IETF
appsawg@jabber.ietf.org
Monday, July 29, 2013< ^ >
stpeter has set the subject to: Applications Area Working Group | visit xmpp:apparea@jabber.ietf.org?join
Room Configuration
Room Occupants

GMT+0
[06:55:29] Scott Hollenbeck joins the room
[06:55:49] Scott Hollenbeck leaves the room
[06:56:41] Franck Martin joins the room
[06:57:32] JcK joins the room
[07:01:02] mnot joins the room
[07:01:21] yuioku.yj joins the room
[07:01:28] m&m joins the room
[07:01:32] yone joins the room
[07:01:49] Scott Hollenbeck joins the room
[07:01:57] <mnot> audio check?
[07:02:18] resnick joins the room
[07:02:25] mnot is your jabber scribe; if you have a question you'd like asked in the room, please say so explicitly.
[07:02:27] <JcK> Audio is working although there is a nasty hiss in the background
[07:02:42] tonyhansen joins the room
[07:02:43] <mnot> thx
[07:02:53] wseltzer joins the room
[07:02:56] m&m leaves the room: Disconnected: connection closed
[07:02:58] <mnot> Chairs: introductions
[07:03:05] m&m joins the room
[07:03:09] <mnot> Slide is "Work, Work, Work..."
[07:03:09] Joe Hildebrand joins the room
[07:03:13] <JcK> Audio just cut out entirely.
[07:03:21] <mnot> anyone else on audio?
[07:03:27] Hugo Kobayashi joins the room
[07:03:31] <mnot> pete at mike
[07:03:35] barryleiba joins the room
[07:03:36] Jim Galvin joins the room
[07:03:39] john.levine joins the room
[07:03:41] <mnot> talking about http-forwarded
[07:03:44] <mnot> coming to completion
[07:03:45] paf joins the room
[07:03:46] jtrentadams joins the room
[07:03:56] <mnot> john, still no audio?
[07:04:09] <JcK> still no audio here.
[07:04:16] <mnot> WGLC right after this meeting for malformed mail and xml mediatypes
[07:04:18] <resnick> Anybody else on the line?
[07:04:35] cabo joins the room
[07:04:43] <mnot> a ticket has been submitted for the audio
[07:04:52] <JcK> thx.  
[07:04:53] <mnot> comments feedback to the mailing list
[07:04:54] Joe Hildebrand leaves the room
[07:05:01] Joe Hildebrand joins the room
[07:05:09] <resnick> Audio seems like a regular issue first session on Monday.
[07:05:18] <mnot> merge-patch and sieve-duplicate drafts
[07:05:24] <mnot> comments, suggestions?
[07:05:51] fneves joins the room
[07:05:51] <mnot> proposed work: service-template
[07:06:19] <mnot> presentations; Service Configuration in Network Function Virtualization
[07:06:40] <mnot> draft-song-appsawg-service-template
[07:07:59] tlr joins the room
[07:08:03] <JcK> I closed at this end and restarted... have sound back, but the hiss (more or a steady static noise) is, subjectively, worse than before.
[07:08:20] <mnot> ok; tell us how you go
[07:09:06] <mnot> slide: "Current Dynamic Service Configuration"
[07:09:12] =JeffH joins the room
[07:09:16] tonyhansen leaves the room
[07:10:13] jtrentadams leaves the room
[07:10:13] <mnot> slide: "Use Case in NFV Context"
[07:11:25] <mnot> slide: "Problem Space - 1"
[07:11:30] tonyhansen joins the room
[07:11:31] jtrentadams joins the room
[07:11:41] <resnick> I will ask at the mic, but where is the apps protocol in this? This sounds like at best an ops or management protocol, but mostly is software running on the NFV controller.
[07:11:57] <mnot> this sounds cloudy
[07:12:15] <barryleiba> SDN-y
[07:12:36] <mnot> the cloud folks claim that as their own... all of your acronyms belong to us
[07:13:03] <resnick> Exactly. I have yet to see cloudy things that actually involve apps-layer protocol (other than HTTP as a substrate).
[07:13:13] <mnot> slide: "Problem Space - 4"
[07:13:37] sftcd joins the room
[07:13:40] <mnot> Oh, I think it could be done here, the question is whether it should; most of those folks aren't here.
[07:13:43] sa10kan3@gmail.com joins the room
[07:14:04] Olafur Gudmundsson joins the room
[07:14:13] <mnot> slide: "Discussions"
[07:14:27] yoav.nir joins the room
[07:15:04] <resnick> Yes, more and more this sounds like ops/mgmt work. Still not seeing apps protocol.
[07:15:20] <barryleiba> I agree
[07:15:28] <mnot> If the OpenStack, CloudStack, etc. folks came here and said "We want to standardise cloud mgmt" it'd be more likely to be real.
