Brief Thoughts on Kinect

Microsoft Research has finally released its free SDK for developing Windows applications that utilize Kinect -- the natural user interface device originally sold as an Xbox 360 peripheral.

I've never thought much of Kinect as a game controller. As much time as Microsoft has spent trying to sell it over the last couple of E3s, the only really great thing they managed to do was sell it as a means of interfacing with the main menu hub. The games have all been iffy, especially attempts to integrate it into the traditional genres with anything other than voice recognition. It's hard to make immersive game experiences without any sort of controller, even if only because of constraints on the camera and physical movement of your avatar around the game environment. The result is sports minigames, dance games, and an assortment of on rails experiences -- not bad intrinsically, but certainly more limited than Microsoft wants us to believe.

Still, the basic interface manipulation powers -- both based on physical, three dimensional gesturing and voice recognition -- is potentially revelatory. Looking through the hacks that have been popping up since the initial release of the device, you can see so many applications for recognition of three dimensional movement of the hands and arms alone -- and this sort of tracking makes full body identification essential so as to not get distracted by other noise. I'm greatly hopeful for the sorts of applications that will result from heavier use of Kinect and other NUI devices in traditional computing.

Still, Microsoft was brilliant in releasing this as a toy first. Even if that isn't ever going to be the best use of the technology, we needed a phase where people were excited about the sheer novelty of this thing in order to develop a proper vocabulary for developing with it. Had this originally shipped as some much vaunted peripheral (a la the Magic Mouse), it would have been laughed at for its lack of support and non-existent consistency of use between applications. Instead, we had a period where that was expected because its primary purpose was as a fun diversion -- a way to sucker people out of their money instead of as a meaningful productivity tool. Months later now, we have experience determining current best practices for recognizing and translating certain behaviors and a proper set of verbs for describing interaction with this thing and an API for productivity development that supports these verbs out of the box.

Kudos, Microsoft. Here's hoping that you can pioneer NUI support in Windows 8 as you continue your positive trend of starting to give a damn about your customers.