Although Microsoft Research is only a part of the software giant's $6-billion-a-year R&D effort, the unit is churning out flashy projects designed to wow users and in the process help vault MSN into a far more competitive position.
On Tuesday, Microsoft opened the doors a bit, putting on a kind of science fair at its campus here, demonstrating 13 of its more promising technologies for an audience of high school students, Silicon Valley residents and journalists. Although the projects on display were positioned as a more or less random sample of what the research group does, it was hard to miss the underlying message: We're going to make MSN cooler than Google or Yahoo! (YHOO).
To be sure, any help MSN can get would be welcome. In the March quarter, for example, the Web portal brought in a relatively paltry $561 million, down from $581 million in the same quarter of 2005. And it swung from a profit of $102 million to a loss of $26 million.
Fair or not, Microsoft has never had the reputation of being a terribly innovative company. And it's difficult for a layperson to evaluate the worth of the company's research. But it certainly appears that Microsoft is mobilizing its considerable resources, and more importantly, its high collective IQ, to get back in the game.
Consider the problem of navigating the Web with a PDA or cell phone. The connection may be fast, and the screen may be bright and colorful, but the keypad is awkward at best, and often nearly unusable.
Voice navigation is one solution, but Microsoft is developing an elegant way to make text input on a handheld fast, easy and accurate. Say a user wants to browse for information on Secretary of State Condoleezza Rice. "The Wild Thing," in reality a complex algorithm linked via the Web to the MSN database, lets him or her type something like this: C* rice (the * is a wild card), and come up with a list of suggestions in order of probability headed by Condoleezza Rice. Using the rankings of various queries, the list even includes common misspellings of her name, as well as guesses that are much further afield.