IDF: Rattner explains Intel's contextually aware computers

Intel's chief technology officer, Justin Rattner, used the last keynote of IDF 2010 to detail the company's work in the field of contextually aware computing.

The concept of contextual computing was first mooted in the early 1990s by Mark Weiser, who ran the Computer Science Laboratory at Xerox PARC for seven years.

The theory suggested that, as the number and power of computers increased, they would have to become more aware of their surroundings and owner's movements in order to be more efficient.

Rattner said that Intel had been working on computers using contextually aware software to learn about their users using a mixture of hardware and software sensors.

"Contextually aware computing will fundamentally change how we use devices. Future devices will constantly learn about you, and will probably even know how you're feeling," he said.

"It will know your likes and dislikes, and will learn where you want to go on the internet."

As the software learns more about the user's preferences it can more accurately predict likely things of interest, according to Rattner.

Those preferences would then be stored in a cloud proxy server and downloaded onto other devices, or even shared with third-party applications.

Rattner showed off a number of applications, such as sensors worn in his socks which recorded his gait and walking style. Similar sensors could help elderly people, perhaps to warn a person if they are about to fall over.

Intel is also working with Fodor's on a contextually aware travel guide, which uses GPS and Fodor databases to provide tourists with information on their location and directions to places of interest, based on the software's knowledge of their preferences.

The prototype software can also be configured to set up an automatic blog, noting your movements, any photographs you've taken and which shops and sites of interest you have visited, and publishing it online.

However, the ultimate goal for many is a direct brain-to-computer interface. Intel is working with Carnegie Mellon University on thought recognition and sensing by computer.

"The ultimate sensor is direct understanding of human thought by machines. It is possible," said Rattner.

However, such advances are a long way off, and Rattner joked that they are unlikely before IDF 2020.