup for Steam games and possibly get an Xbox One or PS4 for the living room.

The idea is cute, but unless they have some stunningly creative solutions to these issues, the entire concept is a non-starter.

[Zaune]

The controller thing I grant you (though configuring one takes what, five minutes per game these days?) but emulation might not be a problem; going by a quick look at the store, they've got three hundred games that are already offered on Linux and another eight hundred that'll work on a Mac, and porting from Mac OS to Linux ought to be fairly painless compared to porting from Windows to Mac OS.

[Hawkwings]

If you are a developer with limited time and funding, which platform are you going to develop for first?

Windows.

As this is such an obvious deduction, I am interested as to what Valve's plan for this is.

[atg]

As this is such an obvious deduction, I am interested as to what Valve's plan for this is.

This to start with:

You can play all your Windows and Mac games on your SteamOS machine, too. Just turn on your existing computer and run Steam as you always have - then your SteamOS machine can stream those games over your home network straight to your TV!

I'd like to see them try to have a compatibility layer like WINE to handle Windows games natively (with the obvious may-or-may-not work disclaimer). There are some 1 man or open source game front ends for emulators that can tell when a particular game needs particular settings for a specific emulator, no reason why Valve could not do the same with WINE.

[InsaneTD]

To me, it sounds like it will just run the games on the computer and then relay the inputs you give and the gui between the systems. But I don't know why anyone would bother. Most LCD TVs have computer inputs now any. Just plug your computer into your tv and use wireless mouse and keyboard.

[Zaune]

To me, it sounds like it will just run the games on the computer and then relay the inputs you give and the gui between the systems. But I don't know why anyone would bother. Most LCD TVs have computer inputs now any. Just plug your computer into your tv and use wireless mouse and keyboard.

There are some games that are suited to being played with a gamepad while sitting on the sofa and others for which you really need a mouse and keyboard. If it's offering to act as a remote access program that doesn't pitch a fit when you run a DirectX game then that's useful all by itself. If it can do all that and run on fairly low-spec hardware then I'm sold already.

Heh. I wsonder if it'll run on a Raspberry Pi?

[bilateralrope]

If you are a developer with limited time and funding, which platform are you going to develop for first?

Windows.

Wouldn't one of the consoles be a better first choice than Windows ?
When I think about titles that don't have a simultaneous release across platforms, the Windows release is usually the one that happens later.

I doubt Valve is trying to make Linux the first choice for a developer. Instead, their first goal may be to just get enough people using Linux that the big titles get a Linux port, or at least official compatibility with WINE, before the game goes on a 75% off sale.

How much of Valve's revenue comes from the people who wait for the 75% off sales before they buy a game ?

[General Zod]

If you are a developer with limited time and funding, which platform are you going to develop for first?

Windows.

Wouldn't one of the consoles be a better first choice than Windows ?

Consoles offer horrible support and they have a reputation of giving indie developers the shaft. Not to mention all the costs and added programming difficulties associated with creating a game for a console and the simple answer is "no".

[Irbis]

The idea is cute, but unless they have some stunningly creative solutions to these issues, the entire concept is a non-starter.

The only way I see anything they promised working is making "linux" closed system similar to Android or Mac OS. Both of them use a lot of linux code, after all.

Except, Valve is tiny compared to Google/Apple, and the sums both companies throw at developing of their software each year are close to Valve's entire worth.

I'd like to see them try to have a compatibility layer like WINE to handle Windows games natively (with the obvious may-or-may-not work disclaimer). There are some 1 man or open source game front ends for emulators that can tell when a particular game needs particular settings for a specific emulator, no reason why Valve could not do the same with WINE.

We're talking about games. They run terribly at emulators.

Not only linux has much worse drivers for high end cards even without emulation penalty, Windows has Direct X to make coding of games easier and faster. Valve would need to replicate it first for their WINE for games to work anywhere near as good as on Windows, and again, resources Microsoft pours into their software are far beyond Valve's capabilities.

