IETF
httpbis
httpbis@jabber.ietf.org
Tuesday, November 5, 2013< ^ >
gmaxwell has set the subject to: HTTPbis Working Group | http://tools.ietf.org/wg/httpbis/agenda?item=agenda-87-httpbis.html
Room Configuration
Room Occupants

GMT+0
[00:02:19] mcmanus leaves the room
[00:55:46] geoffrey.cooper58@gmail.com joins the room
[00:55:54] mcmanus joins the room
[01:02:56] mcmanus leaves the room
[01:05:46] james.godfrey joins the room
[01:32:11] synp joins the room
[01:34:45] SM joins the room
[01:35:53] Andrew Biggs joins the room
[01:38:04] Andrew Biggs leaves the room
[01:39:28] frodek joins the room
[01:44:24] martin.thomson joins the room
[01:45:09] Andrew Biggs joins the room
[01:45:22] Eliot Lear joins the room
[01:45:24] hildjj joins the room
[01:45:36] loreto.salvatore joins the room
[01:45:37] barryleiba joins the room
[01:45:39] fielding joins the room
[01:46:04] <hildjj> http://tools.ietf.org/wg/httpbis/minutes
[01:46:20] <hildjj> Oh, nm.  he's using a different one.
[01:46:22] mcmanus joins the room
[01:46:23] <Eliot Lear> good evening from regency a in vancouver
[01:46:44] <Eliot Lear> i will scribe into jabber
[01:47:00] g.e.montenegro joins the room
[01:47:10] <Eliot Lear> if you want something repeated into the mic and are not in the room, preface your comment with "MIC:"
[01:47:14] <hildjj> http://etherpad.tools.ietf.org:9000/p/ietf88-httpbis-1
[01:47:19] <Eliot Lear> now on draft- −24 change summary
[01:47:20] ted.h joins the room
[01:47:24] Julian joins the room
[01:47:24] Julian leaves the room
[01:47:28] <hildjj> Eliot: I'll be happy to channel to the mic, so you can keep typing.
[01:47:31] martini joins the room
[01:47:34] lef joins the room
[01:47:35] <Eliot Lear> ack, joe
[01:47:35] <barryleiba> "Regency A" reminds me of a highway exit in Hayward, CA: "A STREET DOWNTOWN".
[01:47:35] <martin.thomson> all comments lacking "mic:" shall be ignored
[01:48:03] <Eliot Lear> julian now speaking
[01:48:03] john.levine joins the room
[01:48:03] npdoty@gmail.com joins the room
[01:48:03] <martin.thomson> Barry, it's Canada
[01:48:03] <SM> :-)
[01:48:13] <Eliot Lear> now on slide 2
[01:48:26] <Eliot Lear> now on slide 3
[01:48:31] <Eliot Lear> p1
[01:48:39] <martin.thomson> Though they should have made it "Regency, Eh"
[01:48:39] Hugo Salgado joins the room
[01:48:48] <Eliot Lear> going to slide 7
[01:49:07] <Eliot Lear> last call feedback was editorial
[01:49:10] <Eliot Lear> no individual feedback
[01:49:12] Dan Wing joins the room
[01:49:21] Hugo Salgado leaves the room
[01:49:33] <Eliot Lear> please review drafts!
[01:49:38] <Eliot Lear> and what are the next steps?
[01:49:39] martini leaves the room
[01:49:42] <Eliot Lear> barry coming to the mic
[01:50:14] <Eliot Lear> barry will give plenty of time for the iesg to review the draft set
[01:50:27] <Eliot Lear> won't be on a telechat until mid-december
[01:50:33] Hugo Salgado joins the room
[01:51:12] <Eliot Lear> there are some downrefs, but mark asserts no missing references
[01:51:41] <Eliot Lear> mark now:
[01:51:48] <Eliot Lear> last call ended officially today...
[01:51:55] martini joins the room
[01:51:56] <Eliot Lear> now looking at open tickets on the web
[01:52:40] <Eliot Lear> issue #507
[01:53:08] <SM> http://trac.tools.ietf.org/wg/httpbis/trac/report/22
[01:53:23] tlr joins the room
[01:53:29] <Eliot Lear> this is about consistency between SHOULD v. ought to?
[01:54:18] <Eliot Lear> roy now answering
[01:54:32] Julian joins the room
[01:54:33] Julian leaves the room
[01:54:47] Julian joins the room
[01:54:48] Julian leaves the room
[01:54:49] barryleiba leaves the room
[01:54:59] <Eliot Lear> p1 make MUST, leave p5 alone
[01:55:10] <Eliot Lear> (that was roy)
[01:55:43] <SM> You can have a security consideration and reference it from another draft.
[01:55:48] <Eliot Lear> 1st one is a security issue
[01:56:02] idiliod joins the room
[01:56:15] <Eliot Lear> 2nd one is questionable as to whether it is a security issue
[01:56:37] <Eliot Lear> what about a 206 response?
[01:56:57] <Eliot Lear> both will be made a MUST
[01:57:19] <Eliot Lear> now on content negotiation
[01:57:22] <Eliot Lear> issue 519
[01:57:42] barryleiba joins the room
[01:57:46] geoffrey.cooper58@gmail.com leaves the room
[01:58:09] <barryleiba> Is the mic Roy is on audible in the audio stream?
[01:58:13] <Eliot Lear> henry has studied whether 300 and 406 status codes are used, and in his environment he hasn't seen them
[01:58:20] <SM> The volume is low
[01:58:41] <barryleiba> But it works?
[01:58:52] <SM> Yes, I can understand
[01:58:57] <barryleiba> It sounds like nothing at all in the room.
[01:59:00] <hildjj> i can barely hear him and i'm 4ft behind him IRL
[01:59:04] <Eliot Lear> roy is not sure there is something to fix here
[01:59:14] <barryleiba> I'll ask him to speak closer to it next time.
[01:59:18] <Eliot Lear> mark: we are going to close this
[01:59:19] <SM> Thanks, Barry
[01:59:35] <Eliot Lear> issue 432
[01:59:40] <Eliot Lear> cachability of status codes
[01:59:47] <Eliot Lear> align text
[02:00:23] <Eliot Lear> Section 1.2.1 value MAX_INT?
[02:00:42] Dan Wing leaves the room
[02:00:54] <Eliot Lear> options: 1. leave as is
[02:01:12] <Eliot Lear> 2.  redefine to be 2^31-1
[02:01:26] <Eliot Lear> 3. make it undefined
[02:01:59] wseltzer joins the room
[02:02:03] <Eliot Lear> beware of historic issues if you change it
[02:02:29] <Eliot Lear> sorry we need to explain why it's still there
[02:02:29] <barryleiba> Roy's on a different mic now.  Is it better?
