W3C

– DRAFT –
Voice Interaction

01 December 2021

Attendees

Present
debbie dirk jon
Regrets
-
Chair
debbie
Scribe
ddahl

Meeting minutes

changes to interfaces document

dirk: shows updates from today

https://w3c.github.io/voiceinteraction/voice%20interaction%20drafts/paInterfaces/paInterfaces.htm

dirk: these are in Github

dirk: interface between client and service

dirk: what about the user identification, should that be a parameter or part of the metadata?

jon: some use cases require that we know whether the user is authorized to use the service

debbie: we could postpone this decision

dirk: session id should be required
… if the system requires it there should be a session id

debbie: there could also be a request id

dirk: audio data could be endpointed or streamed

debbie: we need to be able to accomodate both cases

dirk: harder to encode streamed data, have to agree on codec

dirk: we should describe client activation strategy here
… we should decide how to use text that's common to both documents

debbie: we should try to use links if we can

dirk: this might require more work on the architecture document

debbie: we could work on the interfaces and then go back and update architecture

dirk: also changed CSS

debbie: will look into streamed audio data in EMMA

debbie: do we want to standardize low-level interfaces

dirk: we would lose context and history

debbie: should make it usable without those

EMMA

https://www.conversational-technologies.com/emma/emmaexample.html

debbie: interpretations could be nbest

debbie: where is the signal in 2.0

debbie: can we include signal in the document?

debbie: we need to find out if there's an XML format for binary inputs
… that would include images
… InkML could cover ink inputs
… also EmotionML could be part of an EMMA document
… location information
… we can also incorporate Geolocation

dirk: still need to do some research on audio streaming

debbie: there is an audio working group
… some of their work might be useful for representing audio in EMMA

debbie: we can always say that we've enhanced EMMA for streamed audio

dirk: we could mention standards like EMMA, Geolocation, and Audio

debbie: should keep things agnostic to the underlying protocol

dirk: we should standardize names but not protocols

debbie: should we think about MRCP?
… Media Resources Control Protocol

debbie: communication between a voice browser and speech resources

dirk: not that common anymore

dirk: used in concert with SIP calls

ACTION: dirk to update interfaces document

ACTION: debbie to investigate binary XML

Summary of action items

  1. dirk to update interfaces document
  2. debbie to investigate binary XML
Minutes manually created (not a transcript), formatted by scribe.perl version 159 (Fri Nov 5 17:37:14 2021 UTC).

Diagnostics

Succeeded: s/investigate/debbie to investigate

Maybe present: debbie, dirk, jon