INSIDE TALK : Computers Get to Be More Like Their Users : They can operate verbally, visually in ways that have electronic experts abuzz.
Electronics wizards are abuzz over new breakthroughs in some of technology’s most difficult and potentially important areas: the ability of computers to operate more like their human users do--verbally, visually and in other such fundamental ways.
Today, powerful and invaluable though they are, most computers still work in relatively rigid and abstract patterns. They have limited ability to match the flexible, free-flowing, associative manner of thinking and communicating that humans use routinely.
Until recently, for example, voice-recognition technology has been very primitive, able to discern little more than single-digit numbers and simple commands. Computers were easily confused by the ums and ahs that pepper many people’s speech. Similarly, computers have been maddeningly literal-minded, unable to sense what their users want or mean to communicate unless it is expressed exactly as prescribed by the programmer.
Now, insiders say, big steps forward are being taken in many of these areas--with potentially enormous economic and other benefits.
Recent watershed developments in software programs and other technology are enabling computers to recognize a particular speaker and 20,000 or more words.
Researchers predict that by the end of the decade, computers will easily handle much-expanded vocabularies. And they can envision eventual development of computers that can read lips, handwriting and recognize sign language--especially useful in noisy environments.
“There’s a critical moment in the life of a new technology when it suddenly catches on, and we’re at that moment for speech recognition,” says Bishnu Atal, head of the speech research department at AT&T; Bell Laboratories.
The breakthrough technology will mean far more than simply being able to order your computer to dial up a friend’s IBM PC and schedule a lunch date for the two of you.
Scientists predict major advances in the ability of computers to simulate three-dimensional objects or environments, enabling designers--for example--graphically to “step into” a proposed airplane cabin and try out alternative locations for a new galley or emergency door; such computer simulations could save millions of dollars in design changes that now occur after construction has begun.
Al Erisman, director of technology for Boeing Computer Services, a division of the Boeing Co., says all this is possible because of advances in high-speed computing that expedite the conversion of data into images.
Also ahead is an explosion in the growth of image communication, enabling engineers in different parts of the country to hook up their PCs to a central computer and simultaneously view changes that are being made either to text or to computer graphics.
Industry specialists say this could make it easier and faster to launch collaborative projects--such as deciding where to locate a fire wall in a prototype for a new car or how to arrange the room shapes and kitchen layout in a dream house before the carpenters begin.
The new technology also will make voice-recognition computers even more user-friendly--that is, easier for ordinary people to operate.
In one such computer at the Massachusetts Institute of Technology, for instance, you can type out the word hedake --deliberately misspelled but phonetically accurate--and the computer will display headache on the terminal screen, correcting the obvious spelling error.
There are bound to be some glitches before the new technology now emerging from the laboratories has been fully perfected. But scientists insist that they have turned the corner and are ready to apply their new technology to the marketplace.
AT&T;’s Atal says: “The only real limitation is our own imagination.”
