Editor's note: Tadhg Kelly is a veteran game designer and the creator of leading game design blog What Games Are. You can follow him on Twitter here.
There had long been hints that Valve was up to something. Hints like Gabe Newell decrying the state of Windows 8 and speaking in favor of Linux. Hints like Big Picture, the TV-like dashboard. Hints like the Xi3?s boxy little machine. With new consoles and microconsoles starting to pop into existence, this week Valve finally revealed its answer: SteamOS. Steam Machines. Steam Controllers. Boom. Its ambitions are not to launch a console but a whole solution for home gaming entertainment.
In a sense it has to.
Valve's Problems
For Valve and many other studios the decline and fall of the PC is a slowly dawning apocalypse. Sales of gaming PCs may be up while the rest of the market is down but there's a point at which the support that comes from the wider PC ecosystem starts to dwindle. Perhaps video cards start not keeping pace. Perhaps driver software becomes less updated. Perhaps commoditized components stop being cheap and PCs become much more expensive. The knock-on effects of this decline could wipe out Valve, Blizzard and a number of others.
However Valve's second, and largely unique, problem is that a recovery in the PC space implies tighter integration of the platform by Microsoft. With the purchase of most of Nokia and a search for new leadership underway the Wyrm may finally be starting to turn at Big-M. Windows 9 (or 8.5, or whatever) may well turn things around. If it does though it will be for a Microsoft that's far more interested in device-and-service thinking. It will mean more prominence for the Windows Store. It doesn't compete much with Steam today but one day it will.
Finally Valve's third problem is its ecosystem. It has a massive following of both players and developers. Steam is perceived as art house venue of games and every indie wants to be there. That ecosystem has evolved hand-in-hand with the PC paradigm since its inception and is resistant to fundamental change. Steam's adoption took forever and required deeply discounted sales to overcome suspicions. Nowadays it's huge but that doesn't mean its audience is ready for a big shift.
These problems have essentially defined Valve's next generation solution, but they breed problems of their own.
SteamOS
A few weeks ago at LinuxCon 2013, Gabe Newell gave a talk about the future of Linux as a gaming system. Obviously this was something of a preach-to-the-choir crowd and so the talk was warmly received, but outside of its domain does anyone really care? Linux may be powerful but it's also geeky. Getting into it requires rolling up your sleeves and doing some digging. It's cool for motivated people but for everybody else (even the comparatively savvy gamer crowd) it's a chore. This is why, although it has had Linux support for a while, Steam is still largely a Windows client in most people's eyes. It's just easier to make and run games in Microsoft's house.
With SteamOS Valve is trying to be the Android of the game consoles, to do the hard work of taking a cool kernel and make it user-friendly. The company sees itself as providing a nexus solution to everyone's problems, and maybe even harboring hopes of Android-sized success. Manufacturers can use SteamOS as the core of a gaming deck that they would then sell, creating a class of home-hub devices that would be a bit like consoles but also their own thing. There would be choice and range in the market yet a common platform to power them all, much like today's PC. And there would be neat features like streaming games across your home network. Like an Apple TV and iTunes.
Sure, but the big problem with that vision is developer support. The majority of games in the Steam catalog are old, all created for the Windows PC paradigm. They're built to work with mouse, keyboard, DirectX or OpenGL and so on. Support for updating them to work on modern Windows machines is hard enough, but to shift platform? Many don't.
As an anecdotal example, I own about 150 games through my Steam account. Only about half of them work with Mac. The other half are either pending support or simply won't ever have it. The Mac is arguably much more stable and better supported than Linux PCs tend to be, and yet plenty of game makers are not bothered. They can't make the business case for it to be worth the effort. So what hope does Valve have of convincing risk-averse game developers to create full ports of existing games for SteamOS? Or even convincing them to make original games designed specifically for the OS?
On the face of it, very little. Game makers tend to be look-before-you-leap types (as I have often argued, to their cost) which means new platforms often have to deal with chicken/egg paradoxes. You can't get interest in your platform without great games, and you can't get great games on your platform without proving it has significant interest.