[02:02:37] <james.godfrey> yes
[02:02:43] <barryleiba> great
[02:03:37] <Eliot Lear> now long discussion of documenting the mistake but not changing it
[02:04:17] <Eliot Lear> phill now trying to explain logic
[02:04:33] <Eliot Lear> stuart: fix it to MAX_INT
[02:04:47] markh@jit.si joins the room
[02:04:58] <Eliot Lear> those are the open issues
[02:05:19] Dan Wing joins the room
[02:05:42] <Eliot Lear> mark: use of 2119 language
[02:06:07] <Eliot Lear> and we are not going to rathole
[02:06:50] <Eliot Lear> timeframe for moving on: close the next issues and publish −25
[02:07:06] <Eliot Lear> (next "week-ish")
[02:07:22] geoffrey.cooper58@gmail.com joins the room
[02:07:33] <Eliot Lear> then we will hope to be in the RFC Editor by January
[02:07:54] <Eliot Lear> now moving on to HTTP 2
[02:08:06] <Eliot Lear> Seattle review
[02:08:43] <Eliot Lear> nice list of implementations
[02:09:26] wseltzer leaves the room
[02:09:31] <Eliot Lear> plan is for −08 to be implemented after this week
[02:09:34] ted.h leaves the room
[02:09:44] <Eliot Lear> interop for a day
[02:10:11] <Eliot Lear> biggest changes are for HPACK
[02:10:19] <Eliot Lear> started discussions around other than upgrade
[02:10:53] <Eliot Lear> moving right along
[02:11:01] <Eliot Lear> what's new in HTTP/2.0
[02:11:15] martin.thomson leaves the room
[02:11:56] <Eliot Lear> now martin
[02:12:04] <Eliot Lear> "nothing touches the bow tie"
[02:12:37] <Eliot Lear> history of implementation drafts
[02:13:44] ted.h joins the room
[02:13:49] <Eliot Lear> state machine for streams, flow control good
[02:13:54] <Eliot Lear> clarification of errros
[02:14:01] <Eliot Lear> providing diagnostic information
[02:14:19] <Eliot Lear> errors in framing versus malformed http messages
[02:14:30] <Eliot Lear> settings negotiation improved
[02:14:52] <Eliot Lear> sm can you hear?
[02:15:38] <james.godfrey> can make it out, but not very loud
[02:16:08] <Eliot Lear> hoping to simplify header compression
[02:16:18] wseltzer joins the room
[02:16:55] <Eliot Lear> now on what remains
[02:17:17] <Eliot Lear> compression, connection initiation, prioritization, extensibility (or lack thereof)
[02:17:21] <SM> Eliot, the volume is low.  I can still hear
[02:17:45] martini leaves the room
[02:18:09] <Eliot Lear> mark now speaking
[02:18:15] martin.thomson joins the room
[02:19:03] wseltzer leaves the room
[02:19:13] <Eliot Lear> Now on issues discussion
[02:19:21] <Eliot Lear> Issue #1 upgrade mechanism
[02:19:47] <Eliot Lear> will chen:
[02:20:08] <Eliot Lear> beware of ratholing
[02:20:21] <martin.thomson> *will chan
[02:20:27] <Eliot Lear> 16 open design issues
[02:20:56] <Eliot Lear> so now on upgrade
[02:21:07] ted.h leaves the room
[02:21:09] <Eliot Lear> issues relating to ALPN are referred to TLS WG
[02:21:12] martini joins the room
[02:21:16] <Eliot Lear> for http that's us.
[02:21:45] <Eliot Lear> DNS or alternate-protocol:
[02:21:54] <Eliot Lear> mark wrote a draft
[02:22:11] <Eliot Lear> serves as input to optimistic encryption
[02:23:19] <Eliot Lear> draft-nottingham-httpbis-alt-svc-00
[02:23:25] <Eliot Lear> now speaking: mike bishop
[02:23:41] <Eliot Lear> doesn't think this should be in this spec
[02:23:52] ted.h joins the room
[02:23:56] martini leaves the room
[02:23:58] <Eliot Lear> and the server could do this
[02:24:36] martini joins the room
[02:24:50] <Eliot Lear> mark: there are performance implications
[02:25:22] <Eliot Lear> mike: you can already cache the state
[02:26:05] <Eliot Lear> stuart cheshire: clear elegant modular design requires separation of the protocol and how you upgrade to it
[02:26:14] <Eliot Lear> performance requires that you cross those boundaries
[02:26:30] <Eliot Lear> roberto peon:
[02:26:36] loreto.salvatore leaves the room
[02:26:45] <Eliot Lear> how one upgrades and to where one makes a connection
[02:27:30] <Eliot Lear> we're separating the identity of the origin from the host
[02:27:51] <Eliot Lear> martin thomson:
[02:28:01] <Eliot Lear> standing on one foot
[02:28:23] ted.h leaves the room
[02:28:44] <Eliot Lear> worried about mptcp and having middleboxes point things off in different directions
[02:29:00] barryleiba leaves the room
[02:29:10] julian joins the room
[02:29:10] julian leaves the room
[02:30:16] <Eliot Lear> mark: we may not drop upgrade on the floor;
[02:30:40] <Eliot Lear> mark: a number of implementations aren't supporting upgrade
[02:30:50] <Eliot Lear> dan druta:
[02:30:56] martini leaves the room
[02:31:01] <Eliot Lear> lack of transparency when switching protocol
[02:31:04] martini joins the room
[02:31:04] Julian joins the room
[02:31:57] <Eliot Lear> (not clear on his point)
[02:32:59] <Eliot Lear> mike bishop: difference between browser and stack
[02:33:09] <Eliot Lear> ted: lots of corner cases in upgrade
[02:34:24] martini leaves the room
[02:34:42] martini joins the room
[02:34:44] geoffrey.cooper58@gmail.com leaves the room
[02:34:44] <Eliot Lear> get rid of muck around level of indirection around origins/hosts
[02:35:01] <Eliot Lear> so separate that stuff out
[02:35:22] loreto.salvatore joins the room
[02:36:00] <Eliot Lear> ted: worried about checking the wrong thing [against a cert?]
[02:36:23] barryleiba joins the room
[02:36:51] <Eliot Lear> the issue is whether there are difference between certs
[02:37:39] <Eliot Lear> mark: fine with chopping out the indirection
[02:38:14] <Eliot Lear> ted: i want a way to deliver ONLY on http2
[02:38:29] <Eliot Lear> gabriel montenegro:
[02:38:40] ted.h joins the room
[02:38:48] <Eliot Lear> shares same concerns as ted
[02:39:39] <Eliot Lear> does this stuff buy us anything over upgrade?
