Notes, August 20

Present: Peter, Stefan, Harald, Adam (Dan busy, Erik on vacation)
> OK, this week I should be back on track....
>
> Overarching items:
> - Face-to-face. What are our pain points that we have to bring there
> (for 1.0)?
Adam will participate remotely. Peter, Stefan and Harald expect to be there.

Chairs will make a shared doc that we can edit for an ideas list.
> - Pushing a new media capture editors' draft
>
> Mediacapture
> ========
> Pull requests:
> #211 Allow device capabilities to be discoverable - we may back off this
> one.
We can merge this one. We'll get the capabilities of output devices as
soon as we define what they are (and we haven't yet).

> Output is the controversial part, methinks.
> #219 Persist deviceids only after gUM permission grant - seems to have
> consensus. Is it right?
Has been merged.
> #226 typedef MediaStreamError as an object - compound the ugliness of
> the legacy API?
Leave it to Dan. He'll get to it next week.
>   (Will Anne object to the return of MediaStreamError?)
> #231 Answer TAG self-review - does this belong in the doc at all? List
> discussion needed?

The security considerations should answer the questions, but not in
questionnaire form.
Peter will work on this.
Closing.

Adam will push a new version ASAP (with the 24-hour signoff thing).
>
> Issues:
>
> #118 Practical algorithm - HTA to discuss
> #127 Timeout - should classify as "not for #1"
> #189 getUserMedia({}) should be type error
> #193 Adaptive frame rate - HTA to discuss
> #198 MediaTrackCapabilities - see #211
> #212 Privacy/Security review - see #231
> #218 Persist deviceIDs - see #219
> #221 Unconnected tracks - my memory from IETF is gone. Need to refresh
> with Justin.

This was about addTrack(something) in order to get an RTPsender before
the user hits "accept" on an incoming call (we don't want to open
cameras before "accept".
Seems like a good face-to-face meeting.
AI HTA: Check with Pthatcher if he can present the scenario & possible
solutions in Seattle.

> #222 Clarify whether groupId is origin-unique - Stefan? (I think it is)
Yes, it is. Current discussion is "how do we specify it" (definition of
"browsing session").
Stpeter will write up a PR that does the smaller thing (say "it is"),
and file an issue for the larger thing ("where is browsing session
defined").
> #225 tracks vs sinks in section 5 - Dan, anything blocking a PR?
> #227 Update onoverconstrained event handler - Dan?
> #230 WebIDL reference needs updating - Dom seems to advise no action
> now, leave bug open.
> #232 Clone a track: "inherit" not defined (section 4.3) - I think it
> means "copy" in this case. We should change to "copy" (except for
> source, where it means "copy a reference to the same source").
Adam Bergkvist will make a PR for this one. It's IDL language.
> #233 "MediaStream.active" is wrong - I think this is strictly editorial
> (s/MediaStream.active/the active attribute of the mediastream object in
> 4.2.2), but someone must check, and also look for "this error repeats
> itself throughout the draft".
Adam will take care of it.

>
> WebRTC
> =====
> Pull requests: 13
>
> #29 mediadiscarded - Peter, any news?
Closing - will bring to list.
> #236 Operations array -> chain - Adam?
Adam's not happy with using an implementation-specific description.
Wants Chrome feedback.
AI HTA: Ping pthatcher.
> #237 ReplaceTrack - waiting for list discussion
Marked as "F2F topic".
> #240 Move certificate management section - Dan?
Peter will see if he can get it merged.
> #241 Transport objects - Adam? (List discussion seems to have died down,
> I think with rough consensus)
Adam has minor issues (see comments), but looks good. Another pthatcher
ping.
> #242 Remove SyntaxError on malformed ICE - need list discussion and
> assignee.
List discussion petered out after July 17 (IETF started...)
Stefan will read the list and see if he can draw a conclusion.
> #254 RTPEncodingParameters.maxBitrate - list discussion seems to say
> "this topic is complicated", but can we merge this one as "we're sure we
> need this one at least"? Stefan?

One nit needs fixing. We also need a definition of bitrate.
> #255 RTP Sender get capabilities - this removes the "get a blob of SDP
> that you can parse" function and adds some more pre-parsed functions. Is
> it controversial?
Most controversial part is representation of a MIME type (list discussion).
> #256 Fix prose on getStats() w/o selector - don't think this is quite
> right, but makes sense (t
Adam will take a look.
> #258 Codec parameters
Waiting for MIME name in getCap.
> #261 Language on ICE agent being finished - I think this is fine.
stpeter will take a look.
> #262 must -> MUST. Someone needs to review, then merge.
Martin Thomson says it's OK. Some misgivings can be addressed at another
time. Merged.

#266 Tidy up SetLocal/SetRemote language (Adam)

No functional change. Need a reviewer. Perhaps Cullen?
Pinging him.

>
> Issues: 39
>
> I'm out of time for writing agenda. Will return to these later.
>
>
>
>
>


-- 
Surveillance is pervasive. Go Dark.

Received on Thursday, 20 August 2015 15:09:48 UTC