For non-dedicated gaming platforms like PCs, iPhones and Facebook this problem tends to resolve on its own because so many people buy into the platform for other reasons that game makers overcome their skepticism. However dedicated platforms usually have to solve it by opening their wallets. That billion dollars that Microsoft is ploughing into Xbox One for game development? This is why.
If a platform essentially offers to pay game makers enough to make it worth their while, they'll dive in. But is Valve really willing to do that? Sure it has its own key franchise in Half Life 3, but I seriously doubt it will make the mistake of making that game Steam Machine exclusive. If it is relying on providing the platform and then hoping developers will support it, it will need to sweeten that deal somehow.
Steam Machines
Irrespective of the support question, the strategy of being the Android of game consoles reminds me of Microsoft's Windows Media Center adventure. The basic gist of WMC was a mini-PC sitting under your TV and acting as a powerful hub for all your entertainment. It was primarily for movies and recording TV but had other conceivable uses too, such as games.
This idea has always been compelling to some people yet it's never really taken off. One reason is that the hardware has sucked. It's been expensive, underpowered and far more fiddly than it was supposed to be. The living room is not the desktop and does not induce nearly as much tolerance for messing around to make something work. Does anyone really want to have to faff about with their home hub while their family looks on? People want a product that they are confident will not be like that.
Existing console platform manufacturers tend to shy away from multiple-manufacturer ideals for precisely this reason. Similar to the difference between today's iPhone and Android, working with partners tends to detract from control over the user experience and leads to fragmentation. Standards can not be enforced nor compliance ensured. The cognitive load that this places on the user is high, which the user then finds discouraging. The guarantee that every game will just work can not be assured.
So one of my big fears for Steam Machines is that the balance of power between Valve and its partners will favor the partners. The partners may want, for example, to bundle their own dashboards in their machines and develop their own brand relationship with their customers. We see that happening all the time on Android, so why not SteamOS? All that deluded cruftware that you get when you buy an OEM PC? Imagine that under your TV.
My second fear is that Steam Machines will turn out underpowered. Valve may want them to be top-notch but partner manufacturers often have to rationalize the cost of production and shave a bit here and there to make a profit. Ecosystem-led approaches tend to result in halting periods of innovation as a result. For some audiences this wouldn't be a problem, but for PC gamers it is. The perception of being behind is not one they enjoy.
Look at Microsoft's attempts to make Windows-powered tablets and phones work. For years the idea was to partner with everyone and let hardware sort it out, as it had in PC. The result was years of bad Windows Mobile phones and awful touch PCs. Instead the company had to pivot from software to devices and just make the units themselves (or with Nokia). It had to because it's in the nature of OEMs to sit back and make smaller innovations with low-powered (and often just crappy) tech. But to compete with iPad Microsoft needed something much bolder.
While they may have their flaws, one thing that Xbox One and PlayStation 4 will be good at is delivering super-lush (and reliably so) games. Like the Surface or the iPad, the vertical control over the platform has a lot to do with that. I suspect a similar story will play out for Steam Machines, and that Valve will eventually conclude that it should just make its own. It goes against the DNA of the company in many ways (it's used to working with lots of others) but still.
My biggest fear for Steam Machines, however, is price. Sony is able to sell PlayStation 4?s for $400 because it's taking a significant cut from each game. The console is probably being sold at cost price, or even below, because higher prices tend to make customers queasy. Unless a platform has a deeply held marketing story of being much better (say Mac vs PC) then it can be very hard to make a case to customers that they should pay more.
I just have a hard time believing that there are many gamers out there willing to spend $900 on a console, when one of the big cheering points of PS4 vs Xbox One was the $400 vs $500 starting price. Gamers who want big games will be able to buy a dedicated console for half my guesstimated price of a good Steam Machine. Meanwhile the ones who want indie or casual games will be able to buy something an OUYA for $100. All in all I think this hub strategy means Valve's primary customers for Steam Machines will be the people who are already invested in Steam, and only a subset of those who haven't already shelled out a couple of grand on a big PC gaming rig.
