Valve working on 'Steam Box' gaming console with hardware partners | Hacker News Reader

Valve working on 'Steam Box' gaming console with hardware partners

theverge.com

179 pointsKynlyn105 comments

Here is what i hope will pan out:

1. Steam OS will be powered by Linux + OpenGL

2. It will have unparalleled gamer-bias and developer-bias instead of revolving around publishers (meaning its truly gamer focused as what has been Valve's MO)

3. It will be just good enough in the graphics department but will instead lower the friction in the assets pipeline department, virality and player-analytics

4. Graphics will now take a backseat to story, AI, art-direction and pacing (due to the hardware spec being fixed and player measuring tools)

5. crowdsourcing features will be baked in thru APIs (Team Fortress hats as a precursor)

6. social multiplayer features baked in thru APIs

7. a standard controller will be mandated

8. optional accessories like bio-measuring gadgets (measure opponent's sweat, heart-rate and track gaze) will be sold online.

If any of this comes to pass then it would have made PC gaming much much more evolved.

I don't think the hardware spec will be fixed. The article seems to think it will be Android-like. Valve will give guidelines like 'must support OpenGL' and 'needs at least HDMI output and 6 USB ports', but if you want to spring for the Alienware Mega-'Console', then you'll have a bit more power.

And the more I think about this and type my response, the less exciting this sounds... All of this is possible on a PC. What they'll be offering us is basically a pre-loaded, locked-down Linux based OS.

True, but if they invest the resources into making all of these games run on Linux then perhaps those of us with "real" Linux setups will be able to run them too. No more dual booting!

Speaking of Android, I think Google should either either Android or Google TV to turn it into a "console platform" rather than the usual "console-device" we have now.

Graphics on ARM chips are getting good enough for enjoyable 3D games in HD, and soon we'll be able to just hook our phones to a TV and play all the Android games there anyway (some phones already do this).

But Google could encourage manufacturers to build special cheap $100 boxes that are meant only for gaming, too. Then it would just be able to gain market share over the Xbox 360 and PS3 in the same way it gained over the iPhone - by getting all manufacturers to make such devices.

bio measuring gadgets sounds interesting , of course it would be very easy to cheat this and heart rate monitors are not particularly comfortable.

Graphics taking a backseat to the story would be awesome. I think indie games are showing that gamers don't care too much about graphics as long as the gameplay/story are good.

I'd love that as well. But what does Valve has anything to do with it? If it's an open platform, it probably isn't a good idea for them to tell developers to do it their own way. It's up to the developers to decide that.

Indirectly thru the values of this Steam box. If most of the hits on this box are not graphic rich but very gameplay or story rich + hardware is mid-range + shiny tools for story, AI and multiplayer it will inform developers to most of their games in one way.

I second this as well. I think the spot light on eye candy due to mass appeal, for the past few years, has shifted the focus from game play it self which is what games are really about. I'd take a clever and fun game over a glossy one any day.

How do they power it with Linux, yet still support EA's Origin like the article claims they will?

This article doesn't mention a very important and very relevant point (maybe they assume everyone is aware):

Valve are also the developer and publisher of some of the most popular video game titles around. This puts them in an exceptionally good position to build something like this because with the backing of some of their own titles (For example releasing Half Life 3 as a "Steam Box" exclusive) would immediately encourage sales. Portal, Portal 2, Left 4 Dead, Half life and Team Fortress 2 are all hugely popular Valve games and their platform "Steamworks" is again used by some huge names, Call of Duty, Football Manager, the list goes on. They have their platform directly integrated with a large amount of popular games and a lot of new releases go with Steam for their PC releases.

There are some interesting Steamworks and Steam statistics and information in this PDF: http://www.steampowered.com/steamworks/SteamworksBrochure201...

Valve games currently require Steam (on PC) but can be bought anywhere. Making HL3 a Steam Box exclusive would alienate the vast majority of their fan base and kill all the good will they've built over the years.

That's true and I don't believe they ever would, I was using it as an example of the position they're in and the power they hold to make something like this work and be successful. I should have clarified.

it doesn't have to be all or nothing. Since they've got the control - some exclusive content here, financial incentive there, and people will slowly migrate to their platform.

