Open Source at CES

CES, the largest consumer electronics show in the world, held every year in Las Vegas, has just wrapped up. Sadly I’m wasn’t there, but luckily Doc Searls has been at CES keeping track of Linux and Open Source stuff at the show. As usual there are more Linux based devices that you’d realise or that many press outlets (save Doc and a few others) would let you know about. A smattering of devices from ‘Oh, so it runs Linux’ files; Motorola’s settop box and RAZR2 V8, iRobot vacuum cleaners, Dish Networks VIP722 PVR, a whole host of Intel MID devices.

What was interesting is that this seems to be the first year that truly open devices started appearing. Our friends from OpenMoko were there, as were Dash with the Dash Express a GPS based on the OpenMoko hardware. (NB, like Doc, I’d like to give kudos to OpenMoko for running Linux on the laptops on their booth).

Also it would seem Buglabs were one of the hits of the show (they were nominated for Best of CES by C|Net). Their hardware looks really cool with the BUBbase and BUGmodules adding near infinite possibilities for creative usage. I really liked this photo from Doc of Angel Roman’s setup (Angel is a Buglabs engineer), featuring a Lenovo X61 running (Ubuntu) Linux, a Nokia N810, and a Motorola (Linux based) ROKR phone, and the BUGbase and modules (on the left).

Software above the level of a single device

Tim O’Reilly has repeated (the very repeatable) quote from Dave Stutz,

“Useful software written above the level of the single device will command high margins for a long time to come”

This line came at the end of Dave’s parting letter to Microsoft, after he resigned to become a winemaker. I’ve always loved the quote and struggled to understand what it might mean for (open source) software and device development. Tim’s canonical example is three way system of iTunes Music Store (web service), iTunes (PC based software), and iPod (hardware).

Of course the level of integration Apple are able to achieve is made somewhat easier because they control all three layers and have no interest in providing interoperability. In this light Dave’s quote always makes me stop and think about the many devices that are or could be in my home that don’t communicate or work together. Fileservers, laptops, PMPs, cameras, phones, PVRs, game consoles, stereos, all create, store, or consume various types of media. Yet they don’t all talk to each other or even use the same media formats. It doesn’t seem likely that consumers will get all these devices from a single company, hence cross device & manufacturer interoperability has to be the way forward.

For non-networked devices, such as typical PMPs and cameras the laptop will be the integration point, and for Linux, applications such as Rythmbox, Banshee, and F-Spot do a great job of storing, cataloging, and annotating various media. All three work above the level of one device by integrating with web services and using existing media sharing protocols.

Increasingly all devices will become networked and the problem gets significantly harder. I might want to browse photos on my camera via my the computer attached to my TV, or listen to a podcast I downloaded to my laptop on the stereo in the kitchen.

The UPnP standard is aimed to allow applications and devices work together with little or no configuration. We created the GUPnP project to help make software that uses UPnP as easy as possible. UPnP won’t solve all the problems - there’s always the danger that a manufacturer will have a bad implementation, or try to embrace and extend, and there is always DRM to raise it’s ugly head. But UPnP offers the most comprehensive system for connecting these devices together.

Already there are many devices that support the UPnP standard, and hopefully, now that GUPnP has stabilised, we start seeing more use of UPnP in Linux applications, so that one day my house might truely be humming to the sound of software above the level of a single device.