W3C

Web Real-Time Communications Working Group Teleconference

28 Apr 2015

Agenda

See also: IRC log

Attendees

Present
Harald Alvestrand ("hta"), Stefan Håkansson ("stefanh"), Dominique Hazael-Massieux ("dom"), Dan Burnett ("burn"), Adam Roach ("adamR"), Giridhar Mandyam ("gmandyam"), Vivien Lacourba ("vivien"), Cullen Jenning ("fluffy", +1.408.421.aaaa), Justin Uberti ("juberti", +1.425.894.aaee), Mathieu Hofman (+1.805.452.aadd), Göran Eriksson ("goranap", +46.1.07.13.aacc), Eric Rescorla ("ekr"), Milan Patel, Ashok Malhotra, Dan Appelquist ("dka"), Andy Hutton (+44.190.881.aagg), Erik Lagerway ("Lags", +1.604.562.aahh), Mike Champion, Shijun Sun, Tim Panton ("steely_glint"), Martin Thomson ("mt"), Peter Thatcher (+1.425.610.aajj), +1.214.475.aabb, +49.351.49.aaff, +49.351.49.aaii
Chairs
Harald & Stefan
Scribe
Stefan, Dom

Contents


[Scribe is Stefan]

[hta & dom: trying to identify all callers.]

Why we need to recharter

<dom> http://www.w3.org/2015/04/webrtc-rechartering.html

Dom: why? History: orig charter approved 4 years ago
... extended 2 more years
... reason rechartering process started back in Jan
... full recharter since the AC has made it clear they want groups to be reviewed regularly
... given four years without a recharter we decided to recharter
... draft sent to review in Feb
... to the AC
... At the end of the day the Director needs to approve the charter
... but based on input from WG, AC etc.
... The changes shown.

<vivien> [slide 4]

Dom: extend three more years
... more details on deliverables and dependencies
... also a WebRTC NG deliverable
... AC review: 2 formal objections
... dom worked with objectors to solve
... resulted in other objections
... since discussion on the WebRTC mail list

<vivien> [slide 5]

<vivien> [slide 6]

Dom: Peter T proposed a change
...new proposal: extend 2 years (not 3), no NG deliverable, add Liasion to ORTC CG, spelled out a bit working process in the WG

<vivien> [slide 7]

Dom: end date is indicative, can recharter before that
... e.g if we want to start on NG work before the charter end date

<vivien> [slide 8]

Dom: decision process shown.

<vivien> [slide 9]

Dom: consensu today vs. no consensus, objections, etc. (see slide for details)
... questions/comments?

hta: everyone understands the process etc.?

cullen: dom, did the changes directly reflect the objections made by two parties during the ac review?

dom: they were done to address the objections

hta: we had a dialogue with those objecting, and we then we developed proposed changes that we thought would address the objections

cullen: were the changes _only_ made to address the objections, or were there other changes made?

hta: two conditions, one: the objections we though would be addressed by charter changes were made, two: the charter must in the end make sense

hta: next point: chairs to say what they think about the charter process

[Scribe is now Dom]

Chairs perspective

hta: the charter defines the scope of the WG
... the scope of the WG is an API
... for basic real time communication capabilities for JS apps
... we all want this to get done ASAP but not before it's ready
... our first focus is to get a stable spec based on what we currently have
... we have to have a charter to be a valid WG
... the charter describes what we know we need to do
... we don't want to spend too much time talking about the future
... we're close enough to the end (hopefully) that we should be able to start discussing the post 1.0 future
... we don't want to prevent these discussions
... we tried to describe what we thought we wanted to do, without constraining other changes

<ekr> actually I don't necessarily agree with this

Mozilla's position

ekr: I'm not sure I agree with the approach HTA suggested here
... we can either focus on 1.0 and this charter should reflect the work on 1.0 and recharter for NG
... a version of that charter would beefing the current charter as suggested by Cullen
... The other approach is committing to 1.1/NG with a direction
... I sent a proposal toward that
... that integrates guidelines of what -NG would be
... I'm not comfortable with a hybrid charter that contemplates vaguely NG with an odd liaison to ORTC

hta: what do you find odd with a liaison to ORTC?