I think they would release Half Life 3 and other games as "Steam Box 1 Minimum" which is to say, you either buy the box or your computer must meet the minimum conditions of the dedicated box. Valve is devoted to creating an experience, and it's sad that consoles are such a limiting factor in game development, even though it makes sense.

I guess this is their midstep until OnLive-esque cloud gaming is more dependable/available.

But what do consoles limit?

Apart from shiny graphics what things are games trying to do which they need more power for?

Which games are pushing at AI or different forms of man machine interfaces?

When was the last time you saw a console that allows mods to be created?

Battlefield 3 also had toned down graphics on consoles.

Well, this clinches it. The next generation of home consoles from Sony and Microsoft are going to, of necessity, run on minimally-customized, commodity hardware, with the end-user experience as sole differentiator. It has seemed like an obvious choice following the diminishing returns of Sony's adventures in exotic supercomputing architecture, but with a competitor at last combining a stationary-target PC with existing vertical integration and massive mind share and market share, they'll have no choice if they want to remain competitive. This industry is in for interesting times.

And heaven help them all if Apple brings apps to the Apple TV.

As a counter-point, consider the Xbox and the Xbox360. The former was pretty much a garden-variety AT PC in an unusual form-factor with some rudimentary security. The latter is still mostly a PC, but has been customized much further.

Certainly much will remain the same, but in my opinion this progression has illustrated that it behooves console makers to customize to a notable degree.

The 360 bears little resemblance to "a PC". Unless your "PC" is a three-core SMT'd PowerPC with on-die GPU packing stupid-fast DRAM.

(It's not.)

I, for one, will mourn the loss of such exotic supercomputing architecture. Whether it's the massive clustered PS3 supercomputers that people built because of the commodity pricing/loss that Sony was willing to take or the absurdly difficult to utilize twin Hitachi CPU powered Sega Saturn, there is something very appealing (to a big nerd like me) about the sort of exotic hardware that game consoles have used in the past. No more RISCy moves from the console builders I suppose. Ho ho.

Let's say we all move to commodity PC hardware in our consoles. At that point, one of two things will happen: the 'consoles' will have ever-shifting hardware and be complete hell to work for, or developers will code to the hardware and not go beyond, effectively stagnating PC game progress. That feels like a net negative for everyone to me.

You're assuming that the hardware will be upgradable. It won't be. As noted elsewhere, the Xbox basically was a PC.

This console cycle has been characterized by software and online, and that's how it will be from this point out. 360 has a very good OS and networking stack, PS3 not so much. This is why 360 is outselling the PS3, even though it's the weaker-powered of the two.

I don't see this changing in the next cycle either; PS Vita's software looks awful.

The Steam box will be no different; it's about the software, not the hardware.

I think what he is actually doing is taking a loose interpretation of "commodity". That is to say, if consoles are built on "commodity" hardware, that implies different consoles will come from the factory with different hardware that is only partially fungible.

I'm not sure what it is really called, this sort of "fungible but not fungible"- for example, x86 processors are fungible in the functional regard but not in the performance regard.

"massive mind share and market share"

The only thing massive about Valves mind share is maybe the preception of it among HN users.

http://www.google.com/insights/search/#q=%22valve%22%2C%22st...

There's always a tendency to wonder if any new product will be the new category killer. I think we can be confident this isn't. As the article says the Steambox isn't trying to compete with the next xbox, and for good reason -- it would lose horribly.

That doesn't mean it won't be good or I won't get one or it won't be a success. This is a good and logical move for Valve that provides good value for their existing userbase and opportunity for growth.

Way to ruin it with realism! Ultimately Steambox will still use MS Windows and therefor DirectX, so it's not like it's a huge jump from an XBox 360 anyways, right?

The original Xbox consisted of commodity hardware (Intel P3 era Celeron CPU and a GeForce 3-like GPU, if I remember correctly). Although cheap to develop and produce initially, it apparently cost Microsoft dearly later on in the lifecycle because Microsoft didn't own the CPU design, which allowed Intel to charge for more than just manufacturing cost of the CPUs. Compare this to the XBox 360, the design of which Microsoft commissioned from IBM[1]; it's subsequently been through a few die shrinks, reducing production costs. Intel would have to be pretty desperate to license their commodity CPU designs to third parties. So I can't see an Intel-powered console happening anytime soon.

