IETF
rtcweb@jabber.ietf.org
Wednesday, 13 June 2012< ^ >
martin.thomson has set the subject to: RTCWEB WG http://tools.ietf.org/wg/rtcweb/agenda
Room Configuration

GMT+0
[00:13:59] tterribe joins the room
[00:58:09] mreavy joins the room
[02:21:41] mreavy leaves the room
[02:41:04] mreavy joins the room
[02:51:41] mreavy leaves the room
[06:13:13] tterribe leaves the room
[06:22:07] burn joins the room
[06:36:08] Ted Hardie joins the room
[06:36:32] <Ted Hardie> The webex and the hangout should both be up now.
[06:46:45] tuexen joins the room
[06:47:41] JohnLeslie joins the room
[06:51:08] EKR joins the room
[06:51:38] Martin Thomson joins the room
[06:52:42] Cullen Jennings joins the room
[06:59:50] mreavy joins the room
[07:02:29] <Cullen Jennings> who was note taker ?
[07:02:34] <Ted Hardie> Keith
[07:02:38] Anant Narayanan joins the room
[07:02:40] <Cullen Jennings> Thanks Keith
[07:02:43] sal joins the room
[07:03:27] alan.b.johnston joins the room
[07:03:37] tterribe joins the room
[07:03:53] tuexen leaves the room
[07:04:19] lepinski joins the room
[07:08:18] mreavy leaves the room
[07:15:30] tuexen joins the room
[07:16:00] <Cullen Jennings> Ted is dealing with slides and I have a hard time seeing the floor camera in the room when we are looking at slides. Is there someone in the room that could help do the Q management if we need it ?
[07:16:40] TimPanton joins the room
[07:16:57] jesup joins the room
[07:17:13] <tterribe> My jabber connection has been pretty flaky, but I can try.
[07:17:50] <EKR > q+
[07:18:04] mreavy joins the room
[07:18:42] dimartinii joins the room
[07:18:52] <Cullen Jennings> thank tim:
[07:19:10] <Cullen Jennings> Current q is ekr
[07:19:25] <tterribe> I thought ekr just asked his question.
[07:20:02] <Cullen Jennings> I have huge lag on some of my IM
[07:23:03] <Martin Thomson> location of slides?
[07:23:16] <Anant Narayanan> https://www.ietf.org/proceedings/interim/2012/06/12/rtcweb/slides/slides-interim-2012-rtcweb-2-4.pdf
[07:23:21] <Martin Thomson> cheers
[07:23:37] <tterribe> On slide 12 of 18.
[07:24:45] <alan.b.johnston> What is the link to the directory with all the meeting slides?
[07:24:59] <tterribe> http://www.ietf.org/proceedings/interim/2012/06/12/rtcweb/proceedings.html
[07:25:19] Jonathan Lennox joins the room
[07:25:29] <tterribe> On slide 13.
[07:25:29] <alan.b.johnston> Thanks!
[07:25:53] <Cullen Jennings> q+
[07:25:57] <Martin Thomson> q+
[07:25:57] <EKR > q+
[07:26:02] <tterribe> q = Cullen, EKR
[07:26:40] <Cullen Jennings> I think Martin was in that q too - did that show up for you ?
[07:27:00] <tterribe> Oh, hmm, I didn't see it until just now.
[07:27:37] <Martin Thomson> nice. I guess that there are some synchronization issues.
[07:27:40] <Jonathan Lennox> I saw Martin between Culen and EKR
[07:27:59] <Anant Narayanan> Yeah, me too
[07:28:05] <burn> ditto
[07:28:31] <jesup> jabber has delivery issues...
[07:28:45] hardie joins the room
[07:28:46] <EKR > am I ahead of or behind Christer?
[07:28:52] <Martin Thomson> ahead
[07:29:04] Justin Uberti joins the room
[07:29:16] Anant Narayanan leaves the room
[07:29:18] Anant Narayanan joins the room
[07:30:00] <Cullen Jennings> q+
[07:30:46] tterribe leaves the room
[07:31:41] tterribe joins the room
[07:31:46] tterribe leaves the room
[07:31:53] <jesup> q+
[07:31:59] <Martin Thomson> q+
[07:32:38] <hardie> What is the current q state?
[07:32:48] <Cullen Jennings> christer, richard
[07:33:07] <Cullen Jennings> sorry christer, cullen jesup, artihn , richard
[07:33:29] <Cullen Jennings> s/artihn/marthin/
[07:33:53] tuexen leaves the room
[07:33:58] <Martin Thomson> I think that richard was there before I asked
[07:34:16] <Anant Narayanan> q = cullen, jesup, richard, martin
[07:34:21] <Cullen Jennings> q is christer, cullen, jesup, richard, martin
[07:34:32] <Cullen Jennings> anant, plase save me and take over the Q
[07:34:39] <Anant Narayanan> will do! :)
[07:35:06] <hardie> Thanks anant
[07:35:36] hta joins the room
[07:36:30] tterribe joins the room
[07:36:46] tterribe leaves the room
[07:37:03] tterribe joins the room
[07:37:22] hardie leaves the room
[07:37:24] <Anant Narayanan> q = jesup, richard, martin, christer
[07:38:56] sal leaves the room: Replaced by new connection
[07:38:58] sal joins the room
[07:39:08] <TimPanton> Louder please
[07:39:35] <TimPanton> very quiet on webex
[07:39:54] <burn> tim, is that better?
[07:40:01] <TimPanton> a bit.
[07:40:13] <TimPanton> thanks.
[07:42:00] <Spencer Dawkins> am i the only one who keeps hearing something like breathing in webex? i'm not seeing a second speaker, like someone who SHOULD be muted, but it's pretty distracting ...
[07:42:47] <Anant Narayanan> q = martin, christer, magnus
[07:43:23] <TimPanton> Spencer - no the breathing is louder than the speaker for me too.
[07:43:32] <hta> "Ultimately" is not a defined term. There is just "time N" and "time N+t".
[07:43:41] <burn> everyone, please mute
[07:43:46] <burn> (if you are not speaking)
[07:44:01] <EKR > q+
[07:44:03] <Anant Narayanan> q = martin, christer, magnus, paul, ekr
[07:44:04] <burn> chairs, can you announce this on the mic as well so all will be reminded?
[07:44:18] <Cullen Jennings> I'm not listening to webex but if you need me to mute someone on webex, I can.
[07:44:20] <TimPanton> Webex thinks that Kista meetingroom is speaking if that helps.
[07:44:30] <Cullen Jennings> that is the main meeting room
[07:44:46] <TimPanton> and callin-user2
[07:45:55] <burn> tim, there is nothing we can do about the Kista meetingroom. Should not be any breathing there, since no one is next to a mic, but there could be minor room noises
[07:46:32] <Jonathan Lennox> Probably want DTLS handshake as well as ICE to all peers
[07:47:15] <jesup> q+
[07:47:27] <Anant Narayanan> q = martin, christer, magnus, paul, ekr, richard
[07:47:36] <Anant Narayanan> q = christer, magnus, paul, ekr, richard, jesup
[07:48:47] <Martin Thomson> q+
[07:48:55] <Anant Narayanan> q = magnus, paul, ekr, richard, jesup, martin
[07:49:14] <Martin Thomson> q-
[07:49:33] <Anant Narayanan> q = magnus, paul, ekr, richard, jesup
[07:51:19] <Cullen Jennings> q+
[07:51:33] <Anant Narayanan> q = paul, ekr, richard, jesup, cullen
[07:53:24] <Jonathan Lennox> What I said: You've done the ICE exchange, so you have the ufrags.
[07:54:01] <Cullen Jennings> but only after you get the signalling do you have the remote ICE ufrags
[07:54:32] <Jonathan Lennox> True, but you can't get any media until you've done ICE.
