Attendees:
- Robert Brown (Microsoft)
- Dan Burnett (Voxeo)
- Patrick Ehlen (AT&T)
- Michael Johnston (AT&T)
[Milan Young (Nuance) was unavailable this week].
---
The call was 20 minutes. We spent most of the time on the second topic.
---
Topic 1: schedule check
Referring to our schedule http://lists.w3.org/Archives/Public/public-xg-htmlspeech/2011Jun/0002.html, the first draft of the protocol proposal is due next week. We have a rough draft now, but with lots of open issues. So I feel that we're probably about a week behind schedule.
We'll need a longer working session to work through the open issues. We're scheduled to do a status report in the XG's regular call next week. We'll take 15 minutes to do that, then use the remainder of the time to work through open issues.
---
Topic 2: issues with RECOGNIZE
We discussed the concerns people have found with trying to "shoehorn" the RECOGNIZE method to fit the various requirements/scenarios we have. I will redraft to incorporate some of this, mainly to introduce a more generic LISTEN method that works better for continuous recognition (migrating over the work Milan had done on RECOGNIZE), and is more customizable to a variety of scenarios. RECOGNIZE will remain as a one-shot recognition method. We can then review and see if we feel any better about this approach.
---