Having said that, I could imagine AMD licensing a CPU or APU design to Microsoft or Sony[2]. After all, the Wii and Xbox360 already contain custom GPUs of theirs. Still, the current consoles have done well with PPC, so they may just stick with that.

[1] Sony actually paid for some of it because it's basically 3 Cell PPUs glued together, but that's another story

[2] AMD's CPUs are probably not complicated enough for Sony's liking, though. :-P

I could imagine AMD licensing a CPU or APU design to Microsoft or Sony... AMD's CPUs are probably not complicated enough for Sony's liking, though.

http://semiaccurate.com/2012/03/02/sony-playstation-4-will-b...

Hah. Glad I'm not a betting man. :-)

"with the end-user experience as sole differentiator"

Does Kinect not count as a hardware differentiator? For this reason alone MS has the be the favorite among dedicated game boxes.

Nintendo's been the most innovative historically but they've already shown their hand and sadly for them it's holding a ginormous controller that I'm 97% sure is a refurbished Sega Gamegear.

The only question left about Nintendo is who wins the exclusive rights to Mario/Link and co. It's probably worth the most to Microsoft but Apple has a lot of cash overseas they don't know what to do with.

Never underestimate Nintento. When they showed off the Wii Remote back at the TGS, nobody was stunned. Everyone was puzzled or dissapointed, the talks about who will get the rights to their IP was already going on... which was kinda stupid because Nintendo sits on one of the largest piles of liquid money in the gaming industry... they could sit out a whole console generation and still produce another one... don't count them out so fast.

Nintendo's strategy made sense though. A bunch of people didn't want to hear that the gaming paradigm might be changing, but that's besides the point. Personally, I didn't know if they were going to win or lose with their strategy, but I thought that it made sense.

Right now I'm confused about their decision to abandon the Wiimote though. If they're upset that they didn't get as many 3rd-party titles, IIRC that had more to do with the hardware constraints of the platform (I remember it being described as two GameCubes strapped together).

  > they could sit out a whole console generation and still produce
  > another one... don't count them out so fast.

Isn't that the story of Sega though? The Saturn was a flop, and then the DreamCast wasn't enough to rescue them from becoming a software-only company.

I believe the idea is that the Wiimote will continue to be used alongside the new controller in e.g. local multiplayer games. At least, that's what they had set up in some of their E3 demos last year.

Apple purchasing nintendo would be interesting. I don't think Apple has anything to offer in terms of gaming, but they know how to manufacture great hardware (and for less than competitors), and they know how to market and distribute content.

I'm not sure that apple itself has nothing to offer. I have a feeling it is only a matter of time before apple releases something that is of serious interest to gamers.

It may not be a games console as such but think something with a large screen , an app store and some new input method that lends itself well to gaming.

Apple's announcement of what will likely be the iPad 3 coincides with GDC. I would think an iPad 3 + 1080p AppleTV is a compelling console replacement. I bet that is their move: to make your TV a monitor for the iPad. That solves the interface issue with TV in one fell swoop.

I think this could work well for some categories of games, for example strategy games but there are still many popular games this wouldn't work so well for at all(driving games, fps etc)

So either they will focus on a different category of gamer/games or there will have to be some other more game specific input device such as a controller.

Few points:

(1) I don't think Apple would buy Nintendo but...

(2) If Wii U flops worldwide (as many expect) then Nintendo's default option is to go the Sega route and just produce software.

(3) Unlike Sega however, Nintendo's IP is worth substantial market share so I'd expect a major bidding war for exclusivity. Exclusivity would also allow Nintendo to continue to develop for a single mobile and home platform as they do now.

There were rumours a while ago that Valve was working on Linux support for their infamous source engine which powers most if not all their games. What if it wasn't just support for linux but Steam running on a Valve tailored Linux kernel. Obviously most games would still need Windows for DirectX but all of Valves games and many others could work on this super thin Linux kernel.

* http://www.phoronix.com/scan.php?page=news_item&px=ODE4M...

Very unlikely. They specify it will be able to run other services, e.g. EA Origin.

The Source (and Goldsrc) engines support a Linux dedicated server, which explains the job listing. As far as using Linux goes, unlikely, since people won't be able to play any of the games they already own.