[07:55:53] <EKR > q+
[07:56:07] <Anant Narayanan> q = jesup, cullen, ekr
[07:56:21] <Ted Hardie> There will be a chair insert after ekr
[07:56:45] <Anant Narayanan> q = jesup, cullen, ekr, \0
[07:57:29] <EKR > q = jesup, cullen, ekr, /dev/chair
[07:58:23] Magnus W joins the room
[07:59:49] <Anant Narayanan> q = cullen, ekr, ted, [richard]
[08:01:39] <Ted Hardie> Anyone in the jabber have skype and can proxy q requests from partha in skype? It's his only im facility, apparently.
[08:02:35] <Anant Narayanan> I'll get on Skype, send me his ID?
[08:02:39] <EKR > q-
[08:02:57] <Anant Narayanan> q = ted, richard
[08:03:13] <Ted Hardie> Thanks, I'll get it and send it to you.
[08:03:54] <EKR > What I was going to say is: if we *do* decide to support parallel forking, I'm not excited about cloning as the specific mechanism. I'd prefer something more factory like.
[08:05:40] <jesup> +1
[08:05:43] <Neil Stratford> +1
[08:05:45] <Justin Uberti> +1
[08:05:47] <EKR > +1
[08:05:51] <Cullen Jennings> I am sure I do not understand this problem
[08:06:11] <tterribe> I didn't raise my hand, either.
[08:06:34] <Cullen Jennings> +1 do it later
[08:06:35] <Neil Stratford> +1
[08:06:38] <EKR > +1 do it later
[08:06:38] <Justin Uberti> +1
[08:06:39] <jesup> +1 later
[08:06:49] <Martin Thomson> +1 to later, where later === heat death of the universe
[08:07:07] <hta> later >= heat death of the universe?
[08:07:37] <Martin Thomson> HTA: I'm pretty sure that it would be the trigger for that, so...
[08:07:42] <Cullen Jennings> at ietf "now" is aproaching heat death of ...
[08:08:51] <Ted Hardie> @anant sr_parthasarathi is partha's skype id; thanks again.
[08:09:59] <EKR > q+
[08:10:31] <Anant Narayanan> thanks Ted
[08:10:33] <Anant Narayanan> q = ekr
[08:11:58] hardie joins the room
[08:12:04] Gonzalo joins the room
[08:12:18] <Martin Thomson> q+
[08:12:45] <Anant Narayanan> q+
[08:12:50] <Anant Narayanan> q = ekr, martin, anant
[08:13:05] <Cullen Jennings> q+
[08:13:18] <Anant Narayanan> q = ekr, martin, anant, richard, cullen
[08:15:21] <Anant Narayanan> q = ekr, martin, anant, richard, cullen, paul
[08:16:10] hardie leaves the room
[08:16:10] <EKR > oops, q+
[08:16:27] <Ted Hardie> q=anant, richard, cullen, paul, ekr
[08:16:29] <Anant Narayanan> q = martin, anant, richard, cullen, paul, ekr
[08:18:05] <Ted Hardie> @anant, I think we can drop them from the front of the q once they start talking, but it may be a language-specific optimization
[08:18:27] <Ted Hardie> q=richard, cullen, paul, ekr
[08:20:40] <Anant Narayanan> q = cullen, paul, ekr, lennox
[08:20:49] Magnus W leaves the room
[08:20:50] Magnus W joins the room
[08:20:52] <Anant Narayanan> yes, I think it makes sense to drop the person currently speaking from the queue
[08:23:36] <Ted Hardie> Keith goes in the q, just before Richard
[08:23:38] <Anant Narayanan> q = paul, ekr, lennox, richard
[08:23:53] <Anant Narayanan> q = paul, ekr, lennox, keith, richard, christer
[08:28:10] <Anant Narayanan> Partha: "I agree with Paul as "Full" offer is preferred in the mid-session as answerer shall change the negotiation during mid-session".
[08:29:24] <Anant Narayanan> q+
[08:29:27] Magnus W leaves the room
[08:29:29] Magnus W joins the room
[08:29:41] <Anant Narayanan> q = lennox, keith, richard, christer, anant
[08:31:33] Magnus W leaves the room
[08:31:45] Magnus W joins the room
[08:33:44] <Ted Hardie> chair q interrupt after anant
[08:35:06] <Anant Narayanan> q = anant, \0
[08:38:51] alan.b.johnston leaves the room
[08:39:07] <EKR > The word from Webkit seems to be that they have stopped trying to reduce these points.
[08:39:42] <EKR > i.e., reduce the amount of browser fingerprinting info they leak
[08:40:24] alan.b.johnston joins the room
[08:40:47] <Martin Thomson> every new feature increases the surface area by one bit just by existing
[08:41:39] <Ted Hardie> remote please type "need +
[08:41:41] <Martin Thomson> need+
[08:41:42] <Cullen Jennings> +1 I think we need something like sip options
[08:41:45] <EKR > +
[08:41:51] <Justin Uberti> +0.5
[08:42:04] <Ted Hardie> type later + for not doing this now
[08:42:11] <EKR > justin, you'd like an API which gives like random answers some of the time?
[08:42:33] <Jonathan Lennox> EKR: just like OPTIONS!
[08:43:00] <Ted Hardie> Rough consensus to continue the design
[08:43:08] <Martin Thomson> is it a confidence thing where it is available 50% of the time, or is it simply half-arsed where you get half of the capabilities every time?
[08:43:34] <hta> martin, it will be consistent: You get correct answers to half the questions half the time. The rest of the response is randomized.
[08:43:37] <tterribe> I think you should get 75% of the capabilities 75% of the time.
[08:44:03] <tterribe> Well, sqrt(1/2), so 70.1%.
[08:44:17] <Martin Thomson> I was going to call you on that one
[08:46:43] jer_fuller joins the room
[08:49:02] <tterribe> You should have called me on the sqrt(1/2) being 70.7% instead.
[08:51:31] <Jonathan Lennox> Perhaps the answer is a quantum superposition of the right and wrong answers.
[08:51:57] <tterribe> But if we have an API to query the answer, we will change the answer.
[08:52:24] <EKR > It's like the randomized response technique—it's privacy preserving!
[08:52:26] <hta> We need a capability returned in the GetCapabilities call on whether the GetCapabilities call is available.
[08:52:39] <Martin Thomson> Jonathan: but it's always the wrong answer
[08:53:30] <jesup> I didn't even hear the question, so I definitely don't know the answer
[08:53:33] <hta> "All Cretans lie", said the Cretan.
[08:54:15] <Cullen Jennings> q+ once we get to end of slide
[08:54:24] <Martin Thomson> q+ same
[08:54:39] <Ted Hardie> @jesup Prior to getting consensus on the new "Full Offer" within the context of an existing offer, we asked who understood it. Only Richard said yes, so it goes back to the list. Richard will send text to the list, explaining his understanding.
[08:54:41] <Anant Narayanan> q = adam
[08:54:45] <Anant Narayanan> sorry
[08:54:52] <Anant Narayanan> q = cullen, martin, adam
[08:55:22] <EKR > q+
[08:55:32] <Anant Narayanan> q = cullen, martin, adam, ekr
[08:55:49] <Anant Narayanan> q+
[08:55:55] <Anant Narayanan> q = cullen, martin, adam, ekr, anant
[08:56:14] <Anant Narayanan> q = cullen, martin, adam, ekr, anant, dan
[08:58:05] <Anant Narayanan> q = cullen, martin, adam, ekr, anant, dan, hta, richard
[08:59:41] <jesup> q+
[08:59:50] <Anant Narayanan> q = martin, adam, ekr, anant, dan, hta, richard, jesup
[09:03:23] <Ted Hardie> q=ekr, anant, dan, hta, richard, jesup
[09:06:54] <Cullen Jennings> q+
[09:07:04] <Anant Narayanan> q = anant, dan, hta, richard, jesup, cullen
[09:13:42] <EKR > q+
[09:14:16] <Anant Narayanan> q = hta, richard, jesup, cullen, ekr
[09:14:33] <Spencer Dawkins> Umm, did we lose audio on webex?
