- From: Yahn Teisseire <yahn@ukibi.com>
- Date: Wed, 2 Oct 2002 08:44:38 -0400 (EDT)
- To: <www-voice@w3.org>
- Message-ID: <0a0401c26a11$5b2b2e20$e9020a0a@paris.ukibi.com>
Hi,
I'd like to know if that's possible to build a grammar able to recognize
multiple keywords, for example, "offer John a cookie", "send Tabatha
flowers", without building a grammar like :
<grammar>
<![CDATA[[
(offer John cookie) {<action="offer"><contact="John"><object="cookie">}
(offer John flowers)
{<action="offer"><contact="John"><object="flowers">}
(send John cookie) {<action="send "><contact="John"><object="cookie">}
(send John flowers) {<action="send "><contact="John"><object="flowers">}
(offer Tabatha cookie) {<action="offer"><contact="Tabatha
"><object="cookie">}
(offer Tabatha flowers) {<action="offer"><contact="Tabatha
"><object="flowers">}
(send Tabatha cookie) {<action="send "><contact="Tabatha
"><object="cookie">}
(send Tabatha flowers) {<action="send "><contact="Tabatha
"><object="flowers">}
]]]>
</grammar>
because if you have a lot of actions and contacts, the grammar will be too huge to describe all combinations.
Thanks for your help
Received on Wednesday, 2 October 2002 10:20:36 UTC