Not entirely true. Valve's Source games can be easily ported, and other games can be modified slightly to run on WINE. (Similarly how all of EA's Mac games run on WINE instead of being ported)

Haven't they already ported most of their games to mac? It doesn't seem like going to Linux would be that big an effort. Valve's games make up like 10% of steam sales though so they would need to work on pressuring the other 90% of the market that they don't control. Releasing a dedicated box that doesn't work with 90% of the content you have sold doesn't sound like a winner to me.

This sounds interesting and all, but who is going to be their customers? The 30 or so millions of their users obviously already HAVE their own hardware.

So is it going to target the current console owners? And persuade them how? "It's like your box, but you can install mods. It's like a PC but with less freedom."?

Anyway I'm looking forward to the development. The three big console manufacturers need a kick.

Edit: Also Steam doesn't have their own OS obviously. So is every box going to need a Win licence?

Me. Long time Steam user. And, I'll admit up from, Valve fanboy. I'm also a Mac user, and if a Steam box existed, I'd buy it and not look back.

How many of those 30 or so million users are looking to upgrade their video card or buy a new computer? That's your target market.

potential markets:

- people looking for a computer to hook up to their television. I bet there are lots of people out their who would like a nice HTPC but want modern gaming performance as well.

- people with laptops as their primary PC but who also game. How many people have a desktop and a laptop because gaming sucks on most/all laptops? Probably the vast majority of PC gamers. Every single one of these is a potential customer. Use a MacBook Air for work & facebook & casual gaming and get a SteamBox for serious RTS gaming or flight simulators or whatever other kind of games don't work well on consoles or on laptops. You'll no longer have to worry about anti-virus or any of the other crap involved in maintaining a Windows box. You'll just have an OS X box that "just works" and you'll have a Linux box that "just works".

- people who want to buy a PC without paying a license fee to Microsoft. Some of these will install Ubuntu or a pirated Windows on their later, but then support is no longer the Vendor's problem. I suspect it will be much easier to persuade vendors to sell "SteamBox OS" boxes than Linux boxes just because it should reduce support costs.

To summarize: cheapskates and gamers. I think that covers about 99% of the home desktop PC market -- everybody else buys laptops.

I'm a Steam user and I'd probably buy one. I don't own any consoles, and I don't own a Windows machine. I'm tired of booting from OS X into Windows to play Windows-only titles. I'd much rather have an unobtrusive box in the corner that can play all Steam titles.

Everyone is going to be their customer. I imagine this being a standard feature of all TVs in the future. Apple TV is a great idea, but I don't think it will evolve to be a huge market for game developers to target. Steam Box could possibly be 'the' platform on everyone's living room. That's what I read from the article: open platform for hw/sw, standardized specs, standard controller, great content delivery network and easy payment options.. Sounds like success.

Just take a look at community sites like NeoGAF where people own Xboxes, PS3s, Wiis, Vitas, PSPs, iPhones, etc., etc.

I'm sure a lot of people don't take it for granted that they will get the next version of any of these except for the iPhone.

I don't know how many people can imagine the Xbox without Halo, and if Valve can line up some companies or their own games, I wouldn't be surprised if they disrupted the industry. There aren't that many flagship franchises for either of the consoles, and just look at how Nintendo basically don't know how to innovate in videogames anymore and end up releasing Zelda 15, Mario 8, and Mario Kart/Party/Tennis/Golf Umpteen as a result. The iPhone is destroying the handheld market, and I'm sure Activision will figure out how to develop Call of Duty for the new console, if it gains traction.

Valve will allow their customers to play their games on Mac, Windows, and their console, if any versions are available, due to the cross-platform presence of Steam - they have basically already done this with Sony. That's new and very alluring, although I have to wonder how they will manage the disk space of the console. Maybe people will have to buy the physical version of the games and use the Steam key to let them download it to their Mac and PC. The prospect of selling your games directly to your customers with no expenditures save bandwidth through Steam might help the platform earn a profit, too.

Am I the only one that sees this as a bad move? Microsoft had os development, directx, intel this is why they pivoted into the console market. I think valve should expand steam even more - bring steam to xbox, ps3, wii make buying games cheap easy and fast.

