Google’s Executive Chairman Eric Schmidt actually did not say that Chrome OS and Android would remain forever apart.

Google’s Executive Chairman Eric Schmidt actually did not say that Chrome OS and Android would remain forever apart.

more on linux.com

blog.liangzan.net

Wong got hold of the Leap Motion hardware and SDK by participating in Leap Motion’s developer program. He was one the lucky few that was selected. A few weeks later, the Leap motion device was mailed to him, free of charge. It came in a dull black cardboard box. Within the box lies the device, a cable, and a card bearing a message from the founders.

more

Wong’s Conclusion…

As I presented my demo to the local javascript meetup group, everybody was visibly excited by Leap Motion. Leap Motion is an amazing product. Invariably, we wondered what potential applications could be built on Leap Motion. Games are an obvious application. Minority report styled navigation is another. For a long time we have been stuck with the keyboard and mouse. Then touch devices exploded onto the scene. After playing with Leap Motion, I firmly believe it is the dawn of another era of gestures styled devices. I’m living in exciting times.

Development experience

Like a kid with a new toy, I was eager to build applications with it. I chose to try the Javascript API first. The Leap SDK runs a web socket server which allows the Javascript bindings to access the data. Usage of the API is boringly simple. It is an infinite loop which gives you a frame every iteration. The frame is a JSON object which contains the data on the pointers.

The first leap experiment I did was to replicate Mike Bostock’s D3.js particles. I got a trail of particles to follow each pointer.

Gestures naturally come to my mind. The SDK has no provision for gestues yet. I have to recognize it. And that is an Articifial Intelligence problem. I looked for gestures related libraries to build on. And I found the $1 unistroke recognizer. They had a Javascript implementation. I used it to recognize gestures in my next leap experiment. It worked. But it has caveats.

A gesture has a start and an end. Leap runs in an infinite loop. I had to define a start and end. In my gesture experiment, I defined 2 states: fist and point. The fist state literally means clenching into a fist. It acts as a blank state which I use to define the end points of the gesture movement. The point state means an extended finger. In that state, movement is captured. While my experiment is able to recognize gestures, it did not do so cleanly. When I change between a clenched fist to a pointing finger, there are some jitters unavoidably. That polluted my gesture shape. It is very hard to form a clean shape. There should be a better way to define the end point of a gesture.

Experienced with platform integration/prototype hardware
Adept at problem solving in the context of software integration
Embedded Linux
Kernel level/ device driver knowledge
JTAG debugging
ARM MIPS experience

Looking for somone with the following….

Essential:

• OO Design
• HTML/CSS
• LAMP
• AJAX
• IP, UDP, TCP, ICMP & network programming e.g. BSD sockets

You will also posses experience of one or more of the following:

• Java EE
• C/C++
• SIP, RTP, OpenSSL/TLS
• Linux Sysadmin, security, firewalls
• Version control (Git, SVN, Mercurial)