[09:14:39] <Cullen Jennings> uh
[09:15:11] <Spencer Dawkins> back now ...
[09:15:19] <Ted Hardie> thanks for the warning
[09:15:20] <burn> q+
[09:15:29] <Anant Narayanan> q = hta, richard, jesup, cullen, ekr, burn
[09:15:35] <Cullen Jennings> I'm hearing audio on webex
[09:16:21] <Neil Stratford> Is there a use case here for migrating a peer connection between live pages? eg. moving a call out of an 'extension' window into a real pop-out page.
[09:16:29] <EKR > q ?
[09:16:37] <Spencer Dawkins> yeah, it's back. but i got silence, dropped and called back in, and still got silence for another 30 seconds or so.
[09:16:44] <Anant Narayanan> ekr you are already in queue
[09:16:58] <Spencer Dawkins> thanks ...
[09:16:59] <Ted Hardie> q=jesup, cullen, ekr, burn
[09:17:11] <Jonathan Lennox> Neil: do you want that at the mic?
[09:17:12] <dimartinii> +q
[09:17:31] <Anant Narayanan> q = jesup, cullen, ekr, burn, dini
[09:17:35] <Neil Stratford> Jonathan: yes please
[09:17:50] <Ted Hardie> lenox q+
[09:18:13] <Anant Narayanan> q = jesup, cullen, ekr, burn, dini, lennox
[09:18:15] <Jonathan Lennox> The q+ is relayed not me
[09:19:11] <Ted Hardie> understood, but it's you the token will pass to.
[09:19:24] <Jonathan Lennox> ok
[09:24:12] <Martin Thomson> if we're being pedantic, local people also need to announce their name too
[09:25:15] <tterribe> Martin: The scribe has complained to the local people, too.
[09:26:26] <Magnus W> Most are, the notetaker has difficulties to see who is talking in Cambridge
[09:27:56] <Anant Narayanan> q = ekr, burn, dini, lennox
[09:29:45] <Cullen Jennings> q+
[09:29:48] <Anant Narayanan> q = burn, dini, lennox, richard, cullen
[09:30:36] <Ted Hardie> Chair interrupt after cullen
[09:31:07] <Anant Narayanan> q = dini, lennox, richard, cullen,
[09:31:11] <Anant Narayanan> q = dini, lennox, richard, cullen, \0
[09:35:30] <Martin Thomson> I think that we need to signal chair interrupt differently, \0 indicates end of line, but I get the impression that chair interrupt is something else. Maybe just add "chair", or |n if you want to use a graphic.
[09:36:15] dimartini joins the room
[09:36:20] <Anant Narayanan> I assumed a chair interrupt meant end of line
[09:36:52] <EKR > maybe you should use \n to be end of line :)
[09:36:59] <Jonathan Lennox> U+2441 OCR CHAIR: ⑁
[09:37:38] <Anant Narayanan> ♥ the ⑁ symbol
[09:37:43] <EKR > q+
[09:43:51] <EKR > +1 hand
[09:43:52] <Cullen Jennings> +1 hand
[09:43:56] <jesup> +1 hand
[09:44:21] <hta> (new poll starts)
[09:44:24] <jesup> +1 hand
[09:44:24] <Justin Uberti> +1 (by default)
[09:44:26] <EKR > +1 hand
[09:44:33] <mreavy> +1 hand
[09:44:33] <TimPanton> +1 hand
[09:45:34] <Cullen Jennings> + I would use this
[09:45:40] <Justin Uberti> +1 would use
[09:45:40] <Neil Stratford> +1 would use
[09:45:44] <EKR > + I would use some variants of this but not others.
[09:45:45] <jesup> +.75
[09:45:53] <Martin Thomson> +1 ekr
[09:46:01] <mreavy> +1 would use
[09:46:47] <EKR > can someone post a link to the agenda?
[09:47:08] <Anant Narayanan> https://www.ietf.org/proceedings/interim/2012/06/12/rtcweb/proceedings.html
[09:47:21] <Jonathan Lennox> http://www.ietf.org/proceedings/interim/2012/06/12/rtcweb/agenda/agenda-interim-2012-rtcweb-2.txt
[09:47:24] <Anant Narayanan> well https://www.ietf.org/proceedings/interim/2012/06/12/rtcweb/agenda/agenda-interim-2012-rtcweb-2.txt actually
[09:47:51] <Jonathan Lennox> Not sure how well it corresponds to reality at this point
[09:48:31] dimartinii leaves the room: Disconnected: connection closed
[09:48:32] <Ted Hardie> Use the chair slides, rather than the agenda text files; the agenda forced textification that lost some state
[09:49:00] <EKR > q+
[09:51:19] alan.b.johnston leaves the room
[09:56:05] tuexen joins the room
[09:56:59] <EKR > h+
[09:57:59] <Cullen Jennings> ekr raised hand
[09:59:18] tterribe leaves the room: Offline
[09:59:24] Ted Hardie leaves the room
[09:59:28] tterribe joins the room
[09:59:38] hta leaves the room
[10:02:10] EKR leaves the room
[10:02:42] Neil Stratford leaves the room
[10:03:52] Neil Stratford joins the room
[10:07:18] tterribe leaves the room
[10:11:21] tterribe joins the room
[10:11:48] tterribe leaves the room
[10:15:19] tterribe joins the room
[10:15:48] tterribe leaves the room
[10:19:30] tterribe joins the room
[10:19:48] tterribe leaves the room
[10:23:12] tterribe joins the room
[10:23:18] tterribe leaves the room
[10:27:15] tterribe joins the room
[10:27:18] tterribe leaves the room
[10:31:17] tterribe joins the room
[10:31:18] tterribe leaves the room
[10:35:20] tterribe joins the room
[10:35:48] tterribe leaves the room
[10:39:22] tterribe joins the room
[10:39:48] tterribe leaves the room
[10:41:40] burn leaves the room
[10:41:47] burn joins the room
[10:43:25] tterribe joins the room
[10:43:48] tterribe leaves the room
[10:45:29] eburger joins the room
[10:47:28] tterribe joins the room
[10:47:48] tterribe leaves the room
[10:51:31] tterribe joins the room
[10:51:48] tterribe leaves the room
[10:55:34] tterribe joins the room
[10:56:51] hta joins the room
[10:58:33] EKR joins the room
[11:00:46] Neil Stratford leaves the room
[11:00:53] Neil Stratford joins the room
[11:02:01] Ted Hardie joins the room
[11:02:39] dimartini leaves the room
[11:03:31] <Spencer Dawkins> hoo-hoo!
[11:03:37] <Ted Hardie> thanks, spencer
[11:03:56] dimartini joins the room
[11:04:21] JohnLeslie leaves the room
[11:05:04] <Ted Hardie> I will be in full screen mode to enable the webex
[11:05:16] <Ted Hardie> Q management will be done elsewhere.
[11:05:28] <tuexen> webex works fine...
[11:05:29] <Anant Narayanan> I will do Queue management
[11:06:15] hardie joins the room
[11:07:31] <jesup> I'll get onto tsvwg and indicate support
[11:07:41] <dimartini> Am I the only one that get a lot of disconnects on the wifi ?
[11:07:47] <EKR > I'll read and comment
[11:07:52] <hardie> No
[11:07:54] <Anant Narayanan> No, mine's been disconnecting as well
[11:07:57] Magnus W is now known as MagnusW
[11:08:03] <Martin Thomson> the wifi is quite good
[11:08:44] <Jonathan Lennox> For me the wifi per se was good, but the DHCP server kept giving me an in-use IP address
[11:09:13] <dimartini> yes.. I had conflict too
[11:09:18] <hardie> I had to switch to wired to get away from that issue
[11:10:34] <Cullen Jennings> what is the hangouts URL
[11:10:52] <Anant Narayanan> plus.google.com/hangouts/_/google.com/rtcweb
[11:11:09] <hardie> You need to be invited that won't work by irself
[11:11:24] <hardie> Who needs the invite?