Thing is, Valve has mindshare and love amongst gamers that MS couldn't hope to match, and Sony had none of when it brought out the PS.

There will be many gamers (myself included) who will buy this simply because it's Valve and we trust them not only to make it awesome, but to bring awesome games to it as well.

Valve brought Steam (in the form of a subset of the Steamworks APIs) to PS3 for cooperative play in Portal 2. But they can't expand Steam to someone else's platform, those three you mentioned already have their respective stores.

As I see it, a "Steam Box" would merely be a hardware specification that Valve sets and game developers can rely on, ship with Steam preinstalled and supporting their previously announced "big picture mode", and the multi-input controller. But for all intents and purposes, it would be a standard desktop PC. Probably one designed to be used in the living room, but standard hardware nonetheless.

I wrote a post about this a year ago[1], and concluded that it was unlikely that either Microsoft or Apple would sell them OS licenses cheaply enough to make such a product viable.

The third alternative is Valve using Linux, or something, which isn't enormously plausible.

1: http://bbot.org/blog/archives/2011/05/08/valve_isnt_going_to...

I toured the valve offices last summer, and I was told that they were in fact developing their own variation of Linux for internal use.

Gallium does support DX11 with open ATI drivers these days...

What? No it doesn't. There was one experimental commit of the very beginnings of DX support, and it hasn't been worked on at all since.

Why is Linux not plausible?

1) There have been several leaks indicating that Steam for Linux exists, or at least has been prototyped. I imagine that pretty much the only reason why it hasn't been released is because it's a small market with large support requirements.

2) Porting games to a Linux with decent video drivers is no harder than porting to the Mac. Once you've done one, you're 95% of the way to doing the other. The largest task is porting to OpenGL which you have to do to port to PS3 or to Mac anyways.

3) The largest problem game makers have with Linux is supporting a large variety of hardware along with a large variety of distributions. If you can say "We only support SteamLinux on SteamBox", that problem completely disappears.

> "Why is Linux not plausible?"

Short answer: DirectX.

Scream all you want about OpenGL, but MS has honestly created a superior product. Between the hand-wringing over specs and the massive fragmentation via vendor-specific extensions, OpenGL doesn't hold a candle to DirectX. Even old OpenGL stalwarts like Carmack himself have bailed from that wagon.

There's a reason why Mac ports are still relatively rare in the industry, and more importantly, where they do exist, they are universally horrendous in quality and performance.

It should be noted, that while there is agreement that DirectX is a better API. Portability is becoming a much larger concern, especially with mobile gaming. These days, just about every major game engine has the capability of running on OpenGL.

Carmack has acknowledged that DirectX is better, but Rage, for instance, still supports it for systems that don't support DirectX.

Correct me if I'm wrong (written lots of OpenGL and DirectX over the years, but never worked on a console), but doesn't the PS3 use PSGL, an OpenGL offshoot that's not directly compatible with other OpenGL implementations?

Which is to say, porting effort to PS3 does not also give you a PC/Mac OpenGL port.

This is the big problem I see - UE and idTech both support OpenGL, though in reality only UE has any significant licensee base to speak of. There is, however, an awful lot of home-brewing even at the AAA-levels: EA's gone and thrown a lot of weight behind the Frostbite engine, Ubisoft seems quite fond of their Anvil engine (Assassin's Creed + more), and the big massive CoD franchise runs a proprietary engine too.

An OpenGL-only platform might automagically include all UE and idTech licensed games, but that's hardly an impressive snapshot of the gaming industry, particularly the core gaming demographic (as compared to say, the mobile gaming demographic, where OpenGL already reigns supreme). This hypothetical platform won't just leave the vast majority of Steam's existing titles in the cold, but will also have the unenviable task of forcing many devs to provide cross-platform support where none currently exists.

Valve is a juggernaut in the industry - but even that's a very, very tall order. There's certainly a renaissance of interest in OpenGL thanks to the mobile gaming side of things, but I'm extremely skeptical of claims that any player, even someone the stature of Valve, can get a majority of existing Steam devs to sign up to support OpenGL.