ekr: I don't think we should be encouraging multiple specs in this space
... we should not have two groups not collaborating
... the work should happen here
... I propose extending the current charter, or a new charter with specific 1.1 suggestions

Discussion

DKA: Dan Appelquist, Telefonica
... I wanted to first of all that in my capacity as co-chair of the TAG, we had a discussion last week about WebRTC and the rechartering situation
... 1st the TAG supports WebRTC as an important work
... I hope we can focus back on the work ASAP
... Both from the TAG and Telefonica perspective, we would like to see convergence between ORTC and WebRTC 1.1
... maybe this can be done now
... Telefonica didn't object to the proposals; we wouldn't object to any of the proposals on the table
... I don't see a problem with a 1.0 focus with a bit of a staging for 1.1
... that's how I read Dom's proposal
... I think we all have largely the same goals

Mike Champion: Microsoft is one of the formal objector; we worked with Dom and have been very happy with Dom's proposal
... good middle ground on focusing on 1.0 and opening the space for discussion on NG when there is more stability on 1.0 and clarity on the evolution of ORTC
... I do think it's important that ORTC and WebRTC have a lot of common
... the only disagreement is on the peerconnection API
... there is a lot of implementation experience on ORTC that is generating some better understanding of how the high level APIs interact with the lower level protocols
... I'm happy with Dom's proposal and hope we can just move on with it

Cullen: I don't think there is some middle ground — either we do 1.1, or we focus down on 1.0
... I think the time for 1.1 is now; I think ekr's proposal is the best option
... I don't see what Dom's proposal enables that we can't do now
... there is strong support toward an object-based API inspired by the ORTC group
... I don't support Dom's proposal
... is there some concern that ekr's proposal doesn't address?

goran: from Ericsson, the most important thing is to finish 1.0
... no particular preference between ekr [think this meant cullen's] or dom's proposal
... we can definitely live with dom's proposal
... but we need to finish 1.0

<ekr> The notes seem to kind of be missing cullen's comments

hta: [speaking as Google, not chair]
... the most important thing is to get work done as others have said
... integration is going to happen
... this WG has no control on whether ORTC exists or not
... it is a community group which has been granted a name and a space
... I think it would be strange to not acknowledge we talk and work with them
... My fear is that trying to figure the right guidelines for 1.1 will take significant time for wordsmithing
... I think we will be in much better situation to do so in 1/2 year

<Lags> +1 Harald

hta: I'm thus not comfortable with a 1.1-charter ­ — we don't know what shapes it should take yet

juberti: I agree with EKR's framing ­— focus on 1.0 and recharter, or do 1.1
... doing the work on 1.1 now will save us time later
... if we can't get there, then we can fallback on dom's proposal
... I think it would be a denial of reality not to recognize the relationship with ORTC
... clearly we don't want a mixed governance model between the 2 groups
... but recognizing the relationship sounds a good thing
... so would want try 1.1 charter, with fallback on dom's with maybe some massaging

ekr: first on the liaison's text
... would be silly to deny the inspiration from ORTC
... but I don't want to start receiving liaison statements from ORTC CG
... On 1.1 and visibility on what we should be doing — I thought what I've come up with was a reasonable summary of discussions I've had with several people

Dom: @@@
... from my conversation with the objectors, it was too early to understand on what 1.1 will look like

cullen: ekr, justin and I agreed on what the 1.1 should look like
... but if we can't agree with that, I think we should fallback on a more minimal charter update
... would like to get feedback from the ORTC community on ekr's proposal

juberti: if we decide to go down the 1.0 extension path, there has been a concern about competition between WebRTC and ORTC
... not referring to ORTC in the charter will make that perception endure
... we can argue on the specific words of that relationship
... but recognizing it is important
... and that would help avoiding these perceptions of competition

