Thanks for summarizing this. This matches my understanding, and as I see it the main thing we need to resolve to get a stable v1 spec is whether SessionDescription and IceCandidate should be dictionaries or objects. On Tue, Jun 26, 2012 at 5:18 AM, Stefan Hakansson LK < stefan.lk.hakansson@ericsson.com> wrote: > Here is an attempt from the chairs to summarize the status of several > on-going discussions: > > > Renaming to avoid namespace issues (Anant) [ACTION-43] > No appetite for creative solutions like container interfaces. > Observation that others drop many names in the global namespace. > Rough consensus around a prefix for types (most likely RTC or Rtc). > > > Async callbacks for SetLocal/SetRemote calls (Justin) > Initially this seems to be acceptable. Code that doesn’t set them will > still work. > Mainly for error handling purposes - perhaps a more elegant solution > can be found > > > New state names and explicit state diagram for State / IceState > New names and diagrams seem acceptable > > > Constraints -> option dictionaries for Create*/Set* > Discussion shows a need for both optionality and ordering - every > complexity that led to constraints also applies here. > We should keep them as constraints structures. > We should update the registration doc to say that it needs to specify > which API calls they make sense in. > > > Constraints/options for addStream/addTrack > Some options are more related to MediaStream/MediaStreamTracks, than > to session > Need to be defined > > > Format of SessionDescriptions and Ice candidates > These are currently objects > There has been a proposal to move them to dictionaries instead > No consensus yet > > Adding a DataChannel object [ACTION-49] > Proposal written up by Randell > Not clear what the benefits of introducing it is > > Comments/feedback most welcome! > > Stefan for the chairs > >Received on Friday, 13 July 2012 04:02:49 UTC
This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:17:30 UTC