[02:39:54] martini leaves the room
[02:40:00] martini joins the room
[02:40:25] wseltzer joins the room
[02:41:25] martini leaves the room: Replaced by new connection
[02:41:25] martini joins the room
[02:41:32] ted.h leaves the room
[02:43:07] loreto.salvatore leaves the room
[02:43:37] <Eliot Lear> will chen:
[02:43:57] <Eliot Lear> we have no plans to ever support upgrade
[02:44:08] <Eliot Lear> alternate-services is complicated and has lots of corner cases
[02:45:05] <Eliot Lear> chrome and firefox not interested in http2 in the clear
[02:45:14] <Eliot Lear> IE is interested in http2 in the clear
[02:46:05] barryleiba leaves the room
[02:47:57] barryleiba joins the room
[02:49:46] wseltzer leaves the room
[02:51:01] <Eliot Lear> should be standalone spec
[02:51:03] <Eliot Lear> (mark)
[02:53:19] <Eliot Lear> ted: please revise to chop out stuff
[02:53:32] <Eliot Lear> (the indirection stuff)
[02:54:04] loreto.salvatore joins the room
[02:54:06] <Eliot Lear> mark and patrick do a revision; martin to add a reference
[02:54:10] naoki.kambe joins the room
[02:55:25] <Eliot Lear> priority leveling
[02:55:48] <Eliot Lear> this is an outstanding issue- and we're going to talk about this in zürich
[02:57:21] <Eliot Lear> mike bishop: a proxy relaying requests from multiple clients, one browser uses max priority, the other halfway up, the proxy has to start mapping priorities, and there is an explosion of complexity
[02:58:38] <Eliot Lear> martin: this issue is going a bit long
[02:58:57] fielding leaves the room
[02:59:39] <Eliot Lear> will chen:
[02:59:52] martini leaves the room: Replaced by new connection
[02:59:53] martini joins the room
[03:00:08] martini leaves the room
[03:00:10] <Eliot Lear> we had a proposal- it's complicated.  we should get data why the complexity is worth it
[03:00:11] martini joins the room
[03:00:53] <Eliot Lear> will: we don't have data yet on it
[03:01:48] <Eliot Lear> mark: if this gets to the point where this is the only blocker, then we'll be in a bad place.
[03:02:29] <Eliot Lear> roberto: aim for something that might work in the draft now
[03:03:40] <Eliot Lear> issue 95: frame type extensibility
[03:04:36] Dan Wing leaves the room
[03:04:40] <Eliot Lear> please write proposals
[03:06:36] <Eliot Lear> roberto:
[03:06:56] <Eliot Lear> issue; we want the ability to experiment without requiring large changes
[03:07:03] <Eliot Lear> but we have a lack of an example extension
[03:07:47] <Eliot Lear> issue 292 cookie crumbing
[03:08:05] <Eliot Lear> james had an extension to do it
[03:08:40] <Eliot Lear> there was a fairly long conversation about how to do cookie crumbing
[03:08:53] <Eliot Lear> crumbling
[03:08:53] loreto.salvatore leaves the room
[03:09:11] <Eliot Lear> roberto:
[03:09:26] <Eliot Lear> how do we treat nulls in the value stream of the compressor
[03:09:38] <Eliot Lear> order comes up in this as well
[03:09:58] <Eliot Lear> mark: roberto, please raise an issue
[03:10:24] barryleiba leaves the room
[03:12:03] <Eliot Lear> martin: intermediaries have to preserve order
[03:12:13] naoki.kambe leaves the room
[03:12:16] loreto.salvatore joins the room
[03:12:46] barryleiba joins the room
[03:13:26] <Eliot Lear> julian: why are we talking about just about cookies?
[03:13:40] <Eliot Lear> mark: for other values of list notation, concat w/ nulls
[03:14:23] <Eliot Lear> roberto: there are two mechanisms: null mechanism, and when you toggle something on or off, you cause it to be forcefully emitted exactly where you want it to be.
[03:15:00] <Eliot Lear> roberto: we should assume that order doesn't matter and see what breaks
[03:15:35] <Eliot Lear> barry lieba:
[03:15:39] <synp> For now, if your server depends on the order of cookies, don't upgrade to HTTP/2
[03:15:48] ted.h joins the room
[03:16:06] <Eliot Lear> are you assuming that cookies would be treated differently between http2 and http1
[03:16:07] <Eliot Lear> ?
[03:16:19] <Eliot Lear> mike bishop: you have to for transition
[03:16:49] <synp> If it turns out that some servers really need it, add an "ordered-cookie" header and have them use that in HTTP/2
[03:17:12] <Eliot Lear> issue 266
[03:17:23] <Eliot Lear> close in favor of newly opened issue
[03:17:23] hildjj leaves the room
[03:18:40] <Eliot Lear> now alpn string/modifier needs to be defined for huffman
[03:20:48] <Eliot Lear> close it
[03:21:07] <Eliot Lear> alt reference sets
[03:21:35] <Eliot Lear> mike bishop:
[03:21:41] Dan Wing joins the room
[03:22:20] <Eliot Lear> hpack presentation
[03:22:22] <Eliot Lear> herve
[03:23:31] <Eliot Lear> your jabberer is turning into a pumpkin momentarily
[03:24:05] <SM> Thanks for the work, Eliot:)
[03:24:18] martini leaves the room: Replaced by new connection
[03:24:19] martini joins the room
[03:25:57] frodek leaves the room
[03:27:03] <Eliot Lear> welcomes, sm
[03:29:45] barryleiba leaves the room
[03:29:56] ted.h leaves the room
[03:31:06] idiliod leaves the room
[03:32:36] james.godfrey leaves the room
[03:35:00] martini leaves the room
[03:35:15] neilkatin leaves the room
[03:36:15] markh@jit.si leaves the room
[03:36:25] Eliot Lear leaves the room
[03:36:55] loreto.salvatore leaves the room
[03:37:16] tlr leaves the room
[03:37:18] john.levine leaves the room
[03:37:20] synp leaves the room: Computer went to sleep
[03:37:33] lef leaves the room
[03:37:34] Andrew Biggs leaves the room
[03:37:36] SM leaves the room
[03:37:58] martin.thomson leaves the room
[03:38:33] mcmanus leaves the room
[03:38:59] npdoty@gmail.com leaves the room
[03:39:30] Julian leaves the room: Computer went to sleep
[03:41:21] Dan Wing leaves the room
[03:42:34] Dan Wing joins the room
[03:43:37] Dan Wing leaves the room
[03:49:25] Hugo Salgado leaves the room
[03:52:02] g.e.montenegro leaves the room
[03:53:59] mcmanus joins the room
[03:54:34] mcmanus leaves the room
