He called this next generation “spatial computing” and listed numerous attributes: many-core processors; parallel programming; seamlessly connected and fully productive; context-aware and model-based; personalized, humanistic, and adaptive; 3D and immersive; and utilizing speech, vision and gestures.
Mundie’s demos showed some progress in fulfilling Bill Gates’ dream of natural interfaces and seamless computing. The challenge for Microsoft will be turning lab demos into real products and services that can scale. With the Internet as the platform, and not Windows, Microsoft will have many more competitors, and partners, in its quest to realize the vision of spatial computing.
What comes next? Microsoft’s Craig Mundie says spatial computing.
(Credit:
Dan Farber/CNET Networks)
Then he showed how a smart handheld device could be used to navigate in a physical space. Pointing the device at a particular space would show local information, such as when buses were expected to arrive or what stores are having sales that would be of interest to the user based on their profile.
“For a few thousand dollars you could put in an assistant who can guide robotic interaction,” he said. “There is a wealth of opportunity for this, and it will allow people to develop applications and change the way the bulk of the population interacts with computers.”
See also: Mundie: The cloud needs killer apps
The 3D store environment was stitched together with Photosynth technology and interactive. Mundie could “walk” through the store and have a text or voice conversation with a store representative or someone, such as his wife, via his buddy list. In addition, he could watch videos and examine 3D models of the art objects, spinning them around to look at all the different parts of a sculpture.
The prototype system is a resource hog, consuming 40 percent of its eight-core processor system even when idle. Eventually, Mundie said, such a system could be used for rural medical clinics.
Mundie also said that software development hasn’t graduated to become a formal engineering discipline. “The resilience of systems is not up to the task,” he said. “We have to master the transition to a parallel programming environment, with highly distributed, concurrent systems. It’s nascent at this point but it’s required to achieve these capabilities.”
CAMBRIDGE, Mass.–Microsoft Chief Research and Strategy Officer Craig Mundie on Thursday offered a long-term view of where Microsoft and the world of computing are heading over the next few decades. Speaking at the MIT Emerging Technology Conference here, Mundie envisioned a 3D virtual world populated by virtual presences, using a combination of client and cloud services.
(Credit:
Microsoft)
Mundie gave a few examples from Microsoft Research to illustrate the concept of spatial computing. In a few months, the compay plans to test a new virtual reception assistant in some of its campus buildings. The assistant, which takes the form of an avatar, helps schedule shuttle reservations to get people to various locations across the 10-million-square-foot Redmond, Wash., campus.
“Our tools are not designed to address this level of system design,” Mundie explained. “We have to see a paradigm change in the way we write applications.”
In another demo, Mundie offered a glimpse into the future of the live Web. He played out a scenario in which he was in an store, took a picture of a magazine cover on Northwest Indian art with his smartphone, and then placed the phone on a Microsoft Surface technology table when he got to his hotel. The pictures in the phone showed up on the surface table and he dragged them around. The system analyzed the image to determine how to use the photo as a way to pursue next steps in a virtual Web world. The system found the a digital version of the magazine and Mundie proceeded to explore magazine pages. From the magazine image of an art object, he went virtually to the store where the art object was on display.
Microsoft's prototype reception assistant system.
The system includes array microphones and natural language processing by which the avatar listens to the subjects and then interacts with them in real time. The system has been programmed to differentiate people by their clothing. Someone in a suit, for instance, would more likely be a visitor and not a potential shuttle rider.
In addition, creating a rich virtual environment that reflects the real world and is available to billions of people requires a lot of programmers. “If we want a million people to know how to do this, we have to mask complexity,” Mundie said. His goal is to program computers to have the equivalent of human senses that can operate well together. “That’s how we get to natural interfaces,” he said.
(Credit:
Dan Farber/CNET Networks )
Mundie categorized this demo as an illustration of the power of the client and the cloud in spatial computing. “You have to have a to-and-fro between local and centralized data services,” he said.
Programming tools, which have been a strength of Microsoft, will play a crucial role in the emergence of spatial computing. To create a kind of parallel universe–a cyberspace version of the physical world–everyone has to contribute on a continuous basis, Mundie said. Sensors and users will be generating trillions of bits of data, which requires addressing concurrency and complexity in a more loosely coupled, distributed and asynchronous environment, he said.