[11:11:44] <Cullen Jennings> Eric Burger needs invite
[11:12:13] <jesup> q+
[11:12:15] <hardie> I sent him one but I will do it again
[11:12:19] <Anant Narayanan> q = lennox, jesup
[11:12:25] <dimartini> I would like one too.if it is needed
[11:12:44] <Cullen Jennings> q+
[11:12:51] <Anant Narayanan> q = lennox, jesup, cullen
[11:13:01] <Martin Thomson> "the assumption is that the application knows at some level what it is doing" -- one of the funniest things I've heard all day
[11:14:32] <Anant Narayanan> q = cullen, richard, stefan
[11:14:50] <Martin Thomson> q+
[11:14:58] <Anant Narayanan> q = cullen, richard, stefan, martin
[11:15:53] <Anant Narayanan> q+
[11:15:59] <Anant Narayanan> q = cullen, richard, stefan, martin, anant
[11:16:25] <EKR > q+
[11:16:53] <Anant Narayanan> q = cullen, richard, stefan, martin, anant, ekr
[11:19:43] <Anant Narayanan> Ted, there is a request for you to send out an hangout invitation via jabber IM
[11:20:38] <Anant Narayanan> q = richard, stefan, martin, anant, ekr, christer
[11:20:41] <jesup> q+
[11:20:51] <Anant Narayanan> q = richard, stefan, martin, anant, ekr, christer, jesup
[11:21:27] roni_even joins the room
[11:21:40] <hardie> @anant to whom?
[11:21:54] <dimartini> me :)
[11:21:59] <Anant Narayanan> ^
[11:22:03] <hardie> Done
[11:22:54] <Anant Narayanan> q = richard, stefan, martin, anant, ekr, christer, jesup, hta
[11:23:59] <Anant Narayanan> q = richard, stefan, martin, anant, ekr, paul, jesup, hta
[11:25:57] <Anant Narayanan> q = paul, jesup, hta
[11:26:17] <jesup> Example pseudo-SDP:
m=data 12345 SCTP/DTLS 0  1          (or whatever the SDP identifier should be)
a=rtpmap:0 DIAMETER/0
a=fmtp:0 chan=0;random_diameter_parameter=1
a=rtpmap:1 BIDIRECT/1/*
a=fmtp:1 chan=1,name="foo",protocol="ftp",reliable,inorder;chan=2,name="bar",protocol="company/myproto",unreliable;chan=3,name="qwerty",protocol="SIP",partial_xmits,4

[11:27:01] <Martin Thomson> q+
[11:27:16] <Anant Narayanan> q = jesup, hta, martin
[11:27:16] <EKR > I think this should have colons instead of commas
[11:27:32] <Cullen Jennings> q+
[11:27:47] <Anant Narayanan> q = hta, martin, cullen
[11:27:55] <Spencer Dawkins> \i think the question is whether you could prevent coexistence from working ...
[11:28:15] Charles Eckel joins the room
[11:29:23] <tuexen> q+
[11:29:35] <Anant Narayanan> q = hta, martin, cullen, tuexen
[11:29:39] <EKR > q+
[11:29:44] <Anant Narayanan> q = hta, martin, cullen, tuexen, ekr
[11:30:12] <roni_even> so all the protocols will have a MIME type
[11:31:00] <Spencer Dawkins> roni -right. we've done this before - you more often than me.
[11:31:42] <roni_even> we have bfcp h224(far end camera control) and will probably have clue
[11:31:55] <Cullen Jennings> +1 roni
[11:32:11] <Anant Narayanan> q = hta, martin, cullen, tuexen, ekr, ronu
[11:32:15] <Anant Narayanan> s/ronu/roni
[11:32:26] <Anant Narayanan> q = hta, martin, cullen, tuexen, ekr, roni, christer
[11:34:40] <hta> Any time someone invokes the middle box as a reason a thing has to be done a particular way on a fully encrypted channel, can I invoke the Pope?
[11:35:02] <Cullen Jennings> no, middleboxes all look in fully encryped channels all the time
[11:35:29] <hardie> So does the pope, so I think it's fair
[11:35:42] <Cullen Jennings> Good point :-)
[11:37:00] <Cullen Jennings> q+
[11:37:17] <Anant Narayanan> q = roni, christer, cullen, richard
[11:39:43] <Martin Thomson> The only reason that we have this protocol is so that channels (stream pairs) can be assigned names without glare; and the reason we have that is a desire for an API that presents a duplex stream. This is a bizarre mapping. You can address this all at the application layer if you simply identify stream numbers. Names are window dressing.
[11:40:19] <Cullen Jennings> if the names where restricted to be 16 bit integers, could we get rid of all this ?
[11:40:26] <Martin Thomson> my point exactly
[11:40:47] <tuexen> But the channels need to be unidirectional then...
[11:41:06] <Martin Thomson> you can implicitly tie stream 1 in each direction together
[11:41:26] <Cullen Jennings> right
[11:41:39] <hta> Martin, one reason why people wanted names is glare handling.
[11:41:45] <tuexen> But they need the same settings (PR-SCTP policy, PR_SCTP policy value, priority)
[11:41:52] Martin Thomson_ joins the room
[11:41:52] Martin Thomson_ is now known as Martin Thomson__
[11:41:52] Martin Thomson__ is now known as Martin Thomson_
[11:41:52] Martin Thomson_ is now known as Martin Thomson__
[11:41:52] Martin Thomson leaves the room
[11:41:52] Martin Thomson__ is now known as Martin Thomson
[11:41:52] Martin Thomson is now known as Martin Thomson__
[11:41:52] Martin Thomson__ is now known as Martin Thomson_
[11:41:52] Martin Thomson_ is now known as Martin Thomson__
[11:41:58] <jesup> You need to avoid the two sides trying to open stream 1 for a new use at the "same" moment
[11:42:15] <tuexen> Correct.
[11:42:16] Martin Thomson__ is now known as Martin Thomson_
[11:42:16] Martin Thomson_ is now known as Martin Thomson
[11:42:16] Martin Thomson is now known as Martin Thomson__
[11:42:16] Martin Thomson__ is now known as Martin Thomson_
[11:42:16] Martin Thomson_ is now known as Martin Thomson
[11:42:16] Martin Thomson is now known as Martin Thomson__
[11:42:22] Martin Thomson__ is now known as Martin Thomson_
[11:42:22] Martin Thomson_ is now known as Martin Thomson
[11:42:22] Martin Thomson is now known as Martin Thomson__
[11:42:22] Martin Thomson__ is now known as Martin Thomson_
[11:42:22] Martin Thomson_ is now known as Martin Thomson
[11:42:22] Martin Thomson is now known as Martin Thomson__
[11:42:22] Martin Thomson__ is now known as Martin Thomson_
[11:42:22] Martin Thomson_ is now known as Martin Thomson
[11:42:22] Martin Thomson is now known as Martin Thomson__
[11:42:22] Martin Thomson__ is now known as Martin Thomson_
[11:42:22] Martin Thomson_ is now known as Martin Thomson
[11:42:22] Martin Thomson is now known as Martin Thomson__
[11:42:22] Martin Thomson__ is now known as Martin Thomson_
[11:42:22] Martin Thomson_ is now known as Martin Thomson
[11:42:22] Martin Thomson is now known as Martin Thomson__
[11:42:22] Martin Thomson__ is now known as Martin Thomson_
[11:42:22] Martin Thomson_ is now known as Martin Thomson
[11:42:22] Martin Thomson is now known as Martin Thomson__
[11:42:22] Martin Thomson__ leaves the room
[11:42:28] <Cullen Jennings> but what happen if they both open a channel called "one" at same time
[11:42:40] Martin Thomson joins the room
[11:42:40] Martin Thomson is now known as Martin Thomson_
[11:42:40] Martin Thomson_ is now known as Martin Thomson
[11:42:40] Martin Thomson is now known as Martin Thomson_
[11:42:40] Martin Thomson_ is now known as Martin Thomson
[11:42:40] Martin Thomson is now known as Martin Thomson_
[11:42:40] Martin Thomson_ is now known as Martin Thomson
[11:42:40] Martin Thomson is now known as Martin Thomson_
[11:42:40] Martin Thomson_ is now known as Martin Thomson
[11:42:40] Martin Thomson is now known as Martin Thomson_
[11:42:40] Martin Thomson_ is now known as Martin Thomson
[11:42:41] Martin Thomson is now known as Martin Thomson_
[11:42:41] Martin Thomson_ is now known as Martin Thomson
[11:42:43] <hta> Martin of many names.....