So, the PS3 does support PSGL, but that's not actually what you are thinking of. PSGL is a GLES variant (Not exactally, iirc, GLES 2.0 wasn't out then, and GLES 1.0 didn't have the required shader support). If you wrote your game to run using PSGL, you were actually doing pretty good for a full OpenGL Port.

Now, you are actually in fact thinking of LibGCM, which has very few similarities to OpenGL, is much lower level, simpler, and is in fact, the library that most PS3 developers do end up using.

Do however keep in mind, that any major game that runs on Mobile, OS X, or Linux, has pretty much API compatible library usage with OpenGL.

Then while the PS3 doesn't give you OpenGL, any cross platform game that runs on both PS3 and another system generally has a graphics system abstracted enough to make porting to an OpenGL system while not trivial, a rather well contained and limited amount of work. Surely enough to make it warranted to port to whatever graphics API Valve chooses to use, should it be a successful console.

And frankly the Microsoft monopoly on rendering API's. These days modern OpenGL is close enough, to not really matter. Having nearly feature parity, and these days all the fancy stuff happens in shaders anyways, the API just doesn't matter as much as the platform lock-in these days.

Yes, the steam console doesn't automatically get access to the entire library, but keep in mind, Steam on the mac has been hugely successful, and it had a very limited selection of games to begin with. You say that valve can't get a majority of existing devs to sign-up to support OpenGL, frankly they don't have to, rather they have to get just enough Steam devs to do such, and that's something history has shown that they do in fact have influence over.

Frostbite and Anvil run on DirectX and OpenGL-esq systems. IW engine even has games out on Wii. I kind of suspect that all the major engines out there have pretty decent cross-platform support these days. The publishers don't like to be locked into any one console unless they are getting an exclusivity kick-back from Sony, Nintendo or Microsoft.

It might require Valve to get people to recompile their games and tweak some stuff, but like you say, Valve are a juggernaut.

I think you're mistaken, Frostbite doesn't run on any OpenGL-esque systems.

The PS3 is OpenGL based. Specifically it's OpenGL ES with some extra proprietary Sony stuff. The important point is that the engine is not directly tied to any low-level rendering API and clearly has enough abstraction to be ported to multiple platforms with a reasonably low cost of development.

Your first comment isn't entirely accurate. While the PS3 does support PSGS, which is the library you are referring to, engines like FrostBite use the LibGCM interface, which has very little similarity to OpenGL.

The PS3 is not OpenGL based, it provides an optional OpenGL-like layer should titles want to use it but rarely do any these days use it. Frostbite pokes the GPU FIFO directly using the libgcm.

RAGE uses OpenGL on Windows too, and it cost them because the AMD OpenGL drivers were broken when RAGE was released, causing major graphical glitches with AMD cards.

Again, DirectX does have advantages. But they are seeming less and less important in comparison to the platform support these days.

Those "leaks" have been nothing more then people discovering the Linux server files, and assuming that means a client exists as well.

If the point of the steambox is to run games found on steam, it is entirely not plausible. It would end up about like steam on Macs availability would be pitiful.

Modern game consoles run with an OS that acts much more like a hypervisor than an actual operating system. There is a very distinct possibility that valve has choosen to develop the system components required to manage the running game, and the requisite steam services, without actually needing to take an existing off-the-shelf kernel +- any userland. Combine that with the fact that there are tons of other existing kernels which could fit the bill, BSD, l4, even darwin is out there.

And frankly Microsoft could be licensing them some sort of Windows variant OS. Afaik, they maintain a positive relationship with Microsoft, and other than the conflict of interest with consoles, Microsoft knows that Valve is a very powerful player in the market. Also, do recall that Microsoft has licensed out their OS before, it as a different world at the time, but the Dreamcast was using a Microsoft derived kernel.

I doubt that there is much benefit in using a "hypervisor" than just using Linux. On the PS3 with its 256MB of RAM and its 20GB hard drive, ~10 megabytes of RAM for kernel + core libraries and ~5 gigabytes of hard drive is a lot just for the operating system. On a modern machine that's just noise.

As for CPU usage, a properly tuned Linux kernel with unneeded services turned off shouldn't be significantly different from a hypervisor once it's booted.

Valve may try and set things up so that X11 is optional but even that should get mostly out of the way if you use the right libraries. (DRI etc.)

