Summary of decisions made during TPAC WebRTC sessions

This is the summary of decisions made during the sessions at TPAC as 
recorded by the chairs. It has been assembled from the minutes [1][2] as 
well as from private notes.

Anytime a reference to a slide is made, the slides to look at are those 
at [3].

We hope we recorded all decisions and got them right, but please review 
and comment where we did not or where we missed a decision. This is also 
an opportunity for those who could not participate to comment if there 
is any decision made that they are not happy with.

Stefan for the chairs

PS There were a lot of things that were discussed that are not mentioned 
below since we think no conclusion or decision was reached

webrtc-pc
=========
- webrtc-pc Issue #787 "Integrate RTCRtpTransceiver into set 
local/remote steps": will be resolved by referencing the relevant JSEP 
section

- webrtc-pc Issue #645 "public negotiation-needed flag as readonly": 
Will be closed without any action, however a new Issue on 'clarify 
exactly when negotiationneeded fires' will be opened.

- webrtc-pc Issue #746 "Need to specify what happens if 
'createDataChannel' is called with an invalid ID" will be addressed by 
picking a DOMError to report.

- webrtc-pc Issue #727 "removeTrack: throw exception if sender is not in 
connection's set of senders": decided to throw an exception

- webrtc-pc Issue #526 "NetworkError event is not defined and might not 
be needed": RESOLUTION: we will change our close event to match with 
WebSocket’s

- webrtc-pc Issue #799 "Unclear when a DTMFToneChangeEvent is fired with 
an empty string": The solution proposed in PR#807 will be implemented.

- webrtc-pc PR#784 "interToneGap and duration fixes"  will updated so 
that calling insertDTMF with duration and/or interToneGap will _not_ 
change the default need minor fixes and then be merged

- webrtc-pc Issue #714 "STUN/TURN OAuth token auth parameter passing": 
The webrtc chairs will reach out to the IETF TRAM WG to get clarity. 
NOTE: this has been OBE after the meeting as Justin has given input in 
the Issue

- webrtc-pc Issue #305 "Describe what happens when media changes": Of 
the four cases brought up only letterbox/pillar was discussed (the 
others handled by PR#624), the decision being that the UA should center, 
scale and crop (in that order) to make the video fit.

- webrtc-pc error reporting (see e.g. Issues #822 - #830): We will (in 
the absence of good WebIDL support currently) define our own error 
reporting after the model used for mediacapture-main

- webrtc-pc Issue #555 "Sort out requirements around IdpLoginError": 
Decided to go with option 1 (slide 68) based on the new errors to be 
defined (see above)

- webrtc-pc Issue #720/PR #738: The solution in PR#738 was liked and 
adopted. There was some discussion on making getting the fingerprint an 
async operation, but since generateCertificate() is already async, the 
reason was not fully understood at the meeting and the proponent is 
asked to file a new Issue

- webrtc-pc Issue #764 "Specify how an RTCRtpSender should treat an 
ended track": It was decided that an RTCRtpSender should treat ended, 
muted and disabled tracks in the same way: black (video)/silence 
(audio)/zero info (other types) should be transmitted. It was also 
decided that video tracks, when the RTCRtpSender is not transmitting 
anything, should render a "frozen frame" on the remote end if connected 
to a video element. The RTCRtpSender can be made to not transmit e.g. by 
setting RTCRtpSender.enabled = false for all encodings.

- webrtc-pc Issue #698 "JSEP/WebRTC mismatch on empty remote MID" will 
be resolved by referencing the appropriate section in JSEP

- webrtc-pc PR #624 "Upscale allowed": generally liked. Decided to make 
an enum, and other minor updates before adopting

- webrtc-pc Issue #812/PR #818 "RTCIceGatheringState definition": PR 
#818 liked modulo some word smithing. PR #818 will be adopted once updated.

- webrtc-pc issue #561 split of text between documents - agreed with 
split suggested on slides.


webrtc-stats:
============
- webrtc-stats issue #57 QP Statistic - a PR will be generated based on 
the proposal.

- webrtc-stats issue #61 Circuit Breaker (seeing from stats whether its 
trigger conditions were reached) - a PR will be generated based on the 
proposal

- webrtc-stats issue #62 ICE pacing - we’ll ask the IETF ICE WG for 
advice here

- webrtc-stats issue #25 / PR #43 Extension mechanisms - agreed that 
"living document with snapshots" is a procedure we can agree on, bar 
should be similar to IANA "expert review", Cullen to provide proposal 
text for actual procedures.

- webrtc-stats issue #23 / PR #59 Conformance language - agreed that the 
proposed approach is OK, and that we can remove section 8.5 (IANA) from 
webrtc-pc.

- webrtc-stats issue #20 action on non-supported stats - "omit" seems 
right. Stats without a value before some time has passed present a 
challenge (can’t detect non-support if value can be omitted for other 
reasons).

- webrtc-stats issue #26 what happens when things stop - this has 
already been decided (freeze).

- webrtc-stats issue #49 API version flag - we will not add this. Issue 
closed.

- webrtc-stats issue #295 - touches on "get stats on objects", which we 
haven’t specified so far. The "get it all" approach has advantages too. 
Guidance on slide seems reasonable, and will be made into a PR; 
per-object getstats is for future work.


[1] https://www.w3.org/2016/09/22-webrtc-minutes
[2] https://www.w3.org/2016/09/23-webrtc-minutes
[3] 
https://www.w3.org/2011/04/webrtc/wiki/images/1/1e/WebRTCWG-2016-09-22.pdf

Received on Thursday, 6 October 2016 06:52:44 UTC