To put things in perspective, Valve's profits are high double digit millions. The big 3 above? Low to mid double digit billions. No, I can't see it working.

[TronPaul]

The idea is cute, but unless they have some stunningly creative solutions to these issues, the entire concept is a non-starter.

The only way I see anything they promised working is making "linux" closed system similar to Android or Mac OS. Both of them use a lot of linux code, after all.

Except, Valve is tiny compared to Google/Apple, and the sums both companies throw at developing of their software each year are close to Valve's entire worth.

I'd like to see them try to have a compatibility layer like WINE to handle Windows games natively (with the obvious may-or-may-not work disclaimer). There are some 1 man or open source game front ends for emulators that can tell when a particular game needs particular settings for a specific emulator, no reason why Valve could not do the same with WINE.

We're talking about games. They run terribly at emulators.

Not only linux has much worse drivers for high end cards even without emulation penalty, Windows has Direct X to make coding of games easier and faster. Valve would need to replicate it first for their WINE for games to work anywhere near as good as on Windows, and again, resources Microsoft pours into their software are far beyond Valve's capabilities.

To put things in perspective, Valve's profits are high double digit millions. The big 3 above? Low to mid double digit billions. No, I can't see it working.

A. Mac uses a Unix like OS. They don't use Linux at all. There is a difference

B. Linux and Windows have OpenGL, an open source project, which is similar to DirectX. Valve has tested using OpenGL and has said it is faster than DirectX. There are other games that use OpenGL over DirectX. I'm not sure if one can be definitively said to be better or not.

Making a Linux OS is fairly easy since there are so many open source projects you can leverage to flesh out the OS. Really all Valve need to do is write a better Big Picture app that does more than it currently does in Steam and they'll have an "OS" like running XBMC on a raspberryPi is an "OS".

The real challenge is having a buy in from publishers and consumers. Not in building the thing.

[atg]

We're talking about games. They run terribly at emulators.

Not only linux has much worse drivers for high end cards even without emulation penalty, Windows has Direct X to make coding of games easier and faster. Valve would need to replicate it first for their WINE for games to work anywhere near as good as on Windows, and again, resources Microsoft pours into their software are far beyond Valve's capabilities.

To put things in perspective, Valve's profits are high double digit millions. The big 3 above? Low to mid double digit billions. No, I can't see it working.

Ignoring the WINE is not an emulator thing... People have been getting games to work under WINE for years - some games have issues, some don't work, but it would give people an option if they don't have a good gaming machine in the background to stream the Windows games. Valve wouldn't have to do much more than add tweaks - the base is already there and proven.

As for DirectX... OpenGL anyone? You know the thing that provides the graphics API for games on Linux/OSX/iOS/many windows games (including the Source engine that Valve use)/Wii/WiiU/PS3/PS4...

Anyway, the hope would be for publishers to write games native for Linux with WINE as an option for allowing access to the Windows back-catalog where publishers/devs will be disinclined to re-write for Linux.

[bilateralrope]

The idea is cute, but unless they have some stunningly creative solutions to these issues, the entire concept is a non-starter.

The only way I see anything they promised working is making "linux" closed system similar to Android or Mac OS. Both of them use a lot of linux code, after all.

Mac OS was based on BSD, not Linux. The BSD license allows someone to take the code and put it under whichever license you wish, which is how it gets closed off. Linux is under the GPL, which requires that all programs derived from GPL code are also put under the GPL. So it can't be closed off*.

*Unless you manage to get permission from every single person who owns the copyright on the code you use. Which isn't practical.

[Irbis]

A. Mac uses a Unix like OS. They don't use Linux at all. There is a difference

I said it uses a lot of common code, not that is the same thing. And besides, the point was, linux is too fragmented to make game coding and testing on it viable, IMHO. Even Android has base fragmentation problems, and that with huge effort from Google to keep it all together. You need single responsible company-product with fast issue fixing merely as start.

