Valve's Steam Box is becoming a (virtual?) reality in 2013

[STRZ]

Valve has stayed mostly quiet about its plans to enter the hardware business, but in an interview with Kotaku at last night's Video Game Awards, Gabe Newell confirmed the company's plans to sell its own living room PC that could compete with next-generation consoles from Sony and Microsoft. The biggest revelation is that Valve seems set to release its own complete hardware and software solution. When we first reported that the company was working on a "Steam Box" back in March, it appeared that Valve was working on a prototype that would establish a baseline for hardware manufacturers, but it wasn't clear if the company would sell its own product or simply release the designs to others. Newell's comments to Kotaku provide a much clearer picture of what's happening; Newell said that he expects companies to start selling PCs designed for the living room next year — which Kotaku says could have Steam preloaded — and that Valve will create its own distinct package.

http://www.theverge.com/2012/12/8/37443 ... -steam-box

This really blew my mind today, establishing a living room console as baseline hardware requirement for game developers with their own hardware constellation, and let the enthusiasts build their own Linux based consoles with the hardware they want, pushing Linux as a whole as gaming OS. And this is the company wich endorsed the Rift back then on Kickstarter! A VR Enthusiast friendly no nonsense company sets the foundation for a more open and enthusiast friendly gaming environment. This means a lot for VR gaming in the future because Linux, due to it's openess, is potentially a way better playground! I'm thrilled

[cybereality]

This is great news but I don't see anywhere that they will be using Linux or supporting VR. Chances are it will be a Windows box since that's what the bulk of games are supporting.

[EdZ]

My worry is that a static low-cost (because if it were expensive there would be little reason to buy it over a regular PC) platform will not become a performance baseline, but a target ceiling. This is already the case with many multi-platform games: many are designed and optimised to look as good as they can within the confines of PS3/360 hardware, with the PC port being the same engine code with some of the performance/quality tradeoff optimisations turned off (hence the string of PC ports that look average but perform awfully), as opposed to designing things to look and perform as well as possible on high-end hardware with things lopped off and pared down for less powerful systems. Dark Souls is an example of the former method of design taken to the extreme.

[STRZ]

@ Cyberreality

VR may be not initially, but that it will be a Linux box is sure. The Steam beta is running very well on Linux already, thanks to all the work they did in the last months. With their plan to move away from Microsoft and it's Win8 ecosystem and appstore there is no doubt about it that their box will be running Linux.

They have a lot more possibilities than on other platforms, one of them is branding their own operating system and booting straight into Steam. They don't have to pay license fees to other companys if they create a own controller. Every company which want's to sell a force feedback controller/racing wheel for example, has to pay a lot to a third party company, immersion tech. As this company doesn't develop their technology any further (for almost 14 years now), you're stuck with a toy like force feedback featureset wich doesn't cut it if you want real immersion. It's ok for arcade games, but if it comes to simulations it sucks bigtime, you need a massive amount of tweaking as developer to make it usable in a sim. And that is only one example. Windows is toy centric.

These barriers don't exist on a open source operating system, the developers can take direct influence on technology without depending of somebody with other goals in mind.

Windows is supported by Gamers, you're right, but why? Because there was no alternative. The Gamers will go with whatever runs games best, the Valve people know that.

http://www.phoronix.com/scan.php?page=n ... px=MTI0Njk


@ Edz

I don't think that they will depend financially on the hardware itself, a middle class gaming system in a WAF friendly HTPC case isn't so expensive. For 300$ you can build a decent gaming rig nowadays. If they use something like this as their reference system, it still would be 3 x more powerful compared to what you can expect from the upcoming next-gen consoles.
Developers could just use that reference platform like they would adapt to the hardware specs of a console. And the real enthusiasts can build their own with the hardware they want.

To compete with consoles and their comfort you only need a system wich is easy to use, plugging in, maybe auto updating and booting straight into the game center. This is possible with Linux as you can strip away what you don't need, and you can create your own update system, unlike to microsoft where you get only system updates for the Windows base stuff.

[cybereality]

Well if anyone can make Linux a real gaming platform, its Valve. I just wonder about inertia in the dev community. I mean, will all these devs have the motivation to go back into their catalog and port their old games to OpenGL/Linux. That would require significant effort. I can understand maybe for games in development (or just beginning pre-production) it may be easier to make the switch. I just don't see a huge financial benefit for doing this on older titles. Especially ones that sold well already on consoles and Windows, so most of the people that wanted the game already played it. But as a new platform, maybe it could do well.

