- From: Scott McGlashan <scott.mcglashan@pipebeach.com>
- Date: Fri, 22 Nov 2002 13:42:18 +0100
- To: "Jeff Haynie" <jhaynie@vocalocity.net>, <www-voice@w3.org>
The choice is up to the platform (or its ASR/TTS component), both en-uk and en-us are equally valid specifications of en. It can of course make its choice based upon whatever factors it thinks appropriate (locale may be one such factor). The usual problem people face is where the document is specified as en-uk but the platform only supports en-us. In VoiceXML, the languages are passed down the document tree and to the ASR/TTS (sub)documents. In SRGS, section 5.4 states "Conforming Grammar Processor may implement languages by approximate substitutions according to a documented, platform-specific behavior. For example, using a US English speech recognizer to process British English input.". A similar statement will appear in the next version of SSML for TTSs. Hope this helps, Scott Co-Chair, W3C Voice Browser Working Group -----Original Message----- From: Jeff Haynie [mailto:jhaynie@vocalocity.net] Sent: Thursday, November 21, 2002 17:03 To: www-voice@w3.org Subject: Language Identifier Clarification I would like to get better clarification on the use of ISO language identifiers from RFC3066 in xml:lang attributes of grammars, prompts, etc. In the case of using the valid "en" primary language subtag - and when 2 more more languages are supported, such as "en-US" and "en-UK", by the platform - and in the case no root xml:lang is specified, which one takes precendence? I would assume that the platform would make the choice, based on the locale preference of the location. Could the VBWG give further clarification on this logic and their precendence? Thank You. Jeff Jeff Haynie CTO Vocalocity, Inc. www.vocalocity.net 404-487-1200 x1316
Received on Friday, 22 November 2002 07:42:24 UTC