[03:55:50] mcmanus joins the room
[04:08:03] mcmanus leaves the room
[04:09:57] Bjoern joins the room
[04:25:33] Bjoern leaves the room
[06:33:13] lef joins the room
[06:36:30] lef joins the room
[06:46:28] npdoty@gmail.com joins the room
[06:47:34] lef leaves the room
[06:57:17] mcmanus joins the room
[06:57:29] mcmanus joins the room
[06:57:34] mcmanus leaves the room
[06:59:27] frodek joins the room
[07:07:18] Andrew Biggs joins the room
[07:11:35] mcmanus leaves the room
[07:16:27] npdoty@gmail.com leaves the room
[07:17:43] npdoty@gmail.com joins the room
[07:23:53] Andrew Biggs leaves the room
[07:37:11] g.e.montenegro joins the room
[07:38:06] loreto.salvatore joins the room
[07:38:32] Andrew Biggs joins the room
[07:42:43] Andrew Biggs leaves the room
[07:43:35] lef leaves the room
[09:13:36] g.e.montenegro leaves the room
[09:47:01] npdoty@gmail.com leaves the room
[09:47:19] npdoty@gmail.com joins the room
[10:47:10] npdoty@gmail.com leaves the room
[10:47:26] npdoty@gmail.com joins the room
[11:17:14] npdoty@gmail.com leaves the room
[11:17:33] npdoty@gmail.com joins the room
[12:32:24] npdoty@gmail.com leaves the room
[12:32:40] npdoty@gmail.com joins the room
[12:55:13] loreto.salvatore joins the room
[13:10:31] loreto.salvatore leaves the room
[13:32:36] npdoty@gmail.com leaves the room
[13:32:47] npdoty@gmail.com joins the room
[13:55:22] Eliot Lear joins the room
[14:16:54] tlr joins the room
[14:35:47] Eliot Lear leaves the room
[14:38:52] martin.thomson joins the room
[14:41:08] hildjj joins the room
[15:28:31] martini joins the room
[15:29:39] martini leaves the room
[15:29:47] martini joins the room
[15:33:45] Eliot Lear joins the room
[15:35:04] hildjj leaves the room
[15:47:38] frodek leaves the room
[15:53:01] tlr leaves the room
[15:53:28] Bjoern joins the room
[15:53:41] mcmanus joins the room
[15:55:28] loreto.salvatore joins the room
[16:03:54] martin.thomson leaves the room
[16:04:12] martin.thomson joins the room
[16:07:55] loreto.salvatore leaves the room
[16:10:05] martini leaves the room: Replaced by new connection
[16:10:13] martini joins the room
[16:11:06] martini leaves the room
[16:11:11] martini joins the room
[16:14:59] martini leaves the room: Replaced by new connection
[16:15:00] martini joins the room
[16:15:04] martin.thomson leaves the room
[16:18:16] npdoty@gmail.com leaves the room
[16:18:54] npdoty@gmail.com joins the room
[16:32:04] martini leaves the room: Replaced by new connection
[16:32:11] martini joins the room
[16:32:41] martini leaves the room
[16:32:47] martini joins the room
[16:34:36] martini leaves the room
[16:34:39] mcmanus leaves the room
[16:34:42] martini joins the room
[16:35:38] mcmanus joins the room
[16:45:17] Eliot Lear leaves the room
[16:48:07] martini leaves the room: Replaced by new connection
[16:48:16] martini joins the room
[16:49:39] mcmanus leaves the room
[16:50:06] tlr joins the room
[16:50:15] mcmanus joins the room
[16:52:06] hildjj joins the room
[16:52:19] ted.h joins the room
[16:54:26] naoki.kambe joins the room
[16:54:42] kmurchison joins the room
[16:55:06] SM joins the room
[16:57:15] kaduk@jabber.openafs.org/barnowl joins the room
[16:58:39] m&m joins the room
[16:58:47] wseltzer joins the room
[16:58:52] loreto.salvatore joins the room
[17:00:24] richard.barnes joins the room
[17:00:56] kohei.kasamatsu130 joins the room
[17:01:26] sean.turner@jabber.psg.com joins the room
[17:01:34] Tom Yu joins the room
[17:01:53] lef joins the room
[17:02:06] <m&m> etherpad at < http://etherpad.tools.ietf.org:9000/p/ietf88-httpbis-2 >
[17:02:17] Rohan Mahy joins the room
[17:02:19] <hildjj> Please prefix things you want relayed to the microphone with "MIC:"
[17:02:28] barryleiba joins the room
[17:02:38] martin.thomson joins the room
[17:02:39] Tim Bray joins the room
[17:03:04] semery joins the room
[17:03:32] <hildjj> if you are not in the room and you think there's a chance you may want to use this service, please direct message me your name so I can relay it correctly.
[17:04:52] synp joins the room
[17:05:11] jimsch1 joins the room
[17:05:22] sftcd joins the room
[17:05:59] Eliot Lear joins the room
[17:07:14] geoffrey.cooper58@gmail.com joins the room
[17:07:29] satoru.kanno@jabber.org joins the room
[17:11:29] loreto.salvatore leaves the room
[17:13:30] naoki.kambe leaves the room
[17:14:18] naoki.kambe joins the room
[17:15:29] loreto.salvatore joins the room
[17:15:39] <martin.thomson> RFC 4680
[17:15:52] frodek joins the room
[17:17:15] james.godfrey joins the room
[17:17:43] <martin.thomson> I will note that RFC 4680 suffers the same resumption properties as ALPN.
[17:21:12] stpeter joins the room
[17:24:18] tlr leaves the room
[17:24:52] tlr joins the room
[17:25:39] tlr leaves the room
[17:25:41] loreto.salvatore leaves the room
[17:27:56] martini leaves the room: Replaced by new connection
[17:27:57] martini joins the room
[17:28:01] geoffrey.cooper58@gmail.com leaves the room
[17:28:29] wseltzer leaves the room: IETF88
[17:28:43] martini leaves the room: Replaced by new connection
[17:29:08] martini joins the room
[17:29:08] martini leaves the room
[17:29:12] martini joins the room
[17:29:48] Dan Wing joins the room
[17:30:47] loreto.salvatore joins the room
[17:31:33] loreto.salvatore leaves the room
[17:32:47] synp leaves the room: Computer went to sleep
[17:33:51] loreto.salvatore joins the room
[17:34:20] richard.barnes leaves the room
[17:35:06] ted.h leaves the room
[17:37:11] naoki.kambe leaves the room
[17:38:00] cabo joins the room
[17:38:11] Larry Masinter joins the room
[17:38:26] <hildjj> Have you guys considered using the first HTTP header to add to the huffman tables?
[17:38:41] <cabo> The huffman table could have all digits at the same bit length.
[17:38:52] naoki.kambe joins the room
[17:39:01] <cabo> (Helps with PINs and credit card numbers.)