[11:42:57] <jesup> Right now you get two streams named "one"
[11:43:05] Martin Thomson is now known as Martin Thomson_
[11:43:05] Martin Thomson_ is now known as Martin Thomson
[11:43:05] Martin Thomson is now known as Martin Thomson_
[11:43:05] Martin Thomson_ is now known as Martin Thomson
[11:43:05] Martin Thomson is now known as Martin Thomson_
[11:43:05] Martin Thomson_ is now known as Martin Thomson
[11:43:05] Martin Thomson is now known as Martin Thomson_
[11:43:05] Martin Thomson_ is now known as Martin Thomson
[11:43:06] Martin Thomson is now known as Martin Thomson_
[11:43:06] Martin Thomson_ is now known as Martin Thomson
[11:43:06] <tuexen> Currently they end up in two channels with identical names.
[11:43:06] Martin Thomson is now known as Martin Thomson_
[11:43:06] Martin Thomson_ is now known as Martin Thomson
[11:43:06] Martin Thomson is now known as Martin Thomson_
[11:43:06] Martin Thomson_ is now known as Martin Thomson
[11:43:07] Martin Thomson is now known as Martin Thomson_
[11:43:07] Martin Thomson_ is now known as Martin Thomson
[11:43:07] Martin Thomson is now known as Martin Thomson_
[11:43:07] Martin Thomson_ is now known as Martin Thomson
[11:43:07] Martin Thomson is now known as Martin Thomson_
[11:43:08] Martin Thomson_ is now known as Martin Thomson
[11:43:09] Martin Thomson leaves the room
[11:43:38] Martin Thomson joins the room
[11:43:42] <Martin Thomson> this is messed up
[11:43:49] <dimartini> what is going on ?
[11:43:53] <Martin Thomson> speaking specifically about jabber room
[11:43:54] <Cullen Jennings> it seems that in most cases, two stremas named "one" would not be what I want and would be a PITA for the applicaiton to deal with glare
[11:43:54] <jesup> q+
[11:43:58] <Anant Narayanan> q = magnus, jesup
[11:44:00] <hta> If I don't send data, can I claim to have done it in zero time?
[11:44:08] <Martin Thomson> someone sent me a contact request and my client went whacky
[11:44:41] <Martin Thomson> in response to harald, it's bizarre to suggest that you create a glare scenario in order to address glare
[11:44:53] <Anant Narayanan> q = magnus, jesup, richard
[11:44:55] Neil Stratford leaves the room
[11:45:02] Neil Stratford joins the room
[11:45:20] <Anant Narayanan> q = jesup, richard, neil
[11:45:21] dimartini leaves the room
[11:45:48] <Ted Hardie> Is Keith channeling neil, or is he also in the q?
[11:45:56] <hta> TCP does not offer a mechanism for indicating what protocol is run over it. Is this a design mistake?
[11:46:05] dimartini joins the room
[11:46:07] <Anant Narayanan> sorry, it's keith in queue
[11:46:12] <Cullen Jennings> I bet he remembers to say his name
[11:46:13] <Anant Narayanan> q = jesup, richard, keith
[11:46:29] <Martin Thomson> hta, arguably, yes
[11:46:48] <hta> we don't even have a definition of what it means to have "a protocol that runs over the data channel".
[11:46:49] <Martin Thomson> hta, because without that, we're stuck with all these cross protocol attacks
[11:47:35] <hta> martin, what cross protocol attacks (remembering that we have a DTLS-certified connection)?
[11:47:51] <Martin Thomson> I'm talking about TCP
[11:48:54] <Martin Thomson> in this case, the likelihood of someone mistaking a DTLS secured stream as some other sort of DTLS secured stream is slim
[11:48:56] <roni_even> this is why we have payload type in RTP and they are unidirectional negotiated in SDP and are part of the fixed header
[11:49:16] <hta> ah, yes. Cross protocol attacks = "trick X into sending data to Y where X thinks the protocol is A, and Y thinks the protocol is B".
[11:49:19] Jonathan Lennox leaves the room
[11:49:49] <hta> It helps that X and Y have agreed to communicate in this case.
[11:49:56] <Martin Thomson> BTW, ICE depends on the fact that you have a signaling relationship - that's the key mitigation against cross protocol attacks. The fact that you are part of a specific, existing context cuts out a lot of those opportunities for confusion
[11:50:19] <Martin Thomson> hta, it depends on X and Y having agreed to communicate
[11:51:27] <Anant Narayanan> q = roni
[11:52:31] <hta> martin, referent for "it"?
the defense, or the attack?
[11:53:02] <Martin Thomson> the defense, sorry - the same "it" that you used in your previous comment
[11:53:07] <Anant Narayanan> q = adam
[11:53:19] TimPanton leaves the room
[11:53:23] Jonathan Lennox joins the room
[11:53:28] <hardie> q+
[11:53:31] <hta> ok, thanks. English is not an easy language :-)
[11:53:53] <Martin Thomson> it doesn't help that we are tending toward the abstract
[11:53:56] TimPanton joins the room
[11:54:07] <Anant Narayanan> q = adam, ted
[11:55:20] <jesup> q+
[11:55:35] <Anant Narayanan> q = adam, ted, person in red sweater, jesup
[11:56:12] <Anant Narayanan> q = adam, ted, person in red sweater, jesup, paul
[11:56:36] <hardie> q-
[11:56:49] roni_even leaves the room
[11:57:44] <Martin Thomson> that's markus isomaki in red
[11:57:49] <Anant Narayanan> thank you!
[11:58:06] <Anant Narayanan> q = markus, jesup, paul
[11:59:00] <Anant Narayanan> q = jesup, lennox
[11:59:25] <Martin Thomson> SDP is part of the API, therefore all SDP questions should be referred to the W3C
[11:59:30] <Anant Narayanan> q = lennox, \0
[12:03:13] <Martin Thomson> ted's back is really quiet
[12:03:21] <EKR > did ted speak his name?
[12:04:24] <Cullen Jennings> As hat said "I have seen the W3C and they are us"
[12:04:24] <Cullen Jennings> s/hat/hta/
[12:04:27] <Martin Thomson> q+
[12:05:20] <Martin Thomson> q-
[12:05:30] <Martin Thomson> for the record, I almost completely disagree with Cullen
[12:05:54] <Cullen Jennings> dam, that means I am likely wrong
[12:07:33] <hta> Next time, I would like to frame this question as "Here are the current documents. They document a protocol with a known set of properties. Here is a set of proposed changes. Should we make these changes, or not?"
[12:08:13] <Martin Thomson> hta, that's a really easy question to answer every time; unfortunately, it also means that nothing gets done; no proposal is every perfectly acceptable
[12:08:28] <Martin Thomson> s.every.ever
[12:09:21] <hta> Martin, usually it's possible to say "the document after the change is closer to acceptable than before the change" (make the change) or "the change would make it worse" (don't do it), even when neither version is entirely acceptable.
[12:09:44] <hta> But this time, I couldn't tell what the precise proposal for change was, so I was really floundering in trying to figure out what to respond to.
