Speculating about SteamOS

It’s been a tremendously long time since any new content appeared here (I’ve even been considering shutting this web log down) but today I find myself eager to share some thoughts!

What could stir me from my slumber? Somewhat surprisingly I find myself uncharacteristically excited about a gaming platform and, as I no longer share an office with fellow nerds, I have no recourse other than to gush at the internet.

Despite not being a huge gamer (I have likely bought fewer than half a dozen games in the last two years) Valve’s recent announcement of SteamOS has me rather excited. There are three main paths along which my mind wandered when pondering what SteamOS could be:

Firstly, as a self-confessed operating systems nerd (and someone who has spent most of his career on the fringes of that world) I’m excited to see how Valve have architected their Linux-based OS. Whilst it’s possible that they could just choose to ship a spin of an existing Linux distribution (such as Ubuntu or Fedora) I’m hoping they’ve been a little more adventurous.

Thinking about SteamOS for a few minutes whilst browsing their announcement page a system more akin to Android in many ways comes to mind – a highly customised OS platform tuned to the needs of an entertainment device, rather than a stripped down generic OS. I’m certainly not drawing the comparison with Android because I’m expecting Valve to have written huge swathes of OS components that make SteamOS a different beast to a more traditional Linux, quite the opposite – I fully expect them to make use of the excellent Linux userspace which is readily available and widely tested. It’s how they put their OS together and structure things in terms of distributing software, updates and security (more below) which I expect to be different. Particularly as there’s unlikely to be any need to consider ‘legacy’ applications.

It’s perhaps unsurprising that I half-hope, half-expect Valve to be using the tools of the Yocto Project to develop their custom Linux-based OS. Standing on the shoulders of giants in the Yocto community Valve could, with a relatively small team of engineers, create their OS with a focus on the unique requirements of their entertainment platform instead of getting bogged down in the sundry, largely solved, problems of distribution construction.

Secondly, as someone with an increasing interest in information security and secure and trusted Operating Systems I see a good requirement in the platform for interesting security work. I look forward to interesting security solutions which balance human factors (I expect the SteamOS to be primarily driven by a joypad) with the need for a hardened OS which protects a user and their sensitive data (most likely banking details) from attack whilst ensuring the platform remains as non-intrusive as possible.

Finally, as a Linux developer, I’m rather excited about the prospect of folks (particularly some of my friends and family) having Linux-based entertainment machines in their living rooms. I’ve been thinking about a largely vapourware project to build a digital platform to enable pen and paper role players to game with their group, regardless of geographical separation. The project has remained vapourware, at least in part, because each imagined iteration has fallen short of anything tangible – largely because there hasn’t been a compelling platform to target. I have considered (and in some cases prototyped) targeting the XBox, the web and tablets, as well as writing a cross-platform desktop solution, but have hit road blocks when attempting to architect for each of them. (Side-note: huge kudos to the Roll20 team for realising the vision). I hope to be able to target SteamOS using the tools I am familiar with and re-using components of a gaming platform, such as multiplayer chat facilities.

Oh yeah, fourthly – video games! \o/

Update (14/12/13) – the first version of SteamOS was released yesterday and it looks to be a fairly stock GNU/Linux created as a Debian derivative.