[17:39:02] <sftcd> i guess msisdns/phone numbers could be affected here & those are sent in headers in mobile n/ws (maybe less than perviously?)
[17:41:28] ted.h joins the room
[17:41:31] richard.barnes joins the room
[17:42:13] naoki.kambe leaves the room
[17:42:44] <martin.thomson> cabo: that's not a terrible idea
[17:42:59] synp joins the room
[17:45:16] naoki.kambe joins the room
[17:45:46] EKR joins the room
[17:45:54] <EKR> http://cdn.meme.li/i/pfxp3.jpg
[17:47:27] martini leaves the room
[17:47:34] martini joins the room
[17:48:14] naoki.kambe leaves the room
[17:48:37] tlr joins the room
[17:51:00] <stpeter> I assume the chair is scrolling through http://datatracker.ietf.org/doc/draft-ietf-httpbis-header-compression/
[17:51:03] m&m leaves the room: Disconnected: closed
[17:51:18] <ted.h> yes, the -04
[17:51:28] <sftcd> what can't you read that fast?
[17:51:36] m&m joins the room
[17:51:51] <stpeter> sftcd: not from the floor in the back of this little room behind a column
[17:52:08] 197a46ba19cde3e1 joins the room
[17:52:19] <sftcd> there are seats up here btw, 2nd row from front on left
[17:52:34] 197a46ba19cde3e1 leaves the room
[17:52:51] Mani joins the room
[17:52:52] Wolfgang Beck joins the room
[17:55:04] <hildjj> http://tools.ietf.org/html/draft-nottingham-http2-encryption-01
[17:56:03] wseltzer@jabber.org joins the room
[17:56:03] synp leaves the room: Computer went to sleep
[17:56:50] loreto.salvatore leaves the room
[17:58:47] npdoty@gmail.com leaves the room
[17:59:13] barryleiba leaves the room
[17:59:27] <stpeter> https://datatracker.ietf.org/doc/rfc6454/ = The Web Origin Concept
[17:59:33] loreto.salvatore leaves the room
[17:59:43] <Larry Masinter> there's some recent flap about getting Origin updated
[18:00:48] barryleiba joins the room
[18:00:53] <stpeter> are the slides being presented available at https://datatracker.ietf.org/meeting/88/materials.html#wg-httpbis ?
[18:00:58] <Larry Masinter> http://www.ietf.org/mail-archive/web/websec/current/msg01520.html
[18:01:14] naoki.kambe joins the room
[18:01:34] naoki.kambe leaves the room
[18:01:35] neilkatin joins the room
[18:02:02] <Bjoern> (The problems I pointed out with Origin do not really have anything to do with not referencing "WHATWG" documents.)
[18:02:38] naoki.kambe joins the room
[18:05:04] martini leaves the room: Replaced by new connection
[18:05:04] martini joins the room
[18:05:13] <ted.h> Could the security folks just stay looped in?  
[18:05:36] <richard.barnes> ted.h: as opposed to…?
[18:05:52] <ted.h> He said "We'll loop them back in when we get to…"
[18:06:01] <richard.barnes> oh, ok
[18:06:43] martini leaves the room
[18:07:13] naoki.kambe leaves the room
[18:09:29] fielding joins the room
[18:10:16] naoki.kambe joins the room
[18:10:30] <Larry Masinter> Is there any advantage  for not encrypting already encrypted HTTP message bodies?  If bulk of HTTP traffic is already-encrypted video
[18:10:57] <ted.h> Larry, that was a bit too many negatives—can you rephrase?
[18:11:32] <ted.h> Is the question, is the overhead of double encryption an issue?
[18:11:48] <Larry Masinter> ted, yes
[18:12:42] <Larry Masinter> I don't think HTTP2 has really been optimized for video delivery at all
[18:12:46] <Rohan Mahy> i can think of an advantage to encrpyt already encrypted data.  it obscures the names of the resources and hides that the original data was already encrypted.
[18:13:05] <Larry Masinter> you might still want to encrypt the headers
[18:13:11] <Larry Masinter> and certainly the whole request
[18:13:12] <ted.h> Personally, I think the answer is no, but that's partly because there is overhead in splitting out traffic that is already encrypted and partly because the metadata leaks if you leave off the outer encryptiong
[18:13:24] <hildjj> there's a value to chaff.
[18:13:37] <richard.barnes> larry: the encryption is symmetric so there's very little size overhead, and the computation overhead is pretty cheap.  so i'm not super worried
[18:13:45] <Larry Masinter> it's the overhead. i don't think anyone is measuring performance of HTTP2 for video delivery
[18:14:54] naoki.kambe leaves the room
[18:15:21] naoki.kambe joins the room
[18:15:25] <Larry Masinter> question for this discussion: what are the use cases where encryption overhead is significant and unnecessary. Delivery of already-encrypted video is a possible use case
[18:16:04] Tim Bray leaves the room
[18:16:22] naoki.kambe leaves the room
[18:16:35] naoki.kambe joins the room
[18:16:46] naoki.kambe leaves the room
[18:16:49] <richard.barnes> encryption overhead is only really significant when you're doing a bunch of public key operations, and that's addressed by the muxing / connection lifetime in HTTP/2
[18:17:14] ted.h leaves the room
[18:18:32] barryleiba leaves the room
[18:18:37] kohei.kasamatsu130 leaves the room
[18:18:38] richard.barnes leaves the room
[18:18:40] barryleiba joins the room
[18:19:02] <fielding> Nice long line at the mic. Do I really have to go to the mic to remind people that HTTP isn't only used by browsers? I guess not.
[18:19:11] kohei.kasamatsu130 joins the room
[18:19:13] m&m leaves the room: Disconnected: Replaced by new connection
[18:19:13] m&m joins the room
[18:19:47] geoffrey.cooper58@gmail.com joins the room
[18:20:50] richard.barnes joins the room
[18:22:52] kohei.kasamatsu130 leaves the room
[18:23:26] richard.barnes leaves the room
[18:23:38] <sftcd> is that one mic-line or two? the back mic seems to have people but maybe they need to get to the front?
[18:23:45] kohei.kasamatsu130 joins the room
[18:23:54] <fielding> Alternate-Services is just a Link header field (or at least it would be if people weren't so wonky about creating URI schemes that are aliases)
[18:24:38] <sftcd> I didn't know all video was encrypted - is there a MIME type for that?
[18:25:31] <fielding> SHTTP lives!
[18:25:39] EKR leaves the room
[18:26:02] <sean.turner@jabber.psg.com> http://www.iana.org/assignments/media-types/video
[18:27:23] <fielding> You cannot make a logic argument based on your own unwillingness to agree to valid criticism
[18:27:32] Tim Bray joins the room
[18:28:09] EKR joins the room
[18:28:11] <Eliot Lear> if we are going to do OE i would rather see it initiated – and state maintained within a connection.  That way any introduced risk is bounded.
