Scumbria, nearly everyone involved in this antitrust issue is talking about browsers, I believe the real issue is highlighted in this article. The competition and industry is scared to death of Microsoft embedding voice activation into the OS. Winning "standards", is the corporate battle cry of of the 90's. Undoubtedly, the nimble powerhouse Microsoft will come out on top in this battle too.
Bill Gates is an amazing CEO. Instead of crying foul, they should be studying his methods.
Michael
_____________________________________________________________________
Look who's talking about voice-recognition standards
October 30, 1998
PC Week via NewsEdge : NEW YORK--Establishing standards is shaping up as the next hurdle that speech- recognition vendors must vault before they see widespread adoption of the technology.
Sun Microsystems Inc. (Nasdaq:SUNW) released one such standard, its Java Speech API 1.0, this week at the SpeechTek conference here. The API, which (NYSE:IBM) IBM and Lernout & Hauspie Speech Products USA Inc. have already implemented in their respective speech-recognition engines, lets developers write Java applications that can take advantage of underlying speech-recognition technology such as IBM's ViaVoice.
Microsoft Corp. (Nasdaq:MSFT) is promulgating its own API--Speech API 4.0-- which supports C++, ActiveX or Component Object Model development environments. "The Java API vs. the Microsoft API is back to the same issue between C++ and J++," said W.S. "Ozzie" Osborne, general manager of IBM's Speech Systems business unit, who delivered the conference keynote Tuesday.
"We're investing a lot of development into Java because it is cross platform from an applications point of view," Osborne said.
Moving mainstream
The Java API will spur development of speech-recognition applications, which is the next step in moving the technology into the mainstream, pledged Andrew Hunt, Sun's principal investigator for speech applications, in Chelmsford, Mass.
But APIs are only one area that must be addressed in the speech-recognition arena. There is the whole interface paradigm," said Roger Matus, vice president of North American marketing for Dragon Systems Inc.
"[Microsoft's speech] API doesn't even come close to that issue. The real question is, how do you get a common way of doing things, and that's a broader issue than just speech," Matus said.
So what's the holy grail? That would be a universal interface for speech- enabled systems and devices such as VCRs and wearable computers, agreed most participants on a nine-member panel of vendors at the conference.
First things first
Before that can be achieved, there are other standards to tend to, according to Osborne. For example, the PC 99 hardware specification outlined by Intel Corp. and Microsoft needs to expand its audio definition to a stricter noise tolerance level, he said.
Interference from extraneous noise persists as a roadblock to speech- recognition applications, hampering their adoption in areas such as the factory floor, rental car check-ins and border patrols.
IBM is also working with the World Wide Web Consortium to develop speech extensions to Extensible Markup Language.
Another area is a standard for human languages. "We'll see that through APIs," Osborne said. "We'll define some of those interfaces, and some will stay proprietary."
The technology is already being more widely endorsed by PC OEMs, Osborne said. IBM has added speech capabilities to its Aptiva line of consumer PCs and has cut deals with Fujitsu America Inc., six of the top PC makers in China and Packard Bell NEC Inc. in Europe to bundle speech-recognition technology. Deals with more vendors are in the works, he added.
And more software, such as databases, will be speech-enabled. Philips Speech Processing, a division of Dutch consumer electronics giant Philips Electronics N.V., is reportedly talking to the top database vendors about using its speech- recognition engine, and IBM is working on adding speech capabilities to DB2.
"Using voice is a great way to access a database," Osborne said. |