[07:15:45] paf leaves the room: Replaced by new connection
[07:15:46] paf joins the room
[07:15:46] <mnot> slide: "Potential IETF Work"
[07:16:14] paf leaves the room: Replaced by new connection
[07:16:14] paf joins the room
[07:16:35] <mnot> Dave Crocker at mike
[07:16:46] Karen O'Donoghue joins the room
[07:16:47] <mnot> Dave: sounds similar to SDN
[07:17:00] <mnot> Dave: possible that the paradigms could be siilar
[07:17:02] <mnot> similar
[07:17:21] <mnot> Pete Resnick: question - has this work been looked at by OPS? Seems like a better fit
[07:17:34] <mnot> Pete: Also, you described "three interfaces"; this doesn't sound like applications protocols
[07:17:42] <mnot> Pete: So, again, I'm wondering why this is APPS
[07:17:55] <mnot> answer: these are protocols between the user and controller
[07:18:05] <mnot> answer: this is the first time this has been brought to the IETF
[07:18:10] <mnot> Pete: I'd recommend OPS
[07:18:50] <mnot> stpeter: Service Configuration is a doc from Cyrus Daboo; it's now aggregated service discovery
[07:19:05] <mnot> stpeter: we had a BoF about this - how to configure clients dynamically
[07:19:17] <mnot> stpeter: that was directed towards e-mail, IM clients; may be some overlap
[07:19:25] synp joins the room
[07:19:34] yoav.nir leaves the room
[07:19:35] <mnot> other q?
[07:19:52] <mnot> BEGIN: Area meeting
[07:20:00] <mnot> slide: New Working Groups
[07:20:05] <mnot> 5 new groups since Orlando
[07:20:11] <mnot> quick summaries
[07:20:26] <mnot> STOX WG
[07:20:27] paf leaves the room
[07:20:29] paf joins the room
[07:20:30] <mnot> SIP-to-XMPP
[07:20:36] <mnot> started in June
[07:21:08] <mnot> drafts and gateways have been around for years; now formalising
[07:21:18] <mnot> e.g., address mapping
[07:21:48] <mnot> meeting: Thursday
[07:21:54] SM joins the room
[07:21:58] <mnot> Alexey: QRESYNC
[07:22:08] <mnot> Resyncronisation in IMAP
[07:22:23] <mnot> two docs; one extension, one extension to the extension
[07:22:25] <mnot> may combine documents
[07:22:36] <mnot> not planning to meet
[07:22:54] <mnot> SACM - Security Automation and Continuous Monitoring
[07:23:01] <mnot> assessing endpoint security posture
[07:23:21] <mnot> interoperating with an repository of content related to assessing endpoint security posture
[07:23:27] <mnot> meeting on Friday
[07:23:42] <mnot> Barry: How does SACM relate to NEA (?)
[07:24:01] Andrew Sullivan joins the room
[07:24:01] <mnot> answer: closely related; a subset of posture information related over ANEA (?) protocol
[07:24:13] <mnot> Matt Miller: JSON WG
[07:24:22] <mnot> standards-track version of RFC4627
[07:24:27] <mnot> minimal changes
[07:24:37] <mnot> lots of vigorous discussion on list
[07:24:39] synp leaves the room: Computer went to sleep
[07:24:50] Joe Hildebrand leaves the room
[07:24:51] <mnot> Slide: BoF Previews
[07:25:02] Joe Hildebrand joins the room
[07:25:08] <mnot> Barry: how many of you looked at the posters
[07:25:14] <mnot> roughly 1/4 hands up
[07:25:32] Julian joins the room
[07:25:39] <mnot> Trent Adams
[07:25:44] <mnot> DMARC BoF
[07:25:49] <mnot> WG-forming
[07:25:50] jtrentadams leaves the room
[07:25:52] <fneves> Nothing about LMAP?
[07:26:05] <mnot> extensions and BCP for DMARC spec (has been submitted) on individual stream
[07:26:09] <Joe Hildebrand> fneves: chairs must not have been in the room
[07:26:10] Julian leaves the room
[07:26:12] <mnot> no one for LMAP
[07:26:18] <mnot> DMARC on Tuesday
[07:26:43] Julian joins the room
[07:26:44] <mnot> Barry: Base spec is AD-sponsored; contingent on getting a critical mass of reviews
[07:26:44] Julian leaves the room
[07:26:53] Julian joins the room
[07:27:13] <mnot> need to determine whether it's appropriate for a proposed standard
[07:27:26] Julian leaves the room
[07:27:27] jtrentadams joins the room
[07:27:42] lef_jp joins the room
[07:27:43] <mnot> 6TSCH
[07:27:55] <JcK> Audio is barely ok.  The hiss/buzz in the background nearly overwhelms many of the speakers, including the current one.
[07:28:03] <resnick> Pronounced "six-toosh"? :)
[07:28:17] <barryleiba> Quote: "We are looking for the worst acronym."  Response: "You're so there."
