There was substantial progress in the IDNAbis (aka "IDNA2008") working group during the four hours (Monday PM and Tuesday AM), with total failure avoided and mapping on lookup added to the basic IDNA2008 model. Andrew Sullivan scribed day 1, Eric Brunner-Williams did day 2. The real test is whether the choice meets the needs of the Chinese, Japanese and Korean users. The Korean government has a position, and the Joint Engineering Team (lead by Japan) also presented.
The broad choice was
(1) keep 2008, unicode version independent, inclusive, with property-based rules, with lookup mapping suggested (try 2008, if that fails, try 2003) in an appendix
(2) keep 2008, unicode version independent, inclusive, with property-based rules, and add mapping to lookup only, not registration (aka "bundling")
(3) keep 2003, unicode version dependent, exclusive, with codepoint-based rules, with mapping, with specific fixes, e.g., updating the unicode version from 3.x to 5.x, security-specific changes, etc.
The choice was (2).
Peter Dengate-Thrush sat through all four hours. Monday evening he participated in the San Francisco ISOC Chapter founding wine-and-cheese party.
There was no DNS-EXT meeting. There has been a lively discussion of problems arising from the .gov signing. The DNSOP meeting was uneventful, there is a very good draft on timing of signing events and a good "DNSSEC Operations Guide" in draft.