[12:09:44] roni_even joins the room
[12:09:56] <Martin Thomson> right, that's the sort of thing that I'm looking for - this proposed change is headed in a direction that I consider acceptable
[12:14:23] <Cullen Jennings> do we get a coffee break at some point ? <says cullen in begging tone>
[12:14:39] <eburger> I like that idea O:-)
[12:15:11] <Martin Thomson> 😼
[12:16:20] <Jonathan Lennox> The gap in the schedule on the chair slides appears to be at 15:00 CEST (i.e., 45 minutes from now).
[12:16:21] <MagnusW> Another 45 min
[12:17:02] <Cullen Jennings> Sounds good - I will count the seconds
[12:17:08] <Martin Thomson> gah, too long
[12:18:19] <EKR > q+
[12:18:44] <jesup> I'd like a consensus call for this to be adopted as a wg document
[12:18:45] <MagnusW> Martin your presenting before you get coffee
[12:19:10] <Anant Narayanan> q = ekr
[12:19:23] <Martin Thomson> good thing I don't drink coffee, so I'm obviously not in need of a break
[12:19:36] <jesup> q+
[12:19:48] <Anant Narayanan> q = ekr, richard, jesup
[12:19:52] <jesup> "this" == draft-jesup-data-protocol
[12:20:03] <Ted Hardie> @jesup this seems to need to wait until the wg discussion we just deferred to the list, at least to me and Magnus.
[12:20:25] dimartini leaves the room: Disconnected: connection closed
[12:20:36] <Ted Hardie> when you hit the q, can you explain why it needs to be taken up now?
[12:21:49] eburger leaves the room
[12:22:51] <Ted Hardie> we're cutting the q
[12:23:01] <Ted Hardie> @jesup gets the last word
[12:30:28] <Martin Thomson> point of process: we have very little agreement on this stuff. I'd prefer not to see a draft-ietf-anything with a proposal in it.
[12:31:37] tuexen leaves the room
[12:31:39] dimartini joins the room
[12:34:03] <Cullen Jennings> The draft is http://tools.ietf.org/html/draft-thomson-rtcweb-ice-dtls-00
[12:34:03] dimartini leaves the room
[12:34:56] dimartinii joins the room
[12:35:54] <Jonathan Lennox> Can you paste the slides URL?
[12:37:40] <Anant Narayanan> I don't think they were uploaded
[12:40:38] <hta> he said that they were part of the minutes from the last meeting...
[12:41:51] <EKR > q+
[12:43:21] <mreavy> ekr:+1
[12:43:25] <hta> nope, they're not there.
[12:43:32] mary.h.barnes joins the room
[12:44:38] <Cullen Jennings> slides at http://www.ietf.org/proceedings/interim/2012/06/12/rtcweb/slides/slides-interim-2012-rtcweb-2-7.pptx;
[12:45:05] <mary.h.barnes> Thanks!
[12:49:54] eburger joins the room
[12:49:59] <EKR > q+
[12:51:37] <Cullen Jennings> well, that will encourage people not to use SHIM
[12:52:11] Charles Eckel leaves the room
[12:53:13] <jesup> q+
[12:53:22] <Ted Hardie> q=richard, jesup
[12:53:45] <Ted Hardie> q=richard, jesup, lenox
[12:54:00] <Cullen Jennings> addign Justing to q
[12:54:07] <Cullen Jennings> adding justin to Q
[12:54:17] <EKR > q +
[12:55:15] <EKR > For people's reference the two features Justin and I are alluding to are snap start: http://tools.ietf.org/html/draft-agl-tls-snapstart-00 and false start: http://tools.ietf.org/html/draft-bmoeller-tls-falsestart-00
[12:55:35] <Cullen Jennings> q+
[12:55:43] <Anant Narayanan> q = lennox, ekr, cullen
[12:55:55] <EKR > anant: justin is in there somewhere
[12:56:00] <Anant Narayanan> sorry, I lost track of the queue, as somebody was wrong on the internet and had to be corrected!
[12:56:05] <eburger> +1 tp jl
[12:56:12] <Ted Hardie> q ends with cullen.
[12:56:26] <Anant Narayanan> q = lennox, justin, ekr, cullen
[12:56:30] <mreavy> jesup: +100
[12:56:46] <Ted Hardie> Please folks, focus. We have a resolution that the author will caucus with Jonathan and EKR and it will be individual.
[12:57:34] <MagnusW> Cullen regarding SHIM, I think there is a reasonable solution that will not add extra RTT.
[13:01:03] <eburger> yum!
[13:01:04] EKR leaves the room
[13:01:06] <Ted Hardie> See you in 15
[13:02:49] Gonzalo leaves the room
[13:03:27] gcamaril joins the room
[13:05:57] lepinski leaves the room
[13:06:26] lepinski joins the room
[13:08:08] EKR joins the room
[13:08:35] tuexen joins the room
[13:11:41] Neil Stratford leaves the room
[13:11:48] Neil Stratford joins the room
[13:19:11] <eburger> Does ICE go in my coffee?
[13:19:11] dimartinii leaves the room
[13:19:41] <Cullen Jennings> I wanted to stop blocking the camera view of EKR
[13:20:53] aboba@jabber.org joins the room
[13:21:05] dimartini joins the room
[13:21:14] <dimartini> k
[13:21:34] <Ted Hardie> k?
[13:21:52] <Ted Hardie> Do you need to ask a question?
[13:22:44] <dimartini> sorry just testing after a lot of reconnects
[13:22:49] <dimartini> :)
[13:22:51] <Ted Hardie> okay, no problem
[13:30:49] aboba@jabber.org leaves the room
[13:31:53] <jesup> q+
[13:33:56] <Martin Thomson> congestion feedback using RTCP can be conveniently blackholed
[13:34:02] eburger leaves the room
[13:34:30] <jesup> Right, which should cause the CC algo to clamp the sending rate down
[13:35:12] <Martin Thomson> faking CC feedback would be harder, of course
[13:36:14] hardie leaves the room
[13:36:28] hardie joins the room
[13:37:13] Spencer Dawkins leaves the room
[13:37:20] mary.h.barnes leaves the room
[13:37:26] Martin Thomson leaves the room
[13:37:29] Martin Thomson joins the room
[13:37:30] hardie leaves the room
[13:37:32] Spencer Dawkins joins the room
[13:37:55] hardie joins the room
[13:42:27] <Martin Thomson> q+
[13:42:51] <Ted Hardie> q=marting
[13:42:52] <Anant Narayanan> q = martin
[13:44:56] lepinski leaves the room
[13:44:59] <Cullen Jennings> q+
[13:45:40] mary.h.barnes joins the room
[13:45:48] lepinski joins the room
[13:47:35] <Anant Narayanan> q = cullen
[13:48:08] hardie leaves the room
[13:48:31] <Martin Thomson> q+
[13:48:40] <Anant Narayanan> q = martin, lennox
[13:48:47] <Martin Thomson> lennox first
[13:48:54] <Anant Narayanan> q = lennox, martin
[13:49:03] <jesup> q+
[13:49:11] <Anant Narayanan> q = lennox, martin, jesup
[13:52:01] <Anant Narayanan> q = magnus
[13:53:46] <Martin Thomson> q+
[13:54:11] <Anant Narayanan> q = martin
[13:54:54] <jesup> martin+1
[13:55:13] <Anant Narayanan> q = lennox
[13:55:25] <EKR > who is that?
[13:55:38] <Ted Hardie> Muthu
[13:56:52] <Cullen Jennings> q+
[13:57:29] <Anant Narayanan> q = cullen
[13:58:26] <TimPanton> isn't there is a battery life issue on mobile to reducing the timeouts ?
[13:58:55] <Ted Hardie> The key question is whether you would have been able to turn off the radio before and now cannot.
[13:58:55] <Jonathan Lennox> Do you want mic relay
[13:58:58] <Jonathan Lennox> ?