[07:28:18] <mnot> (speaker too fast to scribe)
[07:29:03] <mnot> 6TSCH meeting tomorrow
[07:29:10] <mnot> slides: 6TSCH
[07:29:23] <mnot> slide: IEEE802.15.4e TSCH
[07:29:34] <mnot> nex tslide
[07:29:40] <mnot> 6TSCH Architecture
[07:29:56] Ning Kong joins the room
[07:30:39] Ning Kong leaves the room
[07:30:50] <mnot> BoF: DICE
[07:30:56] <mnot> DTLS in Constrained Environments
[07:31:01] <mnot> meeting on Wednesday
[07:31:16] <mnot> DTLS is must-implement for COAP
[07:31:32] <mnot> if you're a new implementer, DTLS makes brain explode
[07:31:50] <mnot> This BoF is looking to define a minimal profile of DTLS for Internet of Things
[07:32:11] redaka joins the room
[07:32:13] <mnot> Also trying to solve group communications through reuse of the record layer
[07:32:23] <mnot> Wes Eddy: AQM BoF
[07:32:32] <mnot> Active Queue Management and Packet Scheduling
[07:32:55] <mnot> Would like to develop specs for AQM algorithms
[07:33:13] resnick leaves the room
[07:33:18] <JcK> Can't hear this speaker at all - very faint, buzz is louder.
[07:33:29] resnick joins the room
[07:33:33] <mnot> Hopefully, helping to get AQM deployed will require less parameter tuning, help with buffer bloat
[07:34:06] <mnot> There are multiple tickets on the audio; apparently it's the hotel
[07:34:08] <mnot> Alexey at mike
[07:34:30] <mnot> POSH BoF
[07:34:41] <mnot> Multiple domain hosting
[07:34:56] <mnot> assigning trust for certificates
[07:35:03] <mnot> Barry at Mike
[07:35:07] stpeter joins the room
[07:35:15] <mnot> You should all be aware of what's going on in IPR BIS
[07:35:18] <mnot> BCP79 update
[07:35:45] <mnot> There are many clarifications, a few significant changes
[07:35:47] <mnot> this afternoon
[07:36:26] <mnot> Apparently no one can speak to DNSSDEXT and STIR
[07:36:29] <Andrew Sullivan> everyone from DNSSDEXT is probably in HOMENET
[07:36:35] <mnot> Barry: STIR is to secure telephone URIs
[07:36:45] <mnot> (argument in the room about that)
[07:36:57] <mnot> Joe H: It's secure caller ID in a world that includes ?
[07:37:21] <mnot> (frivolity ensues)
[07:37:35] Ning Kong joins the room
[07:37:47] <mnot> DNSSDEXT: It's about service discovery for HOMENET
[07:38:33] tlr leaves the room
[07:38:39] <mnot> Andrew Sullivan: There was some pressure to do multicast DNS outside the LAN context
[07:38:52] <mnot> Andrew: e.g. power management wants to bridge across lan boundaries
[07:39:06] <mnot> Andrew: also, academic users of Bonjour can't use apple TV down the hall
[07:39:14] <mnot> Andrew: this is an attempt to fix that problem
[07:39:30] <mnot> Slide: Miscellaneous
[07:39:32] <mnot> Barry at mike
[07:39:32] paf leaves the room: Replaced by new connection
[07:39:35] <Andrew Sullivan> Apparently, I can talk faster than I can type.
[07:39:36] paf joins the room
[07:39:39] <mnot> discussing Applications Area Directorate
[07:39:54] <mnot> New leader of the directorate is Claudio
[07:40:00] <mnot> Thanks to SM for his long term
[07:40:20] <mnot> Working on getting timing / deadlines right
[07:40:21] tonyhansen leaves the room
[07:40:28] <mnot> giving reviewers time, while still making review useful
[07:40:32] <mnot> Pete at mike
[07:40:37] <mnot> Was talking to Dave last night
[07:40:54] <mnot> Pete: reviews were being assigned late in the process
[07:41:02] <mnot> Pete: IESG could be right after LC ending
[07:41:08] <mnot> Pete: So we got people to start reviews earlier
[07:41:23] <mnot> Pete: earlier reviews / tail heavy process discussion on IETF list
[07:41:38] healthyao2000 joins the room
[07:41:40] <mnot> Pete: separately, we want earlier reviews way before lc
[07:41:42] paf leaves the room: Replaced by new connection
[07:41:45] paf joins the room
[07:41:52] <mnot> Pete: these are two separate things
[07:42:15] <mnot> Barry: the second one isn't an action for the directorate; it's more for WG chairs and doc authors; need to request reviews before LC is over.
[07:42:52] <mnot> Murray: There's an RFC about Security Considerations
[07:43:07] <JcK> Hiss/buzz finally disappeared.   Assuming someone did something, thanks.
[07:43:08] <mnot> Murray: Would it be useful for us to produce a doc about how to make the apps directorate happy
[07:43:09] <mnot> ?
