Sharing voice application processing via markup

摘要

A system is described for processing voice applications comprising a client device ( 10 ) having associated data indicative of its computing capability. The system has access to a plurality of scripts specifying tasks to be performed in a voice-based dialog between a user and the system. The scripts are interpretable at a browser level. A server ( 20 ) selects an appropriate script for the client device ( 10 ) based on the associated data. An interpreter layer processes the selected script to determine a first set of instructions to be performed on the client device ( 10 ) and a second set of instructions to be performed on the server ( 20 ) for the dialog. Computation is thus shared between the client device and the server based on the computational capability of the client.