[18:28:31] Tim Bray leaves the room
[18:29:27] josephyee joins the room
[18:30:48] Tim Bray joins the room
[18:32:11] synp joins the room
[18:32:15] ted.h joins the room
[18:37:23] josephyee leaves the room
[18:38:14] <cabo> Users aren't going to understand "unauthenticated encryption".  Better call it "insecure encryption" or "scrambling" or some such…
[18:38:22] <ted.h> But there's a new carrot:  if you get a TLS cert you get the other advantages of 2.0.  That may very well be insufficient, but that's the reason why you might not believe the current experiment is a valid predictor of what will happen with TLS in 2.0.
[18:38:27] richard.barnes joins the room
[18:38:33] <hildjj> "light obfuscation"
[18:39:25] <richard.barnes> hildjj: TLS_none_with_rot13_none
[18:39:40] mcmanus leaves the room
[18:40:16] <hildjj> it seems reasonable that a new UA would start connecting to 443 no matter what the URI scheme is, and only fall back to 80 if 443 fails and the scheme is http
[18:40:41] <richard.barnes> hildjj: has the nice side effect of increasing latency for unencrypted
[18:40:47] <hildjj> yes.
[18:40:57] <hildjj> If you want your site to be faster, turn on encryption.
[18:41:14] <ted.h> Happy eyeballs for encryption?
[18:41:33] <hildjj> ted: i was about to suggest that, but the "encryption is faster" is a better property. :)
[18:41:33] <fielding> TLS is bullshit security if it isn't backed by verification that the certificate belongs to the desired endpoint.  Existing https is weak enough already given the lax rules on certificate authorities. If we care about actual security and not just window dressing, then we should develop a protocol that includes message-based encryption by either sender.
[18:41:34] npdoty@gmail.com joins the room
[18:42:03] <richard.barnes> fielding: it's not bullshit, it prevents a passive attacker from listening int
[18:42:08] <richard.barnes> s/int/in
[18:42:30] <sftcd> and what's the probability of message-based encryption being done?
[18:42:34] <hildjj> fielding: i could imagine UAs at least doing SSH-style, and perhaps sharing site->cert mappings with friends, looking for discrepencies.
[18:42:36] <geoffrey.cooper58@gmail.com> Screwing up TLS security will come back to haunt us all later. This is a bad idea.
[18:42:59] <sftcd> TLS already supports this
[18:43:10] <sftcd> anon ciphersuites have been around almost for ever
[18:43:36] lef joins the room
[18:43:43] <hildjj> anon ciphersuite solves a separate problem.
[18:43:54] <sftcd> screwing up TLS would be bad but that's not what's under discussion here
[18:43:59] <richard.barnes> sftcd: what's puzzling to me is that the -relaxed thing does not actualy use anon ciphersuites
[18:44:14] <sftcd> it could though right?
[18:44:32] <richard.barnes> yeah, that's wg discussion i guess
[18:44:41] geoffrey.cooper58@gmail.com leaves the room
[18:44:46] <hildjj> anon ciphersuites are good to downgrade people into, if you're the sort that wants to save money on your later attacks.
[18:44:51] richard.barnes leaves the room
[18:45:04] <fielding> richard, it doesn't help anyone to force passive attackers to use active attacks instead - race conditions are not a difficult active attack
[18:45:27] synp leaves the room: Computer went to sleep
[18:45:49] sftcd joins the room
[18:46:00] <fielding> And nothing we discuss here prevents people from using https URIs when they do have a certificate
[18:47:16] sftcd joins the room
[18:47:34] richard.barnes joins the room
[18:48:06] <cabo> It helps a lot to force passive attackers to use active attacks.
[18:48:40] <npdoty@gmail.com> why can't you claim it as a security feature even if the security is not perfect?
[18:48:50] <sftcd> @cabo: yeah the browser might not see it as a security feature, but the spooks would
[18:49:15] <sftcd> some people talk about security in binary ways, which is is wrong 50% of the time
[18:49:26] <npdoty@gmail.com> most privacy and security features in browsers do not protect against every possible attack or attacker, but I still consider them features
[18:50:59] <richard.barnes> npdoty: important thing is being clear in what is being achieved
[18:51:22] <tlr> richard, right -- but not a reason to just do nothing.
[18:51:37] <tlr> And I fear "let’s all move to https" has a too high migration cost
[18:51:41] sftcd leaves the room
[18:51:47] <richard.barnes> tlr: that's exactly the point i was trying to make
[18:51:51] <tlr> (as opposed to "let’s turn on TLS by default")
[18:52:10] hildjj leaves the room
[18:52:23] <richard.barnes> to roberto's point on cost: if we get rid of cleartext, then that cost is just part of the ante for doing HTTP/2 — in order to get all the other benefits, you've gotta eat the crypto
[18:52:33] <richard.barnes> equalizes the playing field
[18:52:50] sftcd leaves the room
[18:53:42] hildjj joins the room
[18:54:54] <fielding> npdoty, because half-assed security has the effect of dissuading people from going to the effort of providing actual security through end-to-end TLS with certificate verification, or the development of other protocols that can secure the user's message against the various attacks.
[18:54:54] wseltzer@jabber.org leaves the room
[18:54:57] <Tim Bray> The arguments I hear from people about how setting up TLS is too expensive/complicated, or screws up their load balancing, are getting weaker and less convincing every year.  I’ve seen a significant move of the needle just in the last couple of years
[18:55:17] wseltzer@jabber.org joins the room
[18:57:40] lef leaves the room
[18:58:43] kohei.kasamatsu130 leaves the room
[18:59:03] kohei.kasamatsu130 joins the room
[18:59:59] Tom Yu leaves the room
[19:00:40] <hildjj> I am no longer convinced that certificates issued by big CAs protect us against active attacks.
[19:00:55] <hildjj> in which case we need out of band mechanisms anyway.
[19:01:06] <fielding> timbray, this is not a discussion about whether or not it is good to use https URIs. It is a discussion about whether it will be possible to not use TLS with http URIs. We are talking about what is the minimum footprint of an HTTP impementation, a protocol that is currently implemented on micro devices.
[19:01:21] <wseltzer@jabber.org> at least not certs+typical user interaction
[19:02:53] <Wolfgang Beck> So it does not replace https but plain text http
[19:03:03] <richard.barnes> wolfgang: exactly
[19:04:31] synp joins the room
[19:05:37] <npdoty@gmail.com> is the concern that servers who learn about the unauthenticated encryption feature will decide not to get a cert and do authenticated encryption when they otherwise would have?