[07:43:15] <Andrew Sullivan> Yes, they just replaced a cable
[07:43:18] <mnot> Barry: We need to publicise the wiki
[07:43:20] <mnot> yay!
[07:43:51] <mnot> Barry: OPenness of IESG Process
[07:44:04] <mnot> Barry: We've been piloting this with specific WGs/Docs
[07:44:07] <JcK> @Andrew: that would explain the brief silence.  
[07:44:07] <stpeter> here is the wiki page: http://trac.tools.ietf.org/area/app/trac/wiki/TypicalAppsAreaIssues
[07:44:18] <mnot> Barry: usually, once a doc enters IESG process, it's opaque
[07:44:24] <stpeter> it's a wiki, feel free to edit :-)
[07:44:45] synp joins the room
[07:44:50] <mnot> Barry: Piloting getting the WG CC:ed on traffic about the doc
[07:44:53] <mnot> e.g., COAP
[07:44:53] synp leaves the room: Computer went to sleep
[07:44:58] <mnot> want to expand it
[07:45:03] <mnot> but we need feedback
[07:45:19] <mnot> Gives the WG more ownership of what happens to the document
[07:45:27] <mnot> makes ADs more responsible for their comments
[07:45:37] <mnot> stpeter at mike
[07:45:55] <mnot> stpeter: I like it; I didn't like that things were opaque when I was on the IESG
[07:45:55] synp joins the room
[07:46:30] <mnot> Pete resnick: We're piloting this so that folks can get used to it both in the IESG and in WGs
[07:46:35] tlr joins the room
[07:46:36] <mnot> Pete: It generates a LOT of e-mail
[07:46:51] <mnot> both because the WG sees IESG comments, and IESG sees responses to e-mail
[07:47:06] <mnot> when we turn on the feature, if every WG member jumps in and comments, it can get unmanageable
[07:47:23] <mnot> one of the things we've been trying (e.g., with JCARD) is to pre-flight and tell people to throttle responses
[07:47:31] <mnot> e.g., see if chair or author can answer question first
[07:47:56] <mnot> Carsten at mike
[07:48:05] <mnot> Carsten: It worked well for COAP
[07:48:14] <mnot> Carsten: Overall, it's the right thing to do; has to be managed.
[07:48:27] <mnot> Carsten: It's a tedious process; it might wear on the WG if it drags out
[07:48:32] Ning Kong leaves the room
[07:48:35] <mnot> Carsten: WG chair has to manage this
[07:49:00] <mnot> Dave Crocker: I've been in favour of more transparency; this has come up over the years and has been rejected.
[07:49:01] bhoeneis joins the room
[07:49:17] <mnot> Dave: Pete said that the way for this to work in scale is for everyone to change (laughter)
[07:49:31] <mnot> Dave: I don't have a suggestion as to how to fix that, but there needs to be some other approach.
[07:49:47] <mnot> Barry: pete didn't say that; he said that as we get started, we need to be careful.
[07:49:55] <mnot> (Dave shakes head in both directions at same time)
[07:50:04] <mnot> Pete: Dave's right, to an extent.
[07:50:22] <mnot> Pete: But, we're piloting it to get a sense of how big a change it is
[07:50:30] <mnot> Pete: It may turn out that it's only a slight change in behaviour.
[07:50:31] <JcK> Mic:  Pete, given some IETF participants, difficult WGs, and passion about some issues,, I can't imagine "please be nice" as sufficient.    I wonder if that process could be put through a moderated channel in the WG -> IESG direction.   Moderating and, if needed, summarzizing, would become part of the WG Chair of Secty job.
[07:50:38] <SM> mic: Why not have the discuss of the DISCUSS be WG-only and then have a summary at the end?
[07:53:49] <Joe Hildebrand> mnot: from above, what I said about STIR was that it was to secure Caller ID in a world that includes VoIP
[07:54:55] synp leaves the room: Computer went to sleep
[07:55:42] <mnot> stPeter at mike
[07:56:00] <JcK> Mic: Pete and Barry, I wouldn't suggest any filtering in the IESG-> WG direction.  In the other one, it would take only a topic that is controversial within the WG in combination with a rant-er to create a massive DoS attack on everyone.    At least I recommend having mechanisms in place that permit switching some sort of moderation process on very quickly.  Like you, I'd prefer to avoid long discussions followed by summarizing if possible.  Been there, done that, and it turns out to be a lot less transparent.