Honestly the main reason I've stuck with Windows all this time is because of the gaming mostly (and peripheral support). If devs started moving to Linux, I'd be happy to make the switch. Especially now since Windows 8 looks like a bust. But there is a lot of inertia to overcome, its not going to happen overnight.

[EdZ]

I don't think that they will depend financially on the hardware itself, a middle class gaming system in a WAF friendly HTPC case isn't so expensive. For 300$ you can build a decent gaming rig nowadays.

'Nowadays' is the important word there. When the 360 and PS3 were released, the hardware in them was comparable to the average non-high-end gaming PCs of the time. Technology has marched on, and consoles have to keep the the same hardware to prevent platform fragmentation. This can be eked out by optimising to precisely known hardware behaviour, but this can only go so far before you start hitting fundamental hardware limits (bus speed, memory capacity, fill rates, etc).

With a tiny HTPC-style gaming PConsole, you have a trade-off: make it a cheap, all-in-one BGA-soldered-to-mainboard affair that keeps the price and size down and provides a known platform to build on. Or make it larger, bulkier, more expensive, and upgradeable (with standard parts, or proprietary, and unpopular, upgrade modules) and lose any sort of platform stability.

[STRZ]

It depends on the game engine of those titles. Unity 4 for example is backwards compatible and has a sort of one click Linux support, and a lot of games are made with this engine in the past. Developers load their game into the engine and have it relatively easy to make a Linux port. Some big game engines are rumoured to support Linux in the near future like Cryngine 3 (due to the demand of StarCitizen Backers), Unreal engine 4, the new id Tech 5 (maybe John Carmack reads this and want to comment on this ) and of course Valves own new engine. And this was already before the Steam console was confirmed.

I mean, most of those companies like Epic, Valve and id have friendly relations anyways and in common that they have endorsed the Rift and are usually more enthusiast friendly, they fit perfectly with the more independent and upcoming crowdfundet companies wich naturally support Linux

There is a cut coming with the new consoles anyway. PS4 will be X86 based as well, instead of the exotic Cell CPU + OpenGL-ES GPU. And supporting native open GL like common Linux distros. Most def. using a more customized embeddet version of Linux as well.

Open GL could be the advantage for Valve instead of a problem because everybody uses it except Microsoft. Especially for those who are still little Tablet gaming studios and want to make something bigger without breaking the bank for usual console licenses or the need to adapt and restrict themselfes to direct x.

With a tiny HTPC-style gaming PConsole, you have a trade-off: make it a cheap, all-in-one BGA-soldered-to-mainboard affair that keeps the price and size down and provides a known platform to build on. Or make it larger, bulkier, more expensive, and upgradeable (with standard parts, or proprietary, and unpopular, upgrade modules) and lose any sort of platform stability.

Platform Stabilty is there if they use their console build like they ship it as reference system for developers. I've seen badass hardware inside those little 50$ ITX cubes, many of them have space for dual slot GPU coolers. If they use something like the AMD Vishera apu like Sony in their PS4, there are motherboards with that apu having a PCIe x16 slot as well. They ship it as a stable tested system, and if you want you just put a bigger GPU inside. I think that there would be no stability problem as long as you stay with one driver for the same hardware line of a manufacturer. Going up from an IGP to a Highend card as long as you don't switch to a newer architecture of that manufacturer.

It's even possible that they license it together with their new controller like Google does with Android, but a tighter defined hardware minimum. So that it will be up to the license partner if it's upgradeable or not.

[MSat]

I thought it was kind of a silly idea at first, but it might actually be a step Steam needs to take in order to survive in the long run. Google, Apple, Amazon and MS all want to control the distribution of software for their OS ecosystems, not to mention the de facto control console manufacturers have over theirs. Apparently, the only way that's possible now a days is to have your own OS, so it only makes sense for a company who's business is games distribution to have their own game playing/purchasing-oriented OS.

It sounds like they want to go the Android route, where numerous manufacturers are creating hardware, while they maintain the OS and make their money through Steam. Perhaps they might build the first reference design, but they're hoping other companies will jump on. I do hope they implement some new standards in terms of hardware design - as to not be constrained by the slowly evolving PC paradigm. I would love to see slide in/out components such as video cards, hard drives, motherboards (which shouldn't be much more than a CPU, ram and interconnect bus). This would allow a more rapid adoption of cutting edge tech then what we traditionally get with PCs. For instance, instead of worrying about socket compatability, CPU manufacturers could implement new technologies as they please, say such as DDR5 memory buses, etc. and just easily remove the old one, and slide in the new one. If the OS is designed right, and Valve works closely with hardware manufacturers, they can still make the whole thing just about as seamless and easy to use as any modern console.