[19:06:12] <EKR> npdoty: I think so, yes
[19:06:31] <ted.h> npdoty:  I think so
[19:07:02] <npdoty@gmail.com> ekr, and is there some evidence that would happen? I don't think browsers are promising lock icons or any additional sugar for the opportunistic unauthenticated case
[19:08:02] <npdoty@gmail.com> that is, servers have all the same incentives to get a cert that they had before (and I hear the performance is even easier) — why would encryption against a purely passive attacker dissuade them?
[19:08:31] <richard.barnes> afaik, there is no evidence, only hypothesis
[19:09:25] <ted.h> npdoty:  Regrettably, I have no data.
[19:10:14] <ted.h> But I think Martin's "simple statement" draft applies.
[19:10:26] <richard.barnes> ted.h: it totally sounds to me like you and roberto are arguing for the status quo, which smells like fail to me
[19:10:40] <ted.h> I am arguing for HTTP/2.0 to be TLS only.
[19:10:49] <npdoty@gmail.com> survey data would certainly be great, but do we even have a single plausible example of a server that would be dissuaded from getting a cert?
[19:10:52] <ted.h> That is not the status quo.
[19:11:00] <richard.barnes> ted.h: TLS only with authentication requried?
[19:11:03] <ted.h> Yes.
[19:11:16] <richard.barnes> then your authentication is going to be shit
[19:11:20] <hildjj> i'm arguing for authentication-required only for http: URIs
[19:11:27] <hildjj> sorry, https
[19:11:37] <hildjj> but always TLS
[19:11:43] <richard.barnes> hildjj: that's where i am
[19:11:52] <hildjj> that also means that we don't need a different upgrade mechanism
[19:12:00] <richard.barnes> TLS-only, but http=>unauthenticated, https=>authenticated
[19:12:14] <hildjj> richard: yes.  same as you.
[19:12:39] <hildjj> plus browser vendors doing interesting out-of-band things to try to detect MITM attacks.
[19:12:57] <ted.h> richard.barnes: I understand that point of view.  But I think having the http URI scheme stay with 1.1 and https moves to 2.0 is a better result.
[19:13:04] <hildjj> for both schemes.
[19:13:43] <ted.h> hildj: do you mean pinning, notary schemes, or something else?
[19:14:32] <richard.barnes> ted.h: i understand why that seems appealing.  i just don't think we have an authentication infrastructure that can provide service that is both high-quality and truly web-scale
[19:14:33] <ted.h> hildj: and will those alleviate richard's concern that my authentication in an authentication-only mode will be shit?
[19:14:34] <hildjj> at least those, probably others. take advantage of the fact that you've got sensors all over the Internet.
[19:14:36] Alex Mayrhofer joins the room
[19:14:44] <richard.barnes> so you'll end up with about the same amount of cleartext as today
[19:15:07] <Alex Mayrhofer> I think that Opportunistic can be useful - look at MTA to MTA traffic in SMTP, which is almost always opportunistic
[19:15:36] <Alex Mayrhofer> it's "better than nothing" - i really like "opportunistic encryption is the new clear text"
[19:15:41] <ted.h> He carries a CA in his pocket?
[19:15:47] <hildjj> ted: i don't know about your other question.  thinking about it.
[19:15:49] <sftcd> a raspberry pi
[19:15:58] <sftcd> and maybe a ca
[19:15:58] <richard.barnes> phb IS a CA
[19:16:12] <sftcd> but NOT the commodo root:-)
[19:16:18] <SM> :D
[19:16:21] <barryleiba> I want to carry a valet in my pocket.
[19:16:32] <barryleiba> Push a button and he comes out and brings me a beer.
[19:16:32] <ted.h> If we treat opportunistically encrypted traffic as cleartext, I think we have moved forward against passive attacks.
[19:16:51] <Alex Mayrhofer> +1, Ted.
[19:16:53] <richard.barnes> ted.h: that's what i want
[19:16:56] <hildjj> yes.
[19:17:19] Larry Masinter leaves the room
[19:17:23] <ted.h> But I think we are risking pervasive attack replacing pervasive passive attacks.
[19:17:37] ted.h points to Martin's draft.
[19:17:53] kohei.kasamatsu130 leaves the room
[19:18:09] <Alex Mayrhofer> were there pervasive active attacks against MTA-to-MTA traffic using opportunistic encryption?
[19:18:12] <neilkatin> Do we expect http2.0 to completely replace http1.1?  If not, then it seems to be fine to say http2.0 supports encryption only; the wg can relax it later if needed...
[19:18:22] <richard.barnes> ted.h: even in that case, you're increasing cost.  maybe marginally, but non-zero
[19:18:36] <ted.h> @neilkatin  thinking on that differs
[19:19:17] <ted.h> @richard.barnes:  you're increasing costs only for the http cleartext mitm; you're reducing it for current TLS-protected mitm.  So it's an estimation game again.
[19:19:18] <Alex Mayrhofer> i never understood why browsers treated encryption with self-signed certs *worse* than clear text, at least on the UI level.
[19:19:27] <richard.barnes> ted.h: yes
[19:19:37] <SM> neilkatin, that was tried for IPv6.
[19:19:45] <cabo> s/unauthenticated encryption/authenticated encryption not rooted in traditional root CAs/
[19:19:45] <neilkatin> could someone in the room transcribe the questions to jabber please?
[19:19:47] <hildjj> alex: because they use https: URIs to get to that state
[19:19:55] <ted.h> 0: Don't know
[19:19:56] <richard.barnes> my estimation is that size(cleartext => unauth) >> size(auth => unauth)
[19:19:59] <npdoty@gmail.com> is there evidence that pervasive active attacks are nearly as easy as pervasive passive attacks? I think research on GFW shows that scalability of MITM attacks is a real challenge even for a well-resourced organization
[19:20:00] <richard.barnes> fwiw
[19:20:04] <ted.h> 1: Do nothing-hope that HTTPS gets more adoption
[19:20:21] <ted.h> 2: Opportunistic encryption without server authentication for HTTP URIs - just for passive attacks
[19:20:49] <neilkatin> @npdoty: I was wondering the same thing; there is lots of discussion of state-level actors as opponents, but that seems a small part of the whole issue (to me at least)
[19:20:59] <ted.h> 3: Opportunistic encryption with server authentication AND downgrade protetection (somehow) for HTTP URLs; no requirement upon HTTP/2 when not available
[19:21:21] <ted.h> 4: Require secure underlying protocol for HTTP/2.0 ( at least in the web browsing case)
[19:21:35] <Alex Mayrhofer> 4. doesn't seem to be exclusive, compared to the others.
[19:21:55] <Alex Mayrhofer> eg. 2. + 4. would be an option.