DKA: I agree with justin about the perception

<Zakim> dka, you wanted to suggest a middle way…

DKA: those of us having worked in this space know there is more overlap and agreement that this perception indicates
... that's one of the reason I supported the idea of having a liaison to ORTC
... not to create a joint decision process, but to send a message to the community that we are working this out
... that the technology is on the right track
... and that the disagreements are being worked out
... and the people making technical, strategic and business decisions can rely on that technology
... Given that there is some support here about including 1.1 in the charter, I'm wondering if there is a middle path — extend for 6 months, create a task force between ORTC and WebRTC to converge toward a charter

hta: we already extended the current charter, for how many months dom ?

Dom: [Scribe missed Dom's comment about the previous group extensions, but they were two extensions:
March 2013 (extension till Feb 2015), March 2015 (extension till June 2015)]

Dom: the advantage of a recharter instead of an extension is that you don't have this 6 months deadline, it was to give us more freedom on what the right timeline for 1.1 would be.

Andy: DKA talked about the message to the community as important
... the community is a bit split and see oRTC and WebRTC as alternative and competing
... the charter should clarify that the WebRTC WG is where the future of WebRTC is being decided
... if we refer to ORTC, we should clarify that

Goran: when it comes to ORTC and governance, it's clear for us that the API is defined in the WG
... but there is a lot of experience and inspiration to be taken from ORTC
... A more tricky question is nailing down 1.0
... I would like to get a lot more focusing on stabilizing 1.0 and scoping down the protocols / API surface

Mike Champion: part of the reason ORTC split off as a CG was because too much energy was expended on the SDP debate
... putting the two together makes sense, but we will be back to what we were initially
... moving energy away from 1.0 to debating various versions of the future
... my best guess is that Microsoft and the rest of the ORTC community could live with a 1.0 focused charter
... but the point made about perception applies
... that we want to show targeting convergence toward a single community
... thus still favour the proposal with a liaison to the CG
... I think getting the group to focus on 1.0 is most important

ekr: my proposal clearly states work on 1.1 would start post-CR
... the source of disagreement that led to the ORTC CG came from the role of SDP in the API
... from what I perceive, these disagreements have now been resolved and my proposal for the 1.1 principles reflect this
... would like to understand what the objections to these principles are

juberti: I just wanted to touch on the point that we don't know enough about RTCWeb to finish 1.0 and describe 1.1
... for 1.1, we want to give more control via the API surface

<fluffy> I agree with what justin just said

juberti: I think we understand better what we want for 1.1 than we knew when we started 1.0

AdamR: I agree with Michael's comment on the perception of ORTC/WebRTC
... my conclusion is very different though
... this should be addressed by folding ORTC into the WG
... I think the wording that EKR's proposal have does a good job of showing where the convergence happens

DKA: I definitely support merging the groups if that can be done — that would be a great message to send
... re distractions of having 1.1 discussions going on in parallel with 1.0 discussions — nothing prevent having different task forces or calls to avoid that
... having a single Wg would be more coherent

burn: I spend a lot of time talking about WebRTC at conferences, esp. about standards
... there is a tremendous benefit of being able to show a path forward
... combining the groups in a single group would be the clearest message
... there is great work going on in the ORTC CG — although I always have to refrain from speaking up too much about it to avoid mis-communications
... would like to get more from the ORTC group into WebRTC
... but the main group is the WebRTC WG
... re having parallel work streams, that can work if people are different — not clear that would work here
... I think all the charters say we should focus on 1.0 until CR
... just want to caution people about parallelizing work

ekr: agree on the dangers of parallelizing
... given the discussions I've heard, there has been support for my proposed charter
... so I would like to ask the chairs to take that up to ask for objections or consensus on it

