Quanta T-Party Project

Project T-Party will address the complexity of today's computing landscape. Whereas people currently have to maintain a wide array of "smart" devices -- from cell phones to computers to personal digital assistants -- on their own, TParty is intended to move such work into the background, making such computing tasks as upgrades and backups more or less invisible to the user.

The work means creating the next generation of platforms for computing and communication. TParty will require reengineering and an extension of the underlying technical infrastructure, the creation of new interfaces and the exploration of new ways of managing and accessing information.

In that sense, T-Party is the natural successor to CSAIL's Project Oxygen. It pioneered ''human-centered computing" innovations, such as reconfigurable microchips and voice-activated software, in an effort to make computers as invisible to users as the air they breathe, in the vision of MIT's late Michael Dertouzos, the project's founder. CSAIL researchers have worked with six business partners on that $30 million project.

The T-Party project was conceived in 2005, shortly after Quanta Chairman Barry Lam heard a talk given by Rodney Brooks, then-director of CSAIL, on the future of computing and communication. Inspired by Brooks’ presentation, Lam approached CSAIL and suggested that CSAIL and Quanta work together to imagine the future of smart devices for communication and information access. T-Party’s research has since moved beyond the original statement of work, which was focused on the development of the next generation of platforms for computing and communication. According to Chris Terman, T-Party research director, “Today, more than three years into this venture, T-Party is less about computing in the sense of operating systems or office applications or the chips inside a computer, and more about the human interface and device independence. We’re looking into how people access information, interact with one another, and build their own virtual world for work and play. We want to make it easy to build that world, share it with others, and connect it to devices that are currently nearby.” Breaking down the barriers The various devices used today each represent a domain – such as a home media center, an office workstation, a smart phone, or a camera – that is compartmentalized, with its own information architecture. These domains are often connected either directly or through the net, but interoperate awkwardly (if at all) because of security concerns or lack of foresight by the designers. T-Party’s goal is to break down the existing boundaries of these architectures and, by developing new universal standards for smart devices, create a single virtual information environment for each user. Independent of a particular device or even the location of the user, this approach would be much more human-centric and user-friendly, and far less cumbersome and restrictive. To this end, CSAIL and Quanta are working on developing devices that can serve our computing and communication needs, but in a more integrated fashion than imaginable today. A future goal, according to Terman, is to develop self-describing devices that can be adopted into the user’s virtual environment. “Imagine if your computer could read the instructions for your new LCD TV and immediately start to use it as an output device for many different types of audio and video streams. The user would only have to say, ‘Use that screen,’ and all the details would be handled automatically – this would be an enabling change for people.” Broad scope, fast results The T-Party project involves nearly 10% of all CSAIL members – PIs, research staff, and students – in areas ranging from systems, networking, security, and human language technology. The technologies they are exploring to support their vision fall into five categories:

(*) Connectivity: The development of direct, secure, authorized, and authenticated access to (mobile) personal devices; new protocols that increase the robustness and performance of wireless networks; networking and data-collection technologies for automotive platforms.
(*) Devices: Smart self-describing device interfaces incorporating an open-source Linux-based software framework.
(*) Applications: Multicore applications with trusted and reliable virtual computation and storage.
(*) Automation: ”Just Play” automated planning technology to build distributed applications on the fly; Web-based front ends.
(*) Natural interactions: The use of human language (speech and gesture) as a central ingredient in a multimodal interface; natural language summarization.

A number of T-Party technologies are finding their way into Quanta prototypes. They include automatic speech understanding technologies for automotive navigation, smart phones, and media centers; technologies for making cameras, smart phones, and similar devices more Internet-friendly; robust wireless advances that greatly improve the throughput of wireless networks; and medical monitoring devices to help doctors dramatically change the effectiveness of how healthcare is delivered.