[ Go to September 1997 Table of Contents ]|
-- by Jim Forbes
We've seen Active Desktop, and we know about Web View. So what's going to be the next big thing from Microsoft? You could be talking to your computer screen pretty soon. And the computer might also learn quite a lot about you and put that knowledge to good use.
According to chairman Bill Gates, Microsoft's research and development folks are hard at work on social interface technology, which will include the use of speech recognition software and hardware, agent technologies and animated characters that react to user commands and help perform tasks. Gates recently predicted that speech will soon be the primary means of controlling a PC and working with software. He added that large chunks of the underlying OS will be dedicated to speech recognition and related functions.
'Speech will blossom'
"As a method of controlling and interacting with computers, [speech] will blossom," said Daniel T. Ling, Microsoft's director of research. "And, in order for this to happen, applications developers, as well as the people who write operating system software, will have to tune their software for speech."
This may take time. Systems with the horsepower needed-think 500-plus MIPS-are only now beginning to appear.
Still, the research is paying off. Products like Microsoft Speech should be integrated into some applications as early as this year. However, the company says it isn't entirely satisfied with current technology, primarily because it was designed for dictation systems rather than standard PCs operated through spoken commands. The researchers insist that the goal is to build overall systems that are easier to use, rather than just more sophisticated dictation programs. In real terms, that means the technology must support at least a 100,000-word vocabulary.
One new development is voice fonts, the ability to imitate someone's phrasing and enunciation. The inevitable next step is voice cloning; this should be supported in PCs using social interfaces.