But there is benefit in not using Linux. Most notably from a business standpoint (GPL). And benefits from not using most of the existing infrastructure (I highly suspect that X11 would be wholly excluded,mrather than just optional.)

I'm not trying to suggest that they will use a hypervisor, but rather suggest that they have a ton more options then seem to be listed here.

I'm not a hardcore gamer, but I still have a Windows box I keep up to spec to play the latest COD et al when they come out. I still also enjoy playing a few old school games that aren't available for consoles, hence I don't have one. The only use I have for this box is gaming, as I do everything else on a nice portable 11" MBA. I would prefer to be able to use the MBA for gaming too, I can connect it to a monitor, mouse, gamepad, etc but it doesn't have the graphics power. I've used OnLive a few times which has been great, but their catalogue is rather limited. Personally I would rather see Valve buy (or build) something like this so I can game anywhere.

(N.b. If you haven't tried OnLive give it a go, they have trials for most games, Just Cause 2 is pretty fun and is a good example of the graphics it can do)

*I also don't have a HD TV which I guess is another reason.

If this is true, maybe that's why Valve has been working on Steam on Linux for so long. People over at Phoronix have been talking about it for years, but nothing official has ever been released.

After all, it doesn't make sense to pay Microsoft the Windows tax for every console, and they couldn't use OSX.

They would have to pay the windows tax, that is where 90% of the games are.

What about WINE?

I would think making Steam performant under Wine/CrossOver would be easier, and would make it available to both Linux and Mac users.

Wouldn't that cause a huge performance hit? I wonder how hard it would be to get a few big name games working in linux.

Wine isn't an emulator, it's an implementation of the Windows API, so there's not necessarily a performance hit.

Some games are known to run faster on WINE than on Windows.

Most games are slower, because the Direct3D implementation isn't as tuned as Windows' is, and much of it is done by translating into OpenGL. However, this hit is a lot smaller than you'd expect.

Before Steam was released for Mac I used Crossover a lot. But the performance was maybe half of the performance in Bootcamp plus loads of graphical bugs. And in some games(HaloPC) the mouse pointer was so bad the game was unplayable. Apparently X11 is single-threaded according to some support article I read. :-/

With Mac-native games from Valve, I get more or less the same performance as in Win7 on the same computer.

Is crossover that much worse on Mac or is it that we're playing different games? I bet the answer is somewhere in the middle. I only use Crossover to play Civ4 on Linux. I reboot to play Skyrim because my video card displays awful banding effects if I'm not in "high" display mode, and crossover only supports Skyrim on medium or low.

I played HaloPC, HL2 and TF2 in Crossover and the performance and correctness was not good.

Random question: When you wrote "Wine isn't an emulator" was that choice of words purposeful or a coincidence? Enjoyed it either way.

Well, Wine is what EA uses to make its games run on Mac OS X (although it doesn't exactly tell you on the box), so I'd assume not.

Not if valve invested time and money into reimplementing directX APIs using OpenGL.

They could also encourage new developments to use some other APIs such as OpenGL and just use DirectX for legacy.

I recently bought a 1.7GB game for my Android phone (Modern Combat 3), that you can optionally connect to a HD monitor using a HDMI out cable.

Quad core mobile chips are already in production and PS3 capable mobile GPUs are 6 months away (ARM Mali T-658).

Why is a console, a console and not simply a power/cooling/connectivity dock for a suitably powerful mobile device ? Id's flagship game, RAGE, has been supremely playable on an iphone for ages now.

Consoles like PS3 and Xbox already lag desktop-grade video hardware by 5-6 years, yet have a large market - so why cant the same be true for mobile hardware ?

The cost of a quad-core SoC isn't so high that it's worth only having one of them, and like you say, you can just connect via HDMI anyway - although that leaves you with a controller problem. Separating people from their phones - docking over by the TV and playing on the couch - is a big hassle, especially if you get a phone call mid-way.

IIRC the mobile RAGE is essentially an dumb on-rails shooter. The PC version is significantly more vast, complicated, and graphically impressive.

the fact that its a dumb shooter has nothing to do with its rendering ability. It's a dumb shooter because people have not learnt to buy $49 mobile games.