[07:56:08] <mnot> Want folks who are in charge of the document to take the lead in these discussions, rather than having random folks jump in
[07:58:36] Ning Kong joins the room
[07:58:41] <mnot> Murray: I'm curious about this for APPSAWG
[07:58:57] <mnot> Pete: Appsawg can figure that out for each doc
[07:59:12] <mnot> Barry: We did it with webfinger - because it was on a separate list
[07:59:26] <mnot> Pete: The IESG is going to do this "send back to the WG thing" more often
[07:59:39] <mnot> pete: When a doc goes into evaluation, and an ad makes a discuss
[07:59:52] <mnot> pete: it becomes the ADs responsibility to work it to conclusion
[08:00:05] <mnot> pete: the responsible ad and shepherd need to make that happen
[08:00:29] <mnot> pete: if there are a bunch of discusses / open issues, it's unwieldy to have all of those discussions
[08:00:41] <mnot> pete: so we want to get the work back into the WG, rather than making it an IESG problem
[08:01:00] <mnot> pete: if there are a number of topics, we'll prefer to send the comments back to the WG, just as if they were LC issues
[08:01:08] g.e.montenegro joins the room
[08:01:08] Floris joins the room
[08:01:26] <mnot> Barry: That means that after the discussion has started, and the authors understand the issues, then it goes back to WG
[08:01:32] paf leaves the room
[08:01:43] Floris leaves the room
[08:02:06] <mnot> Dave Crocker: There's a simple model here that already exists -- if an AD has a discuss, they need to bring it to the WG, just as anyone else would
[08:02:19] <mnot> Dave: That's a much more trivial change, and puts the burden on the complainer
[08:02:23] paf joins the room
[08:02:26] <mnot> Dave: Then it's just WG process
[08:02:32] <mnot> Barry: and that's the preferred mechanism
[08:02:36] paf leaves the room
[08:02:39] kohei.kasamatsu130 joins the room
[08:02:44] <mnot> Dave: that also puts it out of the IESG for the duration of that work
[08:02:52] <mnot> Dave: it burdens the AD, as is appropriate
[08:02:56] paf joins the room
[08:02:58] <mnot> Pete: that's very close to what we have in mind
[08:03:08] kohei.kasamatsu130 leaves the room
[08:03:20] paf leaves the room
[08:04:41] <mnot> Alexey; We should add a state to the data tracker
[08:04:59] <SM> Wait, that requires a RFC update:)
[08:05:17] <mnot> Barry: COAP tried this; was a long document, difficult review. However, it got eight "YES" ballots.
[08:06:26] <mnot> Barry: What should ADs be doing, doing differently, etc.?
[08:06:41] paf joins the room
[08:07:26] <mnot> Larry: multipart/form-data
[08:07:28] paf leaves the room: Replaced by new connection
[08:07:29] paf joins the room
[08:07:34] <mnot> Larry: speced by HTML, but not implemented that way
[08:07:55] <mnot> Larry: willful violations by HTML community
[08:08:13] <mnot> Larry: there's a gap here
[08:08:24] <mnot> Larry: how do we update specs that are old?
[08:08:38] Julian joins the room
[08:08:38] <mnot> Joe H: I support larry's statement
[08:08:59] <mnot> Joe H: We chase people off like this regularly
[08:09:02] <mnot> Barry: How
[08:09:26] <mnot> Joe: E.g., hybi
[08:10:33] <mnot> Eliot at mike
[08:10:43] <SM> That's like saying that there is a high barrier for PS
[08:10:50] <barryleiba> yes
[08:11:33] <mnot> stpeter at mike
[08:13:06] <resnick> Dave Crocker at the mic
[08:14:50] <=JeffH> DaveC: this is a very multi-faceted topic area, and we need to put attention on each of those facets in order to make progress on this overall
[08:16:10] tonyhansen joins the room
[08:16:10] <=JeffH> larryM:  the Web-oriented stuff that's in the IETF is in build/repair cycle -- let's focus on these specs, and see what we can do to address them -- rather than take on the entire PR problems overall
[08:16:12] <SM> What Larry is talking is about maintenance of standards
[08:16:46] <=JeffH> StPete had confessed to OAuth being a particularly bad experience all around and one that is still a fount of ill-will going forward
[08:17:14] <=JeffH> mnot: we failed to revitalize IRI/URI and perhaps we need to "cede the ground" and let others take them over
[08:17:25] <=JeffH> if we can't address them, have others take 'em over
[08:17:31] <=JeffH> don't let 'em rot
[08:17:35] mnot thanks Jeff
[08:17:45] =JeffH welks
[08:17:54] <mnot> We don't have a good view of how operations use these things
[08:18:04] <=JeffH> Doug Otis @mic
[08:18:06] <mnot> When we see a problem, how do we handle it?