Which leads me to…
Steam Controller
For a long time I've wondered who would properly reinvent the modern joypad and do something cool with it. Joypads have been ungainly multi-pad multi-button affairs for a long time, holding on to legacy features like D-pads long past their sell-by date. So I love the idea of the Steam Controller. While the initial renders of it look kinda kooky, I'm excited.
And yet: Do PC gamers want to move away from their mice and keyboards? Irrespective of the problems looming over the PC landscape and how they might be an issue for game makers down the tracks, don't people who go out and buy gaming computers do so because they like to game on computers? Do they actually want to go console at all? Joypads have been available for PC forever yet haven't ever really made the leap into being considered a part of the default spec. They still lack support in many games, and there's a strong resistance to them from the pre-existing community. It simply believes M+K to be best, and for many types of game it is.
And also: Do PC gamers want to move out of their bedrooms and into the communal living room? The guy who plays MMOs night after night in private doesn't necessarily want to do that in shorter bursts on TV because his family wants to be able to watch Netflix. The quirky indie fan might prefer the experience of being absorbed in Amnesia: A Machine for Pigs by a screen close to her face. Closer screens blot out the world more effectively than the at-a-distance feel of TV gaming.
Admittedly that's a somewhat archaic view of modern family life, yet one that still makes me ponder. While it is possible to cross-sell users from one platform to another (Apple and Nintendo do it all the time), with PC gamers there's often a component of the deliberate choice. They have often played many a console game (such as the good folks at Rock Paper Shotgun) but ultimately decided they preferred computer gaming. They are the reason why gaming PCs boom as all other PCs fall.
So I just don't know that joypad gaming will ever be for them.
Streaming and Funding
Valve's biggest asset is the legion of fans willing to give Gabe the benefit of the doubt. That's no small thing. It's other big asset is its methodical approach and willingness to keep driving at a problem until it's solved. It did it with Steam, turning a much-hated e-commerce barrier into a thriving community over the course of half a decade. So while I may have questions over what the company has recently announced, I firmly believe that it will answer them in the long term.
Steam Machine's likely success comes down to the price of the consoles and the support of game makers. This is why perhaps the most interesting announcement from all of this week's news was game streaming. You'll be able to stream your Steam games to your SteamOS console under your TV and play them across your local network. Hopefully that won't be as laggy as remote solutions like OnLive turned out to be.
While customers might struggle to buy into the idea of a $900 console, they'd be much more amenable to a $100 Apple TV-sized streaming unit. I could, for example, see a partnership with folks like OUYA to develop a SteamOS player. I could see something similar for Android tablets, particularly if the Steam Controller could be made to work with those systems. Perhaps by doing something in that vein the company could get access to the living room without all the overhead.
And then what about the actual PC? If Windows is slowly sunsetting that doesn't necessarily mean that the PC form factor will go with it. The PC game player will still want a gaming computer to enjoy her MMOs in her bedroom. While SteamOS has been pitched as more for the living room than the desktop, why not also have a desktop edition?
As for game support Valve has one big advantage that it's not yet using well. Through efforts like Steam Greenlight Valve has brought its community into the publishing process, which has given many indies exposure. The missing piece, which I've previously argued, is crowdfunding. There's no good reason other than reticence why Steam isn't essentially the Kickstarter of games (even more so than Kickstarter itself), and if it were then the resulting funds would encourage game makers to get on board with Steam Machine.
If Valve could pin crowdfunding to its new console platform together then the sky could be the limit. However it remains to be seen whether the company is willing to take that step.
Editor's note: Jack Krawczyk is director of product management at Pandora. Prior to Pandora, he helped build Paid Discovery at StumbleUpon and was a founding member of Google+. Follow him on Twitter 
Google entered the picture and changed direct response advertising from an inferred impact model into a direct keyword to conversion ROI report. The limitation of search, however, is that consumers are most often not in the mindset of making a purchase.
