• Microsoft’s Speech SDK 5.0, a beta release of which is now available, enables developers to integrate speech generation and recognition into Windows applications, using software "engines" from Microsoft or other vendors. Microsoft is using technology from the SDK to speech-enable the Office suite. The SDK includes libraries and documentation for the latest version of the Windows Speech API, version 5.0, which is simpler and more fault-tolerant than version 4.0 but not backward-compatible. New features of the libraries include an XML-based grammar format for expected input phrases, and simpler access to compressed voice data. The SDK delivers test versions of Microsoft speech recognition engines (for Simplified Chinese, U.S. English, and Japanese) and text-to-speech engines (for U.S. English and Japanese). Microsoft expects third parties will supply engines for the new API; Lernout & Hauspie is the only vendor that has announced support to date. The SDK requires Windows 98, NT 4.0, or 2000 with 64MB minimum memory (128MB for speech recognition). The minimum download size is 112MB. For further information and a link to the download site, see www.microsoft.com/iit.