[08:18:09] mnot again
[08:18:30] <mnot> Keith Moore: publish a two page RFC that says consensus is that change control is ceded
[08:18:38] <mnot> Pete Resnick (no hat)
[08:18:59] <mnot> Ceding control is important; Dave made a reference to what's appropriate for here and what's not
[08:19:13] <mnot> we do a bad job in making a conscious decision about what's appropriate to work on
[08:19:25] <mnot> e.g., APIs, internal library specs
[08:19:40] <mnot> I think we never really made a conscious decision
[08:19:53] <mnot> Personally, there are lots of good reasons to be doing APIs, software specs
[08:20:11] <mnot> Some groups do this well, e.g., W3C
[08:20:26] <mnot> So I think it's a good idea to cede things to people who are more suited to doing it
[08:20:54] <mnot> Larry masinter: I don't think that the reality is nearly as neat as "If the mime specs in their applicability to the Web" then you probably don't want to give up control of MIME
[08:20:55] tlr leaves the room
[08:21:11] <mnot> Larry: There are lots of applications that use URIs that don't want to give up control, because they're not part of the Web
[08:21:24] <mnot> Larry: I don't think the problem can be solved by just saying that "we give this up."
[08:21:45] <mnot> Larry: More coordination is necessary
[08:22:29] <mnot> Andrew Sullivan: One of the problems in these maintenance cases is that it's hard, and nobody wants to do it, but nobody wants to give it up, because others will do it wrong.
[08:22:46] <mnot> Andrew: when people complain about it, they really mean that somebody else should do it, but boss them around
[08:23:03] <mnot> Andrew: so we should find people to do the work here, and shield them.
[08:23:08] resnick leaves the room
[08:23:13] <mnot> Andrew: Or we need to step up and take the heat.
[08:23:18] resnick joins the room
[08:23:52] <mnot> Paul Hoffman: I'm concerned that people are talking about ceding control. We don't share well, but that's not the only people who can do work. Things can be done through the independent submission editor.
[08:24:08] <mnot> Paul Hoffman: E.g., having someone work on form data is often enough.
[08:24:57] <mnot> Paul: no org needed
[08:25:09] <mnot> Thomas Roessler at mice
[08:25:22] <mnot> TLR: we have a shared interest in these things happening (W3C)
[08:25:28] <mnot> TLR: resources are liited
[08:25:37] <mnot> limited
[08:25:44] <mnot> TLR: It's fine if it happens at the IETF
[08:25:54] <mnot> TLR: right answer is to take it case by case
[08:26:03] <resnick> The non-maintenance cases are also interesting. OAuth and Hybi are a different sort of problem.
[08:26:08] <mnot> TLR: sometimes, W3C might need soething slightly different
[08:26:38] <mnot> TLR: need open, nuanced, detailed conversations
[08:27:11] <mnot> Eliot Lear; Agree with Thomas Roessler
[08:27:26] <mnot> Eliot: We're often introspective
[08:27:59] tlr joins the room
[08:27:59] <mnot> Eliot: Doing maintenance work - "I don't need anybody; I need SOMEBODY"
[08:28:40] <mnot> Keith Moore - I should point out that ceding control is a corner case, not a general solution
[08:29:05] <mnot> Keith - This organisation isn't structured to encourage maintenance
[08:29:24] <mnot> Keith - we try to solve these problems with the same approaches for 20 years
[08:29:31] <mnot> Keith - we need to explore here
[08:29:32] <JcK> Mic: The other issue here is that a notion keeps rearing his head (in different forms) that the IETF is in control, or should be in control, of the Internet and that we can prevent something from happening by (actively or passively) refusing to deal with it.  We need to beat that back every time it shows up -- it really gets in the way of both maintenance and new work.
[08:30:56] healthyao2000 leaves the room
[08:31:32] <=JeffH> mnot @ mic:  there's issues with docs being maintained by folks other than the originals.  also there's a NIH problem here, they run into roadblocks,  feel bad about oauth, need to figure out way to help folks come here and give them the aircover they need to be successful
[08:31:55] <JcK> thx
[08:32:24] paf leaves the room
[08:32:25] paf joins the room
[08:32:42] <mnot> stpeter: It's not fun to do maintenance. If people don't get feedback in, the process doesn't work.
[08:32:44] <=JeffH> stpeter: maintenance is hard and stuff, but it has to happen, dunno how to figure out whether folks who come do shiny new things will hang around and maintain it -- dunno how to figure this out
[08:32:54] =JeffH back to mnot
[08:33:08] <mnot> Dave Crocker - Maintenance is complementary to the point Pete raised about getting something out quickly
[08:33:20] <mnot> Dave - we have a tone that we review things by starting fresh every time
[08:33:41] <mnot> Dave - improvement in an incremental way is not in the review model
[08:33:51] <mnot> Dave - especially in maintenance activities, this is the wrong thing to do
[08:34:13] <mnot> Dave - people doing products don't want a new spec each six months
[08:34:49] <mnot> Dave - it could help enormously if we understand the core and get it out quickly, rather than trying to ship a complete spec
[08:35:07] <mnot> Dave - in six months, you add to the core, not revise the core
[08:35:27] <mnot> Larry Masinter - I want to get back to my use case of multipart/form-data
[08:35:34] <mnot> Larry - a WG seems like overkill
[08:35:41] <mnot> Larry - maybe a "Web-related specs" WG?