Since it seems highly unlikely that they would be able to port their whole catalog, either they're looking to start completely fresh, or they're going to implement some sort of Windows API bridge. I'm curious what this means for the future of linux. Will they be pushing for changes to OpenGL, or will they create a more modern graphics library? I don't doubt this is all possible, and as Cybereality said, if anyone can do it, it's Valve.

[STRZ]

Since it seems highly unlikely that they would be able to port their whole catalog, either they're looking to start completely fresh, or they're going to implement some sort of Windows API bridge.

The source engine games are beeing ported, some run already in the Steam Linux Beta. The rest would be the business of the opther companies most likely. But it could be possible that they will help once their own stuff is done. If you build your own Linux box, you can use the Wine to simulate a Windows environment, works for many games, even worked years ago with Steam itself. For Android it's even easier with AndroVM including Google appstore acess, and a program to run Mac binaries was announced these days.

I'm curious what this means for the future of linux. Will they be pushing for changes to OpenGL, or will they create a more modern graphics library? I don't doubt this is all possible, and as Cybereality said, if anyone can do it, it's Valve.

They have one of the best low level x86 assembly coders worldwide in their team (Abrash), and other very talented newly aquired people for their project! They can accomplish everything they want with those guys, i think.

But it wouldn't make sense to create a "new OpenGL" if there is nothing really wrong about openGL. Further it would hurt the portability thing between future consoles.

Direct X 11 vs. OpenGL side by side comparison: http://www.youtube.com/watch?v=HC3JGG6xHN8

Please have a read http://blog.wolfire.com/2010/01/Why-you ... ot-DirectX

Btw, the game "Overgrowth" of the developer writing this blog rocks, check it out. http://www.youtube.com/user/wolfiregames

[Fredz]

They have one of the best low level x86 assembly coders worldwide in their team (Abrash), and other very talented newly aquired people for their project! They can accomplish everything they want with those guys, i think.

Michael Abrash is not involved in low level programming at Valve, he explained that there were already talented people working on that and that his help would be more appropriate elsewhere. He works exclusively on AR now.

Also I doubt anyone is still using x86 assembler in game programming now, but people at Valve seem nevertheless very talented, so they'll probably come with interesting things for their box.

[WiredEarp]

I think a new OpenGL based implementation, that is pushed by a big vendor, but maintains the cross platform abilities of OpenGL, is a great idea.
OpenGL as it is has suffered greatly over the last 10 years from DirectX competition, and I dont think it will be able to increase its share much until we wean ourselves off Microsoft, without having a big name pushing a new version of it.

[MSat]

While I don't doubt the same final results can be had in OpenGL as with DX, but it is my understanding that DX along with vendor driver's make certain aspects of graphics programming easier. That's not to say that some additional graphics libraries can't be used between the application and OGL, but as far as I'm aware, those don't exist - at least not as a cohesive package. If Valve wants to make a game-oriented OS, then this is mandatory IMO.

[STRZ]

@ MSat

Valve has Sam Latinga, the creator of the SDL wich is used with openGL, something similar to the DX environment but free and open source software.

http://www.phoronix.com/scan.php?page=a ... _sdl&num=1
http://www.phoronix.com/scan.php?page=n ... px=MTE0MDU
http://en.wikipedia.org/wiki/Simple_DirectMedia_Layer
http://en.wikipedia.org/wiki/Sam_Lantinga

It's impressive because Valve has end to end control over their stuff, they control for gaming important core technologies of the operating system, as well as they are designing cutting edge game engines and hardware control periphals on the other end. This wouldn't be possible in a Windows environment, because Microsoft dictates every aspect of the core functionality. I think this is a big reason why PC Gaming is stuck for so many years.

Another big advantage over other consoles is, that if you would develop for the Steam console, your game would appear on the Linux desktop Steam Client as well, with no additional effort because it uses the same client. And at a lower cost, because publishing your game for usual consoles is very expensive. As independent developer or one man team it's nearly impossible. I think that there will be A LOT of crowd funded games from smaller independent studios for the Steam console/linux Desktop.

Michael Abrash is not involved in low level programming at Valve, he explained that there were already talented people working on that and that his help would be more appropriate elsewhere. He works exclusively on AR now.

Allright, thanks for the clarification.