On the Steam Box

Posted by Jack on 2012-12-11 at 02:52
Tagged: gaming

The Steam Box has been making the news lately, most recently with a confirmation from the venerable Gabe that it's a thing that may exist in our plane of reality. There's also been a lot of talk lately about the Steam Linux beta (that I got in to, hooray!) and the release of Steam's Big Picture which is a console like controller interface intended for big screens (i.e. TVs).

Because of the timing, a lot of these Steam Box announcements, and a lot of the buzz around it has been that this box will run Linux. As such, in the gaming community there's a lot of "whoa, Linux... what does that mean for us?" and in the Linux community there's a lot of "finally Linux games!" I love Steam, I own a handful of titles in it, and I'm extremely pleased that they've put out a native Linux client (even in beta form). However, I am totally unconvinced that this is going to be a Linux based Steam machine, despite the timing. Here's my logic.

Why Not Linux?

I think that it's likely that a Valve console would take advantage of the huge Steam library. They've put almost a decade of effort into turning Steam into a slick, painless, even fun experience and they've sold millions of games. Steam is now a big release platform for a lot of AAA publishers (Bethesda, id, Eidos, Firaxis, Gearbox, etc.) as well as a load of indie publishers that wouldn't have found nearly the following if it wasn't as painless to find out about them and pay them.

It seems to me that Valve would be making a huge mistake if they're not parlaying that massive, successful library directly into their prospective console. Bringing a handful of AAA launch titles into your living room on day one makes the Steam Box just another console that differentiates itself with maybe a handful of Valve exclusives (Half-life 3, anybody?). Bringing 2000 mature, well-loved and already purchased games into your living room with a cheap box and promising all the future PC releases would be killer. In addition, a Steam Box that was just a Windows machine with a slick interface would suddenly become the defacto PC spec - solving an issue that game devs have struggled with since the very beginning of PC gaming, namely how to deal with the thousands of different hardware and resource configurations. Finally, it would have the added benefit that literally any game that runs on Windows would effectively work out of the box (perhaps with a little tweaking for the spec, or any novel input devices, but without the pain of a full port).

There is a whole lot of greatness in bringing a cheap, well configured and standardized PC into the living room with a giant library of already working and popular games.

Unfortunately, with a Linux based OS on it, this is impossible. Valve titles would be ported, of course (and in fact that seems like it will happen regardless thanks to the Steam Linux beta) and a fair number of indie games already have ports, but what about the library titles that would strengthen a console release? How many developers can Valve convince into doing a free port of an older game? I believe the answer is very few because for most of these old games there's no profit in it. Technology like Wine could be used, but if your focus is on solid gaming experience that's a whole new set of problems. As such, using a Linux based OS would almost completely obviate the advantage of the Steam library.

Why I Could Be (And Hope I Am) Wrong

First of all, perhaps I'm overstating the likelihood of the Steam library coming to the living room. A lot of the older, unportable games would not be controller friendly and if they're aiming for a more traditional controller oriented approach (instead of controller, mouse, and keyboard) they'd be worthless and hard to play even if they ran perfectly. Not to mention the fact that, by definition, the entire existing library already runs elsewhere which doesn't really give the Steam Box any draw over a gaming PC except perhaps to those without the cash for an expensive pre-built or the know-how to build their own.

In essence, maybe losing compatibility wouldn't be the worst thing in the world and if that's the case Linux starts making a whole lot more sense. You get a very mature stack from kernel through display and because you're targeting a single hardware configuration you can create a stable, well-tested release on top of open source components fairly easily. Compared to the amount of time it would take to custom develop the entire stack, the bugfixing would take a trivial amount of time and effort.

Second, Linux is free as in freedom. What other console developer would be able to glean fixes and features from unpaid volunteers?

Third, Linux is free as in beer. A Steam library compatible approach would have to come packaged with a Windows license which easily adds $100 to the unsubsidized price tag of the device. Ideally, if they chose to go this route, they could get a deal from Microsoft. The Dreamcast, for example, ran a version of WinCE developed by Microsoft that was seamless. That was before the Xbox hit the scene however and I highly doubt Microsoft would be so amenable these days. On the other hand, if the alternative is to have a Linux box running AAA games, they might be better served by giving Valve a deal to maintain their edge in the desktop space. Either way, though the box becomes more expensive and they lose the advantage of the open source stack.

The final, and best fact for the possibility of Linux on the console is that, if they chose to ignore compatibility, and went the more traditional console route with a release in 2014 / 2015, they'd have plenty of time to rally support for Linux titles, get the already existing Linux ports lined up and polished, and - in the end - come to the table with a more extensive library than any of the competitors.


Windows* Massive Steam library already polished and working
* Easy to productize fast
* Already mature
* Take a price hit on licensing / packaging Windows
* Less fine grain control of software stack
Linux* Free (as in beer) - lower end price
* Open source - get fixes from volunteers
* Already mature
* Not many existing ports in the catalog
* Brand new platform without much industry expertise

I'm still not convinced that this isn't going to end up being Valve's effort to further monetize their work on the existing Steam library, but if they are willing to start a serious console from scratch then Linux is cleary the way to go. None of the information we already know about the device seems to indicate which approach Valve is favoring and as such I think it's premature to make assumptions.