DKA: nothing bad to say about EKR's proposal
... I have not been heard a lot in the WebRTC space and I recognize that
... some of this is coming from the fact that Telefonica is highly invested in this technological space
... we would like to see the community coming together
... we're betting our technology platform on it
... this to explain the context of my comments

Mike: I would like to remind people that EKR's proposal is less than 12 hours old
... I would be unhappy to go in that direction without more time
... putting the groups back together given the past tensions around SDP doesn't necessarily sound like a recipe for success
... but I'm not saying we would object to it — we need time to consider it

ekr: that's totally fair
... I think we have more consensus now on what the relationship with SDP now than we had before
... I'm more than happy to sit down and work out specific changes

Mike: I don't understand why you have the pen on the charter rather than dom

ekr: I'm trying to get agreement with interested parties

mike: but we had been working with Dom on getting something we thought was acceptable
... anyway, we'll need more time to look into this

juberti: agree the proposal is fairly new; but I think it's a substantive proposal to get convergence
... but would still be interested in getting the rough sense of the participants on it even if non-committal

goran: Ericsson would also need more time to look into this

goran: in particular, the description around SDP — we want to allow innovation in this space
... @@@missed

<ekr> goran: I do want to apologize for only getting this this so recently

<ekr> … and a major point of this text is to enable innovation without having it be in SDP

hta: with my chair's hat
... there is no charter that can say there is common control between the CG and the WG

<goranap> @ekr thanks for clarifying, :-)

hta: people seem to be reasonably happy with including a 1.1 definition along the lines of what EKR has proposed

<ekr> goranap: If it doesn't say that, please tell me, b/c that's what it's supposed to say :)

hta: the people that have been most active in ORTC seems to be indicating that they would be interested to continue the work under the WebRTC umbrella when we come to the stage of 1.1/NG
... and so we should try to add that
... and yes, we should not try to boil the ocean

<ekr> +1

hta: can people live with something going in that direction?

mike: the devil is in the details

<burn> +1 to hta

<ekr> I think they may have lived for more like 11 hours :)

mike: the ORTC has had the goals of allowing shims with WebRTC
... but the surface of the shim is what is controversial

<steely_glint> ekr - I'm really not keen on the 'low level' moniker - but it isn't worth butting in now.

<ekr> seely_glint: feel free to suggest some new text.

Dom: I am fairly sure I can ask the director for another 6 months extension, but we should then consider it as a strong deadline and we won't get another extension

<dka> Suggest if we do that (the 6 mo extension) we should commit to a task force with a call schedue to work on the next charter – with as much transparency as possible.

Dom: Ekr, your proposal doesn't include the -NG deliverable is that missing from your proposal?

ekr: yes I omitted it by mistake

<juberti> I think that is the right thing for a 1.1 recharter - finish 1.0 and then explicitly plan to merge ORTC as part of 1.1.

ErikL: I would like to mirror the comments that DKA made
... I would love to see the two groups coming togehter
... I'm not sure if this is possible
... I like the spirit's of ekr's proposal, but it's very fresh and the devil is in the details
... extending for 6 months and allowing 1.0 to make its way to CR
... and enabling to look at 1.1 after that
... but getting the 2 groups together would be ideal
... the liaison with the CG would benefit the WG without implying the decision of the WG would require the CG approval

EKR: one more thing: this is an attempt to distill the principles people agree upon
... regardless of whether we adopt this in the charter or not, I really want to get feedback on these principles
... because we will need to do so sooner or later

juberti: if we can't agree on these principles, we're much further apart than we though
... I'm not hearing anyone freaking about, although they need more time

Dom: @@@

harald: in 6 months we would need to have a new charter

<dka> my suggestion was agreeing a new charter (which merges the 2 groups) in 6 months - not to commit to a 1.0 CR in 6 months

harald: the chairs will thus look into a charter that has 1.1 in it

<dka> I am happy to commit my time to this.

<dka> thanks all

<burn> I am happy to help as well

<steely_glint> @ekr - just drop the words.

<burn> bye

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/04/29 11:45:46 $