B. Linux and Windows have OpenGL, an open source project, which is similar to DirectX. Valve has tested using OpenGL and has said it is faster than DirectX. There are other games that use OpenGL over DirectX. I'm not sure if one can be definitively said to be better or not.

OpenGL might or might not have advantages in some situations, but Direct X does far more (not only graphical side of things, also sound and a lot more besides) and anyway, for a game to use OpenGL in comparable way instead you need to do extensive recoding.

Making a Linux OS is fairly easy since there are so many open source projects you can leverage to flesh out the OS. Really all Valve need to do is write a better Big Picture app that does more than it currently does in Steam and they'll have an "OS" like running XBMC on a raspberryPi is an "OS".

Making linux fork is easy, making anyone want to use it or develop for it is much harder.

Look at this, grand total of 0.7% of Steam users even installed, much less played anything on linux Steam client. And a lot of this number have to be Team Fortress 2 players trying to run game (once) on linux client for Tux promotional item.

Ignoring the WINE is not an emulator thing... People have been getting games to work under WINE for years - some games have issues, some don't work, but it would give people an option if they don't have a good gaming machine in the background to stream the Windows games.

It's not strictly an emulator, but it still tries to translate Windows code to linux one on the fly. That's what I meant by producing whole linux Direct X equivalent, it's a must for even trying to run modern game and so far results had been poor.

As far as getting games to work, I was under impression they have mixed success with games using Direct X 2-3 generations back, that is, now using Direct X 8, and even only some of these run, do it slowly and with problems. Did I miss big breakthrough recently?

Anyway, the hope would be for publishers to write games native for Linux with WINE as an option for allowing access to the Windows back-catalog where publishers/devs will be disinclined to re-write for Linux.

Writing games for 0.7% of users while most game development studios do it in Direct X, not OpenGL - someone would need to pay to cover the development cost. Who? Valve? That would piss off their Windows users, as most of the money would come out of their own pocket.

I have a feeling Gabe Newell is taking the whole "sky is falling, Microsoft is opening their own app market, how dare they!" way too far.

[atg]

I'm curious, do you have any figures on directx v.s. opengl?

Because the way I see it is that:
Wii/WiiU/PS3/PS4/iOS/Android/Linux/OSX support opengl.
Xbox 360/XboxOne supports DirectX.
Windows supports both.

I dont see it being that big an issue for devs to write for linux (especially if this SteamOS becomes the default 'game' distro). Many indie game devs (including many 1-2 man shops that I know or have met personally and have had successfull published games) are able to code for Windows/Linux/OSX on their extremely limited resources, are you saying a big game dev cannot do the same?

[bilateralrope]

Not only linux has much worse drivers for high end cards even without emulation penalty,

Speaking of drivers:

Nvidia seeks peace with Linux, pledges help on open source driver (Updated)
Torvalds isn't yet ready to apologize for giving Nvidia the finger.

by Jon Brodkin - Sept 25 2013, 4:24am NZST

Open Source

130
Will Nvidia give Linus a reason to lower his finger?
aaltouniversityace

Editor's note: See the update at the end of this article for Linus Torvalds' response to Nvidia.

Few companies have been the target of as much criticism in the Linux community as Nvidia. Linus Torvalds himself last year called Nvidia the "single worst company" Linux developers have ever worked with, giving the company his middle finger in a public talk.

Nvidia is now trying to get on Linux developers' good side. Yesterday, Nvidia's Andy Ritger e-mailed developers of Nouveau, an open source driver for Nvidia cards that is built by reverse engineering Nvidia's proprietary drivers. Ritger wrote that "NVIDIA is releasing public documentation on certain aspects of our GPUs, with the intent to address areas that impact the out-of-the-box usability of NVIDIA GPUs with Nouveau. We intend to provide more documentation over time, and guidance in additional areas as we are able."