[08:36:35] <mnot> Tony H at mic
[08:36:40] <mnot> Tony - we had YAM
[08:37:03] <mnot> Tony - it was oriented around what Larry was talking about, in the mail space
[08:37:10] <JcK> Mic (if possible, otherwise ignore) I mostly agree with Dave, but am also reminded of an observation that, in info systems projects, there are people who are good "starters", including initial and conceptual design, and ones who are good "finishers", including spec-maintainers.  The IETF has tended to focus on and reward the former, which correlates with an industry tradition of maintainers not getting any respect.
[08:37:44] <Andrew Sullivan> mic line was closed
[08:37:47] <mnot> Tony - One of those docs was one I was responsible for
[08:37:51] <JcK> Ack
[08:38:09] m&m leaves the room: Disconnected: connection closed
[08:38:33] <barryleiba> YAM failed for specific reasons, beyond just that it was a maintenance WG.
[08:38:53] <SM> Gerontology
[08:39:10] <Andrew Sullivan> There is the other fact that many people who are interested in maintenance are also not always good spec people.
[08:39:45] <Andrew Sullivan> The thing about the "starters" is that they often want to write down exactly what you absolutely must know to make the protocol work
[08:39:52] <Andrew Sullivan> because that's all they know.
[08:40:05] <=JeffH> mnot: thinks need really good coord to make Larry's  suggestion to work, but is plausible, really need aircover, interesting in working with Ads and all to try to make this work
[08:40:15] <mnot> Update from JOSE WG
[08:40:26] <Andrew Sullivan> But maintainers are often interested in completeness and closure
[08:40:34] <mnot> Karen: JOSE is developing secure update format for JSON
[08:40:39] <Andrew Sullivan> so you end up with one of those maintenance tasks that never ships
[08:40:43] <mnot> Karen: we want early review
[08:40:53] <mnot> Karen: four key specs approaching WGLC
[08:40:58] <mnot> Karen: mature enough for review
[08:41:12] <JcK> YAM and the maintenance problems are also both symptoms of one of the reasons we don't advance more docs past Proposed Standard.  Not the whole story in either case -- another issue is the knowledge that any attempt to maintain and clean up a document almost always results in reopening old issues and being nit-picked to death about irrelevancies.  
@Andrew: exactly... and why we need to have more of them around and reinforce their work.
[08:41:25] <mnot> Karen: Will send a note to list
[08:41:49] <mnot> Barry: getting feedback that apps folks don't see applicability of JOSE
[08:42:03] <Andrew Sullivan> @JcK yes, but we also need to guide them so that they can learn how much an enemy of the good "perfect" can be
[08:42:09] <mnot> stpeter: Matt Miller and I use it in POSH
[08:42:12] <Andrew Sullivan> I had one example I'm thinking of in particular
[08:42:22] <mnot> stpeter: Matt also discovered that BrowserID folks are using an early version
[08:42:47] <mnot> barry: it would be good if that were in the doc in an implementation section, or shepherd write-up
[08:43:00] <mnot> Murray: WEIRDS as well
[08:43:02] <Andrew Sullivan> where a document that was originally 7 pages _with all the boilerplate_
[08:43:05] <mnot> CBOR
[08:43:09] <mnot> Paul's slides
[08:43:16] <Andrew Sullivan> turned into more than 25 pages of dense, sometimes inconsistent text
[08:43:35] <Andrew Sullivan> in an effort to "clarify", which actually meant "completely describe under all cases".
[08:44:07] <Andrew Sullivan> that was in no sense an improvement, and the document ended up dying.  It's too bad: we had a frustrated maintainer who, I suspect, is still angry, and a lot of useful work lost
[08:44:26] <Andrew Sullivan> but that was not good maintenance, and I still think it was the right thing to do.
[08:44:36] paf leaves the room: Replaced by new connection
[08:44:36] paf joins the room
[08:44:52] <mnot> CBOR Design and Overview
[08:45:03] <mnot> Slide: Why
[08:45:34] <JcK> @Andrew: yes.  And that guidance issue is the more positive side of that YAM-ish problem.  And, sometimes, that 7 -> 25 page problem is a symptom of improper scoping or scope explosion between earlier and laterr versions.  It is important to keep those distinctions in mind.
[08:46:30] <Andrew Sullivan> @JcK: agree.  Perhaps we could get together a list of successful maintenance examples, and we could point people at that as an illustration?
[08:46:47] m&m joins the room
[08:46:57] <Andrew Sullivan> things like 821/822 and descendants are probably _not_ the examples we want :)
[08:47:07] <mnot> stpeter at mic: There are other similar things - why another?