A mobile phone already exposes a touch-screen interface/proximity sensor/accelerometer/gps/bluetooth/etc The interface is richer, but different - in the same a dualshock is different from a keyboard. You do not need to separate a user from a phone if they have a power cord trailing to their phone. But actually, you dont even need to do that. Your phone can remain docked and your joystick can have a mic to answer calls.

I do understand that a mobile chipset may trail PC chipsets - but I'm comparing mobiles to consoles, which already trail and are still accepted.

Rage on iPhone is a rail shooter, it's not the same game as the PC/console version. It's been stripped down to the exact minimum it needs to support that predefined path through the game, and while that's great fun on a mobile, I don't think everyone would be totally impressed if that was the direction the next generation of consoles took.

It's a rail shooter because of control issues, not power. Drop your phone in a dock attached to a gamepad, and those issues disappear.

According to Carmack, it's on rails to reduce file size (texture LOD):http://toucharcade.com/2010/11/19/john-carmack-on-rage/

And then some. On the 360 Rage comes on 3 8GB DVDs. The megatexturing method that Carmack favours is intensely data-heavy.

Look at screenshots of both, the iphone version looks great for a portable game but there are differences in the level of detail.

Most iphone/android fps games look like PC games did ~2003

Hmmm. They must have something special planned because I doubt it's another console.

http://www.chrisnorstrom.com/2011/02/creation-the-steam-cons...

(above) What I predict Big Picture mode is suppose to be. But seriously, how is a $300-$600 console suppose to compete with a $800-$2,000 mid to high end gaming pc? And shouldn't Valve be working on bringing movie and music to steam first before taking this on? I don't know, but I'm dying to find out.

Why would Valve be interested in selling this to people with a gaming PC? Those people already play games on a PC, and are unlikely to give up their keyboard/mouse setups.

Your scenario is highly unlikely.

Thanks for the downvote and not understanding the concept.

Many console games outsell their PC games, PC and Console games can both reach the same graphical qualities. A lot of poeple play PC games with a controller because it's so much easier than with a mouse and keyboard (except for sim games, strategy games, and first person shooters). In many cases the only difference between console gaming and PC gaming is the setting. Console gaming takes place in a living room with a big screen tv and nice comfy chairs with room for others, PC gaming takes place in your home office or personal room and it's not as relaxing or natural feeling nor is there room for others. So by streaming a PC into the living room instead of buying more hardware to maintain and take care and hook up and make room for. Just stream your $1,000 PC to the living room.

If they are too loose with their hardware specifications then writing software for it will be a nightmare. Ideally it would be a rigidly specified but open platform. Actually, I wouldn't mind if it was a locked down platform like the iPhone but I wouldn't have to buy an expensive dev kit to develop for it. I've always wanted to develop something for the XBox or PS2/3 but couldn't because I would need the big bucks for the dev kit and approval to be a part of the developer program. Apple's proven that their model of development works.

Ha, about five years ago I thought about how I wanted Valve to compete with Alienware in the gaming PC business, and how they would brand their rig the "Steam Box". Interesting to see something happening here. Also, I bet a Valve-branded "SteamBook" laptop computer would be spiffy.

A bold move but it makes perfect sense. Manufacturers bring no value in the games business, its now all about contents and distribution. Apple and Valve will have growing power over the old business models.

I want Steam OS on my desktop. Optimized for gaming and little else.

Semi-related on Win7 (Pro/Ultimate/Enterprise only) you can use the group policy object editor to launch something other than explorer.exe as the shell.

In the past I've used this to launch directly into Steam. My friend has a spare PC in his living room mean for games with an open invitation for anyone to use, but didn't like people abusing it/his internet (We have data quota here, iirc his is 75GB a month then he gets throttled to 65Kb/s), so he asked if I could make it for steam games only. Now when you turn it on, you get a steam login prompt, and basically nothing else. All the installed games are available to anyone who owns them on steam, and work normally.

Hitting Ctrl+Alt+Del, opening the task manager, hitting File->New Task and running explorer.exe gets you the start bar back, so its not locked in or anything.

In general its pretty useful if you have a Windows box that's used for basically one thing, like XBMC or Steam.

Is there a noticeable performance difference? Sounds cool

KB+M and user mods please. I'll drop xbox in a heartbeat.

Reply on news.ycombinator.com