W3C

- DRAFT -

voice interaction community group

21 Oct 2020

Attendees

Present
debbie, jon, dirk
Regrets
Chair
debbie
Scribe
ddahl

Contents


<scribe> scribe: ddahl

IPA architecture 1.1 changes

dirk: reviews minimum architecture

jon: explain context in more detail

dirk: could be use to follow up references
... dialogs are like apps

debbie: difference between dialogs and dialog management
... dialogs are domain/application specific components that do a task

dirk: where should a full-blown IPA be added?

jon: should IPA also be in the blue box
... IPA should be in both boxes

dirk: provider selection service could be just pure data
... another perspective is to just bypass the orange box

debbie: is there room for standardization there? or just the green box to the blue box

dirk: send and receive audio
... but if user deviates from specific application, what happens
... this would break the direct connection

debbie: you would have to rely on the IPA to know when to give up
... if direct connection fails, what happens?

dirk: then we would revert to the provider selection service
... need to have direct connection to hook in Google or Siri
... add IPA to other diagram and explain context box

debbie: suggest captions for the figures
... also check for alt text
... should have noinput and error
... is the Dialog Manager responsible for the Dialog Strategy?
... or should a Dialog be able to choose its own Dialog Strategy?
... based on the application?
... like a travel planning dialog lends itself to a frame-based strategy
... could the DM have the ability to follow different strategies?

dirk: yes, that would ease the developer's job

debbie: the dialog would have to say what kind of dialog it was
... like a VoiceXML form
... one and done could be modeled with a degenerate state-based dialog
... is the DM just responsible for the UX?
... DM and Dialog both contribute to the UX

dirk: DM defines the playground, but the Dialog implementation also is important

debbie: we could think of components as black boxes or talk about internal structure
... maybe we should still think of them as black boxes

dirk: just define interfaces

debbie: we could give examples
... like VoiceXML as a dialog

dirk: we should go one step further with this, more work on context, support for other IPA's, walkthrough for another IPA provider

debbie: last thing to do is review use cases
... should review abstract and Introduction

<scribe> ACTION: debbie to review intro material

<scribe> ACTION: dirk update document based on today's discussion

jon: will ask for review regarding context

debbie: context is always there in diagrams, but often poorly defined

jon: OVN will discuss next week

dirk: also added table of abbreviations

debbie: make sure to explain that NLG includes TTS as we used it here

dirk: need to add IPA

debbie: should eventually have references

Summary of Action Items

[NEW] ACTION: debbie to review intro material
[NEW] ACTION: dirk update document based on today's discussion
 

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version (CVS log)
$Date: 2020/10/21 16:01:42 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision of Date 
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Succeeded: s/boy/box/
Succeeded: s/use/used/
Present: debbie jon dirk
Found Scribe: ddahl
Inferring ScribeNick: ddahl

WARNING: No date found!  Assuming today.  (Hint: Specify
the W3C IRC log URL, and the date will be determined from that.)
Or specify the date like this:
<dbooth> Date: 12 Sep 2002

People with action items: debbie dirk

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)


[End of scribe.perl diagnostic output]