[13:59:09] <TimPanton> Let's see if it comes up.
[13:59:13] <jesup> q+
[13:59:22] <Anant Narayanan> q = cullen, jesup
[14:00:39] <Ted Hardie> At that, you will never turn off the radio while this is going. If there are real flows going, though, that is not a big deal; I'm also not clear that the time the radios would be off prior to the previous consent check is that big a deal.
[14:00:42] <Anant Narayanan> q = cullen, jesup, hta
[14:00:55] <Anant Narayanan> q = jesup, hta, ted
[14:00:58] <Justin Uberti> q+
[14:01:08] <Anant Narayanan> q = jesup, hta, ted, justin, lennox
[14:02:54] <Martin Thomson> RTO needs a minimum for the initial checks, but agree that zero is OK once the flow is up
[14:03:13] <Anant Narayanan> q = hta, ted, justin, lennox, keith
[14:04:32] <Anant Narayanan> q = justin, lennox, keith
[14:05:42] <Martin Thomson> q+
[14:05:48] <Anant Narayanan> q = justin, lennox, keith, martin
[14:07:18] aboba joins the room
[14:08:27] <Ted Hardie> q+
[14:08:42] <Anant Narayanan> q = keith, martin, ted
[14:10:30] <aboba> Mic: Agree with Jonathan on use of a liveness check/ICE restart at 5 sec... but use for consent (with automatic turnoff) at that interval is problematic.
[14:11:00] <Ted Hardie> @aboba that's not the proposal; the consent is on a previous slide, which is not at that rate
[14:11:09] burn leaves the room
[14:11:15] burn joins the room
[14:11:24] <aboba> Ted: OK, thanks for the clarification.
[14:11:36] <Martin Thomson> q-
[14:11:42] <Anant Narayanan> q = ted
[14:13:31] <Anant Narayanan> q = ted, richard
[14:14:26] hta leaves the room
[14:14:58] <Cullen Jennings> q+
[14:15:04] <Anant Narayanan> q = richard, cullen
[14:15:08] hta joins the room
[14:17:13] <Cullen Jennings> q- I will send it to the list, drop me from q
[14:17:19] <Anant Narayanan> q = nil;
[14:17:32] <Ted Hardie> Let's end the q here, and move on.
[14:19:10] <Ted Hardie> @aboba does this answer your previous objection?
[14:19:48] <Anant Narayanan> q = christer
[14:20:05] <Anant Narayanan> q = nil;
[14:20:40] <Martin Thomson> +1 to this proposal II
[14:21:46] <Ted Hardie> Muthu speaking
[14:22:20] <Martin Thomson> q+
[14:22:31] roni_even leaves the room
[14:22:43] <Anant Narayanan> q = martin
[14:23:20] <Cullen Jennings> q+
[14:23:27] <jesup> +1
[14:24:50] <Martin Thomson> q+
[14:25:16] <Ted Hardie> q=oscar
[14:25:16] <Jonathan Lennox> Not mic-worthy: "When either timer expires start a STUN transaction" is wrong. If you have a transaction live you shouldn't start another one. (This also tweaks the "If transaction was started by Tc, stop sending" rule.)
[14:25:31] <jesup> q+
[14:26:01] <EKR > lenox: totally right.
[14:26:15] <Anant Narayanan> q = oscar, jesup
[14:26:21] <Ted Hardie> q end
[14:26:22] <Cullen Jennings> ok chair closing Q after curren jesup
[14:26:44] <EKR > I looked at this algorithm and was like "I will be sorry if I try to cover the case where a transactin is already in progress"
[14:26:45] hta leaves the room
[14:27:11] hta joins the room
[14:31:50] <aboba> Bernard: I'm in favor of the proposal.
[14:32:36] <jesup> +1 in favor
[14:32:38] <Martin Thomson> in favour
[14:32:39] <EKR > +1
[14:32:42] <mreavy> +1 in favor
[14:32:43] <Justin Uberti> +1
[14:35:50] <hta> the WEBRTC WG demands that there be an explicit liaison statement detailing this requirement before it will take action on it .... NOT.
[14:36:19] <tterribe> ekr put some text on a slide. How more official does it get?
[14:37:34] <Martin Thomson> q+
[14:37:43] <Anant Narayanan> q = richard, martin
[14:40:04] <Anant Narayanan> Question on behalf of Partha: "As parallel forking is outside the scope of JSEP... how is it applicable?"
[14:40:54] <Cullen Jennings> q+
[14:42:04] <aboba> If Bob1, Bob2, Bob3 actually are legitimately responding to ICE, then is this an attack?? The attacker could be a proxy that supports parallel forking....
[14:44:11] <aboba> So the attack is that there is no authentication??
[14:44:22] <Martin Thomson> no authentication of Bob
[14:44:24] <Martin Thomson> yes
[14:44:42] <aboba> Can't that be solved by requiring u-frag/password?
[14:44:56] <Martin Thomson> That's what I think EKR is planning to suggest
[14:45:36] <aboba> Ah yes.. *unique* u-frag/passwords.
[14:45:41] <Cullen Jennings> chair cutting off q after richard
[14:45:43] <Anant Narayanan> q = richard, \0
[14:45:47] <Martin Thomson> Bob's ufrag is what Alice will have to use and it will have to be unique ...ish
[14:45:59] <aboba> Got it.
[14:46:43] <Martin Thomson> Re: jonathan's statement - this is true, though this has implications for rtcweb regarding the selection of ufrag, as we've discussed here
[14:47:11] tuexen leaves the room
[14:48:43] <Martin Thomson> q+
[14:49:12] <Cullen Jennings> q+
[14:49:19] <Anant Narayanan> q = martin, cullen
[14:49:55] <Martin Thomson> q-
[14:50:51] <aboba> SDP is not trusted....
[14:51:08] <Cullen Jennings> I need way more coffee and way more sleep :-)
[14:55:03] mary.h.barnes leaves the room
[14:55:35] <Martin Thomson> assertion could be a JSON object rather than a string
[14:57:24] <jesup> cullen++ with s/coffee/diet dew/
[14:58:36] <Martin Thomson> are there any restrictions on using curly braces on SDP lines?
[14:59:04] <hta> we could take a lead from the W3C and percent-encode the curly brackets....
[14:59:08] <Jonathan Lennox> Can we just turn all of SDP into base64ed JSON in a= lines?
[14:59:31] <Martin Thomson> puke
[14:59:38] Neil Stratford leaves the room
[14:59:40] Neil Stratford joins the room
[14:59:49] <hta> Jonathan, carrying SDP in SDP? Watch out, or you'll create an informational black hole....
[14:59:53] <Jonathan Lennox> Repulsive syntax in SDP? We can't have *that*.
[15:00:05] <dimartini> lol
[15:00:10] <Cullen Jennings> so we will have a base64 encoded veriosn of a base64 encoded assertion -
[15:00:32] <Martin Thomson> JWT is already base64 in base64
[15:01:17] <jesup> It's base64 all the way down
[15:01:23] <Jonathan Lennox> SDP att-value is byte-string, which is any byte except NUL, CR, or LF
[15:01:31] <Ted Hardie> escaped stringified jason structures and "readable" don't really belong in the same utterance
[15:01:37] <Martin Thomson> size=(4/3)^{n}
[15:01:47] <Cullen Jennings> q+
[15:02:01] <hta> the mapping of : ( to :( never seemed more appropriate!
[15:02:16] gcamaril leaves the room
[15:02:35] <EKR > lennox: in that case, we might be ok.
[15:03:23] <Ted Hardie> It can be a more general attestation, right?
[15:03:29] <Justin Uberti> q+
[15:03:43] <Martin Thomson> ted, that depends largely on the trust model, which hasn't been resolved
[15:03:44] <Ted Hardie> q+
[15:03:58] <EKR > ted: yes it oculd.
[15:04:09] <Ted Hardie> Well, I think the placement and the "jason structure" would at least handle Belay-like structures.
