[voiceinteraction] potential properties for data messages between components

At the last call I took an action to propose some properties for inclusion in data messages transmitted between architectural
components.
Here are some ideas, based on the W3C Multimodal Architecture https://www.w3.org/TR/mmi-arch/  and EMMA
(https://www.w3.org/TR/2015/WD-emma20-20150908/)
We can review these in the call tomorrow.

Data exchange properties
1. From MMI Architecture
name of request
source
target
request id

2. From EMMA
a. ASR properties
tokens
lattice
language
confidence
timestamps
medium
mode
signal 
dialog turn

b. Natural language properties
interpretation(s)
id
grammar
model
alternatives (nbest)

Received on Tuesday, 7 September 2021 20:37:10 UTC