« IT Services Unprofitable for CGEY   |   Main   |   Oracle ruling news analyses »

IBM open sources 2 chunks of speech-recognition code

The New York Times has a piece entitled Speech Code From I.B.M. to Become Open Source, where they report IBM is donating two separate blocks of code related to speech recognition to open source groups. “IBM is donating code that it estimates cost the company $10 million to develop. One collection of speech software for handling basic words for dates, time and locations, like cities and states, will go to the Apache Software Foundation. The company is also contributing speech-editing tools to a second open-source group, the Eclipse Foundation.”
The article continues to give additonal insight into IBM’s motivations for this gift. “”We’re trying to spur the industry around open standards to get more and more speech application development,” said Steven A. Mills, the senior vice president in charge of I.B.M.’s software business. “Our code contribution is about getting that ecosystem going. If that happens, we think it will bring more business opportunities to I.B.M.”
This is development is important because it will speed development of technologies to bridge the important gap preventing voice from being effective input to systems and a mode of interaction. Speech recognition and processing still have a long way to go and a huge variety in specific applications. Open source development excels in applications where a high degree of customization and tweaking are important: we’ll see the value when open source developers start hacking to bring speech recognition to such wide ranging things as hobbyist (home electronics), improved usability for the blind, and as well as general interactions with devices.