[15:04:17] <EKR > hardie: yes
[15:04:24] <jesup> ted++ on readable JSON
[15:04:32] <Martin Thomson> the more that I think about the trust model, the more I consider that this is a harder problem that I first thought, even for first party assertions
[15:04:45] <Ted Hardie> "I yam what aI yam"?
[15:05:06] <Ted Hardie> Wow, so tired can't even make pop culture references.
[15:08:41] <hta> and whether or not the size of the SDP blows up beyond the 64 Kbyte UDP limit...
[15:08:49] tuexen joins the room
[15:09:29] <Jonathan Lennox> Would it be interesting to have an identity assertion which was "you can actually trust the X.509 cert, it's signed by a well-known CA"?
[15:10:06] <hta> "you can actually trust the X.509 cert, it's the result of a Flame MD5 collision attack with a Microsoft certificate"?
[15:10:27] <Jonathan Lennox> That's a separate problem. :-)
[15:10:48] <Martin Thomson> I'm not sure that it is necessary to rely on this mechanism when you have a trusted domain on the other end.
[15:10:49] <Ted Hardie> Real Cryptographers (™) have touched this assertion. Tremble, and obey!
[15:11:01] <hta> sorry, not helpful. We should probably have identity assertions that leverage the CA tree.
[15:11:17] <hta> The other snarky version is "You can really trust this one, it's signed by DigiNotar".....
[15:11:21] <EKR > hta: we should be helpful
[15:11:24] <Martin Thomson> If the certificate offered is a valid certificate for a domain, then it should be OK to say that you are talking to that "domain"
[15:12:30] <Martin Thomson> the implications of having the IdP proxy load from cache are interesting when it comes to revocation
[15:12:51] <Martin Thomson> q+
[15:12:58] <hta> I think it's a truism that if the certificate validates according to one's validation procedures, the certificate is then as trustworthy as anything that's validated according to those procedures.
[15:13:10] <Martin Thomson> hta, agree
[15:13:46] <Jonathan Lennox> hta: but it's helpful to communicate which validation procedure you expect to work. (Normally I wouldn't bother checking the DTLS cert against CAs, since I'm expecting it to be self-signed.)
[15:13:53] <Ted Hardie> And this has the time of the attestation, right?
[15:14:03] <Martin Thomson> I would hope so
[15:14:06] <Ted Hardie> So you can determine whether it is within the validity period?
[15:14:42] <hta> woo, universal clock requirement ahead! (but validation periods are significantly longer than what you expect clock skew to be, even when someone's attacking your clock scheme...)
[15:15:31] <Ted Hardie> It doesn't have to be universal, because you check it with the attesting authority; unless they have experienced clock skew, it's not an issue, right?
[15:16:31] <hta> A live connection with the attesting authority will solve a lot of attacks. We can even ask it for a signed statement about what day it is.
[15:20:02] <Martin Thomson> it's not necessarily true that the fingerprint is fixed for the entire SDP either - you could use a different certificate for each transport. That said, it's not really all that useful.
[15:20:39] <Jonathan Lennox> I think we've only got a single fingerprint per m= line, modulo capneg. Is that what you mean?
[15:20:49] <Jonathan Lennox> You can have different fingerprints for different m= lines.
[15:21:02] <Martin Thomson> that's it - how the browser manages to do that is another question
[15:21:35] <Jonathan Lennox> You mean what the chrome looks like? "I've verified that the audio is from EKR and the video is from Martin"?
[15:21:54] <Jonathan Lennox> "According to Facebook and Google respectively"?
[15:21:56] <Martin Thomson> on many levels, this is challenging
[15:23:56] <Cullen Jennings> + backed
[15:24:01] <aboba> Bernard: -baked
[15:24:08] <mreavy> +baked
[15:24:09] <jesup> +baked
[15:24:11] <EKR > + baked
[15:24:18] <Cullen Jennings> + ekr is baked too
[15:24:18] <Justin Uberti> baked
[15:24:45] <aboba> not baked
[15:26:32] lepinski leaves the room
[15:31:57] <Cullen Jennings> q+
[15:33:18] EKR leaves the room
[15:33:20] Martin Thomson leaves the room
[15:33:30] Martin Thomson joins the room
[15:38:46] eburger joins the room
[15:40:15] <eburger> zzzz
[15:40:26] <aboba> zzzzzzzzzzzzzzzzzzzzzzzzzzz
[15:40:41] <aboba> -conscious
[15:41:29] <aboba> need coffee NOW!
[15:41:30] <Martin Thomson> +zzzzzz
[15:42:01] <hta> the chairs may be heading for a sharp finish in 18 minutes and 23 seconds.
[15:42:10] <eburger> Thanks
[15:43:36] <Anant Narayanan> ekr: comment about that last slide, we need to ask for a reason string
[15:44:18] <Anant Narayanan> in addition to a privacy policy, asking the developer to provide a reason string for why they need camera/mic access is legally enforceable (i.e. if malice was found later, legal action can be taken for misrepresentation)
[15:44:45] <Anant Narayanan> so we should add that to getUserMedia, but that's a W3C action item
[15:44:48] <eburger> @anantn: a catalog of IANA-registered strings or an opaque string?
[15:44:51] <Martin Thomson> anant, you cannot make that a blanket statement. it might be true in the US, but it's almost certainly not true elsewhere
[15:45:05] <Anant Narayanan> @Martin, true in the US & Europe AFAIK.
[15:45:09] <Anant Narayanan> @eburger: opaque.
[15:45:17] EKR joins the room
[15:45:56] <Martin Thomson> I heard 5i, which is an irrational number
[15:46:47] <mreavy> :-)
[15:47:22] <eburger> Thanks, and Awesome job finishing on time!
[15:47:23] EKR leaves the room
[15:47:30] EKR joins the room
[15:47:31] <Martin Thomson> actually, correction (blame tiredness)...i is imaginary, not irrational
[15:47:36] EKR leaves the room
[15:47:40] <Anant Narayanan> thanks all!
[15:47:51] EKR joins the room
[15:47:52] EKR leaves the room
[15:47:54] eburger leaves the room
[15:47:57] hta leaves the room
[15:48:10] tterribe leaves the room
[15:48:17] Ted Hardie leaves the room
[15:48:38] Jonathan Lennox leaves the room
[15:48:44] <aboba> Thanks everyone for finishing before I fell asleep on my keyboard!
[15:48:55] aboba leaves the room
[15:48:59] Cullen Jennings leaves the room
[15:49:01] Neil Stratford leaves the room
[15:49:03] Anant Narayanan leaves the room
[15:49:09] jer_fuller leaves the room
[15:50:18] burn leaves the room
[15:51:42] sal leaves the room
[15:52:20] mreavy leaves the room
[15:52:25] tuexen leaves the room
[15:55:53] EKR joins the room
[15:58:19] Justin Uberti leaves the room
[15:59:20] jesup leaves the room
[15:59:28] Martin Thomson leaves the room
[15:59:39] TimPanton leaves the room
[16:01:45] Justin Uberti joins the room
[16:03:10] Justin Uberti leaves the room
[16:03:58] EKR leaves the room
[16:05:09] MagnusW leaves the room
[16:17:23] Martin Thomson joins the room
[16:19:16] dimartini leaves the room: Disconnected: connection closed
[16:20:25] Martin Thomson leaves the room
[16:41:14] jesup joins the room
[16:51:21] jesup leaves the room
[17:43:30] EKR joins the room
[17:43:33] Martin Thomson joins the room
[17:45:21] mreavy joins the room
[17:50:51] mreavy leaves the room
[17:55:46] EKR leaves the room
[17:57:56] Martin Thomson leaves the room
[17:58:17] Neil Stratford joins the room
[18:00:03] EKR joins the room
[18:12:41] EKR leaves the room
[21:22:19] hta joins the room
[21:54:05] hta leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!