IBM has been demonstrating the speech navigation and voice dictation capabilities of the next version of OS/2 this week at Spring Comdex in Chicago.
The OS, code-named Merlin, which goes out to 10,000 beta users at the end of this month, will ship with a microphone lodged in a headset, which will contain a built-in noise-cancellation function to help screen out typical office background noise, according to Wally Casey, vice-president of client product management in IBM's software group.
Users will be able to use voice to navigate through applications. For instance, users can go through their email, saying "discard" or "send", or cruise the Internet by issuing commands like "scroll down", Casey says.
"Literally, the idea of people with their feet up, surfing the 'Net," is a reality with the voice capabilities of Merlin, Casey says.
Merlin will offer 97% accuracy in voice recognition right out of the box, Casey claims.
Moreover, Merlin does not require users to verbally go through menu hierarchies to access a particular function, according to Casey. For example, in viewing a multipage document, users can say "last page", as opposed to issuing the separate commands as reflected in the menus of "view" followed by "last page", Casey says.
Though Merlin can interpret navigation commands spoken in a normal voice, dictation must still be given with significant, unnatural pauses between words. Nonetheless, voice recognition is fundamental to what US and European users surveyed by IBM say they want, Casey says.
IBM has also announced that Merlin will ship with a CD-ROM containing every known device driver. Some people say that OS/2 was hard to install because of finding the right device driver, Casey says. "We've decided to change that once and for all."