Quotes of the Day

It was quite a sight. Last week at the post-doctoral science fair that is Microsoft's annual TechFest, Bill Gates was standing with his hands in his pockets, stomping his feet on the floor, staring at the image of the contents of an Outlook email inbox folder projected on the wall. But this was no temper tantrum about spam or inadvertently lost messages. Gates was trying out a demo that lets people do multi-limbed multitasking.

The device dubbed Step Mail is an electronic hopscotch-like floor pad by which users who have their hands full with other tasks at their desks can control their email with their feet. As an audience of assistants, members of the research team and star-struck employees watched, Gates, clad in a maroon crew neck sweater and khaki colored trousers, tried out the gizmo. Standing on the dance pad, he stomped on the down arrow and watched the cursor move down the list. His right loafer hopped on the right arrow until it moved up to the delete function marked by X and then he stomped forcefully and obliterated the entry.

What the Sundance Film Festival is to Hollywood, TechFest, now in its sixth year, is to Microsoft. It's where the company's 700 or so imagination entrepreneurs from the Microsoft Research (MSR) division have for the past six years unveiled their envelope-pushing inventions and futuristic prototypes for consideration by more than 6,000 company employees on the products side of the business.

Microsoft Research, started in 1991 as the first of its kind corporate computer science research lab, may well be more dynamic than any university computer science department. That's the way Rick Rashid, the senior VP in charge who was a professor at Carnegie Mellon, designed it . "For me basic research has to be first and foremost about moving the state of the art forward in computer science. We publish in peer reviewed journals just like professors so that our research is subject to the same purity of process that a university would have," says Rashid. "It's not about doing things that are necessarily going to turn into products—product groups do a pretty good job of that. Researchers are about trying things when you have no idea if they work and seeing where the ideas take you."

Researcher A.J. Brush's team attracted attention for LINC, a demo of a digital calendar aimed at families who, what else, frantically multitask. Operated with a sleek stylus pen, LINC looks like and is as easy to use as a paper calendar that hangs on most kitchen walls, crammed with daily to-do lists and annotated with scribbled margin notes, asteriks and arrows. It's designed for use in the kitchen, but the data is downloadable to everyoneís cell phone as well as their work, school and home computers. The idea is to keep the entries legible while keeping mom, dad and all the kids on the same page, literally and electronically, all the time. Each person can add, switch or delete entries and the changes are automatically incorporated onto all of the calendars on the networked devices.

The Adaptive Technologies Group has developed a prototype that deploys a web-enabled wireless device with photo-taking function—ideally a cellphone camera—to search the web using images, rather than keywords. The wi fi camera sends images to the PC which uploads it to a server that then tries to recognize the picture by searching through a database of images. The server does all the search work and sends the results back to the user's cell phone. "They always say a picture is worth a thousand words," says team member Larry Zitnick. "So imagine I'm at the grocery store, if I'm looking at a product from the shelves, like cereal, and I want more information about it, I take my cell phone out, take a picture of the front of the box and up will pop more information about the product such as prices at other stores, manufacturers coupons, or detailed nutrition information. Then, as I'm eating the cereal, I see information on the back of the box on the King Kong movie. So I take a picture of it and it takes me to the king kong movie website. You can imagine this for all sorts of scenarios."

Anoop Gupta, corporate vice president for the unified communications group, oversees the exchange of ideas between the researchers and product specialists. He's a former tenured professor of computer science at Stanford who spent four years at MSR and also spent a year and a half in the coveted role of Gates' technical assistant. Gupta's job is to see that researchers get their ideas in front of the people in charge of the Windows Media Products that use that technology. Back in January of 2003 Gupta's group came up with a telephony prototype that connected the phone to the PC. When a call comes to the user's phone, it rings on the PC, and shows the user who's calling. A prompt allows the user to transfer the call to his mobile phone or home or office phone and then take the call.

The feature was so popular that within 6 months of installing the prototype on computers at Microsoft in January, over 7,000 employees were enthusaistically using it. By August of that year, the company executives decided to market it. Meantime, Gupta's MSR group co-authored a paper with a computer science professor from the University of Calgary that was presented at a leading conference on computer and human interaction, reporting on the development of the prototype and the results of user studies with the enhanced telephony device. The resulting product, Microsoft Office Communicator, which brings together email, IM telephony and other features on the PC, began shipping last May and is currently in the hands of more than 10 million users, he says. Microsoft's goal is to attract 200 million. If collaboration between research and the business side can keep producing that kind of growth at Microsoft, Bill Gates will most certainly have a lot of reasons to stomp his feet and dance on the floor, whether or not StepMail eventually makes it to market.

• SONJA STEPTOE
• At the company's annual TechFest, researchers unveil their latest inventions and prototypes