The first step was releasing documentation of the Device Control Block (DCB) layout in Nvidia's VBIOS, describing the board's topology and display connectors. Ritger continued:

I suspect much of the information in that document is not news for the Nouveau community, but hopefully it will be helpful to confirm your understanding or flesh out the implementation of a few unhandled cases.

A few of us who work on NVIDIA's proprietary Linux GPU driver will pay attention to nouveau at lists.freedesktop.org and try to chime in when we can.

If there are specific areas of documentation that would most help you, that feedback would help NVIDIA prioritize our documentation efforts.

If you have specific questions for NVIDIA, you can ask here, or direct them to: open-gpu-doc at nvidia.com. I can't promise we'll be able to answer everything, but we'll provide best-effort in areas where we are able.

The gesture was well-received. In response, Maarten Lankhorst of Canonical wrote, "You rock!" Lankhorst added, "Our biggest struggle at the moment is the video clocking and power management, which is highly device specific and depends on configuration too. A complete video bios documentation would be nice too, I understand that will take a bit longer than just the dcb."

I popped into the Nouveau IRC channel this morning to get the developers' take on the significance of the news. They confirmed that they already knew much of the information in the DCB documentation but said that it may "help us find a few corner cases for cards we don't own" and allow them "to handle uncommon cases we haven't seen in the wild."

One developer said, "The stuff Nvidia wants to offer in the future is much more interesting, really. That does sound like a good start, though."

Nvidia's move this week (on the same day that Valve announced a new Linux-based operating system for gaming) is a change of heart from the position it took last year after Torvalds' criticism. At the time, Nvidia told Tom's Hardware, "While we understand that some people would prefer us to provide detailed documentation on all of our GPU internals or be more active in Linux kernel community development discussions, we have made a decision to support Linux on our GPUs by leveraging Nvidia common code rather than the Linux common infrastructure. While this may not please everyone, it does allow us to provide the most consistent GPU experience to our customers, regardless of platform or operating system."

We've e-mailed Ritger, Torvalds, and Linux kernel maintainer Greg Kroah-Hartman to get more information and reactions, and we will update this post if we hear back from them.

UPDATE: Torvalds has responded to Ars, saying he's optimistic but not quite ready to apologize to Nvidia. "We'll see," Torvalds wrote in an e-mail. "I'm cautiously optimistic that this is a real shift in how Nvidia perceives Linux. The actual docs released so far are fairly limited, and in themselves they wouldn't be a big thing, but if Nvidia really does follow up and start opening up more, that would certainly be great.

"They've already been much better in the ARM SoC space than they were on the more traditional GPU side, and I really hope that some day I can just apologize for ever giving them the finger."

Kroah-Hartman responded as well, saying, "It's very significant, and very nice to see happen."

UPDATE 2: Ritger responded to our questions regarding what Nvidia plans to do next. He wrote that "more BIOS-related information is in the pipeline," and that "Our goal is for the Nouveau driver to give NVIDIA users a reasonable out-of-the-box experience. This entails things like successful GPU initialization, display configuration, and basic 2D and 3D rendering. The DCB and other BIOS-related information will hopefully help improve some scenarios where Nouveau had initialization problems, or display enumeration sorts of challenges."

What information Nvidia releases will be "largely based on feedback from the Nouveau community," he also wrote. "So far, feedback from Nouveau guys has been that the DCB spec is actually more useful than I expected it to be, which is great news."

The timing of this announcement suggests that Nvidia believe that Valve might be successful here.

[Irbis]

Regarding latest announcement - meh. Valve, like I predicted, is too cheap to even produce reference version of their SteamBox (like Google does with Nexus) and promise that they maybe will produce a few units for testing in close future, in their famous Valve time

Come on, if there is one thing Steve Jobs did well, it was creating hype for his new product. Why this announcement wasn't paired with video of Gabe Nevell "playing" some custom-made video posing as game with elegant box standing next to Quad-HD TV?

I'm curious, do you have any figures on directx v.s. opengl?