[19:22:52] synp leaves the room: Computer went to sleep
[19:23:43] <sftcd> I think "2 or 3" is hummable I can't see how to distinguish them at this point
[19:23:50] <Wolfgang Beck> State level opponents control CAs, not much difference between opp. enc. http and https for them..
[19:24:25] kohei.kasamatsu130 joins the room
[19:24:38] <ted.h> richard.barnes:  (returning to your estimation).  In terms of sites (or even resources), I think that might be true, but the value of the ones going from ( auth > unauth) may be balanced or higher.
[19:24:51] <cabo> @Wolfgang: This is detectable.  There is a big difference between undetectable and detectable.
[19:27:02] <fielding> That would be "forbid the use of http URIs -- force https only"
[19:27:40] jimsch1 leaves the room
[19:27:40] satoru.kanno@jabber.org leaves the room
[19:28:05] Eliot Lear leaves the room
[19:28:10] hildjj leaves the room
[19:28:11] EKR leaves the room
[19:28:15] Tim Bray leaves the room
[19:28:26] sean.turner@jabber.psg.com leaves the room
[19:28:29] Wolfgang Beck leaves the room
[19:28:30] martin.thomson leaves the room
[19:28:31] richard.barnes leaves the room
[19:28:36] semery leaves the room
[19:28:36] ted.h leaves the room
[19:28:48] SM leaves the room
[19:28:51] kohei.kasamatsu130 leaves the room
[19:29:01] m&m leaves the room
[19:29:03] neilkatin leaves the room
[19:29:09] Dan Wing leaves the room
[19:29:10] frodek leaves the room
[19:29:10] lef leaves the room
[19:29:10] cabo leaves the room
[19:29:15] sftcd leaves the room
[19:29:19] Rohan Mahy leaves the room
[19:29:29] fielding leaves the room
[19:29:42] stpeter leaves the room
[19:29:44] Alex Mayrhofer leaves the room
[19:29:48] kmurchison leaves the room
[19:32:56] lef joins the room
[19:35:10] Mani leaves the room
[19:35:40] lef leaves the room
[19:36:03] npdoty@gmail.com leaves the room
[19:40:48] lef joins the room
[19:44:04] Eliot Lear joins the room
[19:45:32] Bjoern leaves the room: Replaced by new connection
[19:46:57] frodek joins the room
[19:49:53] Eliot Lear leaves the room
[19:53:04] sean.turner@jabber.psg.com joins the room
[19:53:31] martin.thomson joins the room
[19:54:07] barryleiba leaves the room
[19:55:40] lef leaves the room
[19:55:46] tlr leaves the room
[19:56:56] tlr joins the room
[19:58:35] lef joins the room
[20:00:06] EKR joins the room
[20:00:35] sean.turner@jabber.psg.com leaves the room
[20:04:01] Eliot Lear joins the room
[20:07:56] jimsch1 joins the room
[20:08:02] jimsch1 leaves the room
[20:10:03] tlr leaves the room
[20:11:10] wseltzer@jabber.org leaves the room
[20:13:09] stpeter joins the room
[20:14:00] richard.barnes joins the room
[20:19:42] Bjoern joins the room
[20:36:03] Tim Bray joins the room
[20:38:39] cabo joins the room
[20:44:59] m&m joins the room
[20:47:40] lef leaves the room
[20:49:38] m&m leaves the room
[20:49:44] stpeter leaves the room
[20:50:21] james.godfrey leaves the room
[20:50:54] hildjj joins the room
[20:52:21] Tim Bray leaves the room
[20:54:49] EKR leaves the room
[20:55:39] martin.thomson leaves the room
[20:55:58] Eliot Lear leaves the room
[20:58:08] stpeter joins the room
[20:58:50] richard.barnes leaves the room
[20:59:41] stpeter leaves the room
[20:59:42] stpeter joins the room
[20:59:44] hildjj leaves the room
[21:00:17] Tim Bray joins the room
[21:02:13] martin.thomson joins the room
[21:02:45] tlr joins the room
[21:02:45] Rohan Mahy joins the room
[21:02:45] wseltzer joins the room
[21:03:28] stpeter leaves the room
[21:03:47] lef joins the room
[21:04:49] EKR joins the room
[21:11:26] Dan Wing joins the room
[21:11:36] richard.barnes joins the room
[21:18:34] mcmanus joins the room
[21:20:56] cabo leaves the room
[21:22:03] mcmanus leaves the room
[21:24:26] tlr leaves the room
[21:26:06] Eliot Lear joins the room
[21:26:47] tlr joins the room
[21:27:30] Rohan Mahy leaves the room
[21:31:53] Mani joins the room
[21:32:05] Mani leaves the room
[21:34:17] ted.h joins the room
[21:38:04] ted.h leaves the room
[21:39:33] Dan Wing leaves the room
[21:39:49] richard.barnes leaves the room
[21:52:49] Eliot Lear leaves the room
[21:53:57] Eliot Lear joins the room
[21:54:09] Eliot Lear leaves the room
[21:59:00] frodek leaves the room
[21:59:23] Tim Bray leaves the room
[21:59:40] tlr leaves the room
[22:01:15] tlr joins the room
[22:04:06] EKR leaves the room
[22:04:41] lef leaves the room
[22:06:21] tlr leaves the room
[22:07:13] martin.thomson leaves the room
[22:08:53] wseltzer leaves the room
[22:14:04] Tim Bray joins the room
[22:16:02] Tim Bray leaves the room
[22:17:53] npdoty@gmail.com joins the room
[22:18:17] martin.thomson joins the room
[22:20:56] npdoty@gmail.com leaves the room
[22:22:48] mcmanus joins the room
[22:23:16] Eliot Lear joins the room
[22:27:03] EKR joins the room
[22:30:13] tlr joins the room
[22:31:51] wseltzer joins the room
[22:31:53] lef joins the room
[22:42:50] Larry Masinter joins the room
[22:44:15] tlr leaves the room
[22:44:18] EKR leaves the room
[22:46:42] tlr joins the room
[22:52:40] tlr leaves the room
[22:53:18] satoru.kanno@jabber.org joins the room
[22:53:19] tlr joins the room
[23:14:11] mcmanus leaves the room
[23:17:32] Tim Bray joins the room
[23:25:29] Tim Bray leaves the room
[23:27:21] Bjoern leaves the room
[23:29:12] EKR joins the room
[23:35:27] Eliot Lear leaves the room
[23:48:24] EKR leaves the room
[23:51:04] tlr leaves the room
[23:55:12] wseltzer leaves the room
[23:56:22] martin.thomson leaves the room
[23:58:15] Larry Masinter leaves the room
[23:59:12] satoru.kanno@jabber.org leaves the room
[23:59:42] lef leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!