Cool day. I went to a SIGGRAPH meeting featuring a talk by Johnny Lee (HCI researcher and one of the core contributors to the Kinect project at Microsoft) that I had heard about from qDot (the other OpenKinect maintainer) three hours before the event. Thanks to the fact that I'm already in Silicon Valley, I was able to go anyway.
It was cool hearing about how Microsoft had to learn how to deal with taking research where things are often poorly or underspecified and being able to integrate such a team with their software engineering prowess where they have a lot of skill at successfully solving well-defined problems. One thing Johnny noted in particular was the fact that traditional software engineering expects linear progress over time - you expect to accomplish twice as much work in two months with the same manpower as in one month. That doesn't hold true for research - it's much more a series of quantum leaps, and much harder to predict future performance. Now that I've heard about some more of the difficulties involved in the Kinect project, I'm even more impressed by what Microsoft was able to deliver. Congratulations, Johnny - you accomplished your goal of commoditizing depth cameras. :)
After the talk, I hung around and chatted with the other attendees and Johnny about all manner of HCI-related things until about 10pm. When I got home, there were 20 snails in our front walkway. I counted them.