Because the way I see it is that:
Wii/WiiU/PS3/PS4/iOS/Android/Linux/OSX support opengl.
Xbox 360/XboxOne supports DirectX.
Windows supports both.

I wouldn't exactly count consoles on OpenGL side - their code is so different from PC one it makes little difference what one component is, you need different skills anyway. As for OS X, even very powerful Apple couldn't make gaming on their computers a thing. Almost every time I read about Mac versions of games, there are complains how slow and glitchy they are. The only success they had are small games-apps on their mobile devices, I just can't see Valve being more successful.

I dont see it being that big an issue for devs to write for linux (especially if this SteamOS becomes the default 'game' distro). Many indie game devs (including many 1-2 man shops that I know or have met personally and have had successfull published games) are able to code for Windows/Linux/OSX on their extremely limited resources, are you saying a big game dev cannot do the same?

Question of scale. If big developer already has a team best at doing things in Direct X, retraining/replacing it all is going to be much more painful process than changing a pair of guys. Plus, as I said, Direct X does more than OpenGL so instead of one team of programmers you now need two, one graphical and other doing rest of things Direct X one did.

Then, there is the fact a lot of big developers have close contacts with Microsoft by way of various deals to make games on Xbox, and would write game in Direct X even if OpenGL was exactly as good. Making 2 versions of game, PC/Xbox is faster and cheaper due to common code, if SteamOS becomes a thing big games might become more expensive because now you will need to code 2 very different versions.

The timing of this announcement suggests that Nvidia believe that Valve might be successful here.

So, to sum things up, linux Nvidia drivers sucked, now they will suck a bit less, but linux team already deciphered most of this and what they need are things hidden deeper Nvidia would rather not share.

And anyway, even if linux guys had same documentation as Nvidia team writing drivers for Windows, I think dedicated team with a lot of experience with their product would still do better job than one doing job on the side.

[Irbis]

So, third announcement: Steam Controller. Interesting, and more innovative than the other two, but to be honest, PC saw thousands of interesting controller ideas that failed to gain any traction.

[General Zod]

So, third announcement: Steam Controller. Interesting, and more innovative than the other two, but to be honest, PC saw thousands of interesting controller ideas that failed to gain any traction.

What's most interesting about this is the simplicity and the ability to work with all sorts of games. How many of those other controllers were either too specialized or had poor build quality?

[Irbis]

What's most interesting about this is the simplicity and the ability to work with all sorts of games. How many of those other controllers were either too specialized or had poor build quality?

Upon thinking a bit, there are two very bad features though:

A) Thumb is your strongest finger. This thing has no thumb button group, unlike all other gamepads with 4-6, meaning it will be very weak in games that pads were strongest in, like fighting games or platformers. The four buttons it has are awkwardly behind touchpads, meaning you can't touch one if you want to press a button, or both if you want to press AX or BY combination common on gamepads. No moving while trying to kick in Mortal Kombat or trying to jump in Mario? Say what?

B) No D-pad. Again, heavy duty, quick and responsive steering. I just don't see touchpads doing it, there is a reason even gamepads with 2 analog sticks keep the d-pads.

One good feature is buttons at the back, but to me it feels like poor man's mouse and keyboard rolled into gamepad shape, not proper gamepad.

[Zaune]

Minor necro: iBuyPower show off a prototype SteamOS-based gaming machine. I'm reserving judgement until we know more about the hardware, but it certainly looks cool.

[Irbis]

"Prototype" is really a bit much to say - they're just PCs. What it is to prototype? It's just a box. Like they always were. Eh, as much as I hate Apple, their 'trash can' PC is interesting idea with unique air flow, while all parts of Steam Box offering outside of controller look from past decade or two so far.

Also, IMHO extremely dubious claim: "hardware will run all Steam titles in 1080p resolution at 60fps". Steam OS titles? Because 'all Steam titles' would require it to be a machine with four digit price, first one not being '1'.

[InsaneTD]

I hope that isn't the final colour.