essay

SteamOS: The Questions

Here’s the thing: I don’t trust Steam. Which is not to write that I don’t use Steam or even prefer it often over trying to figure out dependency issues with games that don’t use it. Games for Windows was frequently a mess in that regard — and recently closed down anyway. And even Desura, which is actually pretty good most of the time, fails to grab attention away from the 500-pound gorilla in the room that is Steam.

So, we all deal with Steam. If you want to play computer games, and especially those AAA games, you will end up using Steam. It’s become the de facto and even in most circles the lingua franca of playing computer games. Most people use Steam — millions upon millions, in fact.

And therein lies the problem, at least for me. Systems oppress. Maybe they don’t mean to, maybe they do. Yet, the community practices of the environment, from the user interface to the way it handles data, informs how people understand the system and each other in the process. The system mediates and thus becomes the language.

To borrow a very Foucauldian idea, systems ‘normalize’ its users over time. So, Steam, with its years of practices and millions of users, have a very smooth system. Many of the “bumps” (conflict and often critical areas) have been ironed out. Steam users, for the most part, have an idea of what a “game” is, for example — even if I and many others disagree with them.

The plan, then, of an operating system based on this concerns me. Even though it is Linux-based and supposedly has “openness” at its heart, it’s still Steam. But now, of course, at a system level. Not just interfacing through Windows or OS X, but handling the hardware too.

It’s easy to boast of higher speeds then. Operating systems are usually designed for general usage, with balance toward both small processes and potentially huge ones. The system handles everything (mostly) equally and users get a good (for relative values) experience overall.

Optimize that for games and things improve. Without the overhead of additional processes, more space and time can be spent on a game itself, reducing the operating system to its barest levels. It’s how consoles — except maybe Xbox One and PS4 — work. Everything is designed with gaming in mind — from the hardware right on through the OS layer.

From a performance side, everything reads great. Valve, through Steam, has effectively created a pseudo-console. Instead of having to worry about handling the hardware (which carry huge initial and ongoing costs), they get the users to do that themselves. Just install this OS and you have a machine dedicated to playing games. Everything’s great.

Except, it isn’t.

Because that was my very first response to this news: who is this for?

What is the target market for this exactly? Users will need an additional machine to the one they have. One for playing games and another, running this SteamOS, to act as host for the “in-home streaming” (whatever that means). So, that’s two fairly new systems. And a decent network. And the ability to get everything running.

Because, as I can attest from years of experience, Linux isn’t the most user friendly of kernels (few are). Sure, collections like Ubuntu have improved the process of installing a Linux-based system tremendously. And it has come a very far way from the days of using floppy disks and FTP sites (back when I first stared working with it). All you need now is a DVD or decent connection to download a client.

Then there is the openness issue. How open is “open”?

Steam has done a great job with its Workshop functionality. Now, instead of potentially downloading and installing extra programs and modifications for games already in Steam, it’s all part of the same system. Which is also the problem too: it’s all the same system.

Everything that happens, and that now includes many development tools too, is enclosed by Steam. They have complete control over the process. They handle the updates — and the access.

What that means for an operating system remains the central issue for me. Like I had to sign to use their services, will there be a EULA for the operating system too? Will I be allowed to change things? Can I change the libraries or drivers?

And, as a way to close this up, what does this mean for the future of computer games? If nothing else, this is a major step toward a centralized Steam platform. Not just in the case of having different, Windows or OS X, versions, but designing for the SteamOS too. And the potentially thousands if not millions of different hardware configurations underneath that.

Just because it is Linux-based doesn’t mean driver issues go away — in some cases it might even get worse. SteamOS will, supposedly, try to handle all that and wort out the mess for the user. But that doesn’t mean it won’t be there.

Given the fears and concerns several employees at Valve have voiced about Windows 8, I can understand this move for them. If the problem is the operating system itself, merely replace it with one of your own. You don’t need to worry about it then, you are it.

Still, this is greater power for Steam. It’s not just a launcher anyway. Nor a service on a system. It’s the OS itself. For good or bad, it’s complete control over a user’s machine.