[08:47:25] <tlr> stpeter: http://xkcd.com/927/
[08:47:29] <mnot> paul: that's not for this description; we aren't saying people should do this. We compare in the document
[08:47:39] paf leaves the room: Replaced by new connection
[08:47:39] paf joins the room
[08:48:12] <mnot> Slide: What
[08:48:40] <mnot> Now in IETF LC
[08:48:58] <stpeter> https://datatracker.ietf.org/doc/draft-bormann-cbor/
[08:49:31] <mnot> Slide: Design Goals (1 of 3)
[08:51:54] <mnot> Slide: Design Goals (3 of 3)
[08:52:44] paf leaves the room: Replaced by new connection
[08:52:44] paf joins the room
[08:53:49] <mnot> Slide: Very quick overview of the format
[08:55:09] <mnot> Slide: Semantics of Additional Information
[08:55:58] <mnot> Slide: Some Features
[08:57:48] paf leaves the room: Replaced by new connection
[08:57:49] paf joins the room
[08:58:44] <mnot> Slide: What's in the Document
[08:58:56] <JcK> Please push Paul closer to the mic or vice versa -- he is getting more and more faint as he goes along.  thx
[08:59:34] <JcK> that is better
[09:00:36] <mnot> Slide: History
[09:01:48] paf leaves the room: Replaced by new connection
[09:01:48] paf joins the room
[09:01:56] <mnot> Marc Blanchett at mic
[09:02:07] <mnot> Have you considered adding a tag for IP addresses?
[09:02:21] <mnot> Paul: we haven't, but we could. If we don't, it could be added to the registry later.
[09:03:05] andrey.uzunov joins the room
[09:03:43] <mnot> Larry Masinter: updating BCP70 - would it be appropriate to include this?
[09:03:51] <mnot> Paul: Maybe; too early to tell
[09:04:08] <mnot> Dave Crocker - Jack H proposed something similar for what became RFC????
[09:04:17] <mnot> Dave - it's nice to see it back, haunting the IETF
[09:04:22] <resnick> 733
[09:04:23] paf leaves the room: Replaced by new connection
[09:04:23] paf joins the room
[09:04:38] <mnot> Barry - How many peole have commented? Read? (not many, a few)
[09:05:00] <mnot> Barry: docs that are AD-sponsored need a good set of eyes on them
[09:05:13] <Andrew Sullivan> Why "good"?  Why not just "not bad"?
[09:06:13] <resnick> Determine consensus from review alone. Needs to be serious consideration, IMO.
[09:06:19] <mnot> Andrew Biggs - Aggregated Service Discovery
[09:06:22] paf leaves the room: Replaced by new connection
[09:06:22] paf joins the room
[09:07:06] <mnot> Slide: Quick Recap of IETF86 BoF
[09:07:14] andrey.uzunov leaves the room
[09:10:26] <mnot> Slide: Progress
[09:10:39] nemo joins the room
[09:13:09] nemo leaves the room
[09:16:36] <mnot> John Levine: Extending the DNS by adding new RR Types
[09:17:19] <Andrew Sullivan> I'm going to avoid going to the mic, but I think this draft is a good idea
[09:17:33] <Andrew Sullivan> and I think it's an application
[09:17:39] <Andrew Sullivan> so I think it belongs here :)
[09:18:19] <mnot> Slide: Observations
[09:18:55] <mnot> Slide: DNS extension language
[09:19:25] <mnot> Slide: Stored in the DNS
[09:19:35] <JcK> @Andrew: perhaps DNSEXT should take it up.   Whoops! :-).  But I agree... probably a good idea
[09:19:43] redaka leaves the room
[09:19:55] <Andrew Sullivan> I specifically think that DNSEXT should _not_ have taken it up
[09:20:04] <Andrew Sullivan> that was a sure way to ensure it would die
[09:20:29] <JcK> @Andrew: yep
[09:21:02] <mnot> Dave Crocker at mic
[09:21:05] <JcK> For the record, these slides are not accessible from the meeting materials page and the presentation was a little hard to follow from the audio alone.
[09:22:47] <mnot> Patrick at mic
[09:22:56] <JcK> Mic (if possible): While I'm generally opposed to putting things in the DNS "because it is there" and "because we can't think of a better place", putting this sort of info there seems no less appropriate than using a server bootstrap mechanism.
[09:23:03] <Andrew Sullivan> mic line closed
[09:23:07] <Andrew Sullivan> (2 mins ago)
[09:23:42] <Andrew Sullivan> Patrik was impersonating Bad Idea Fairy at mic?
[09:23:45] <JcK> @andrew: Again, understood,  Hence the "if possible"
[09:24:53] <mnot> Andrew and John Levine
[09:25:06] <mnot> Putting Administrative Boundaries in the DNS
[09:25:20] <mnot> E.g., issuing certs, setting cookies
[09:27:29] Olafur Gudmundsson leaves the room
[09:28:17] <stpeter> https://datatracker.ietf.org/doc/draft-levine-orgboundary/ is John's draft
[09:28:28] <stpeter> https://datatracker.ietf.org/doc/draft-sullivan-domain-policy-authority/ is Andrew's (and Jeff's)
[09:30:01] <mnot> read drafts; about a quarter of the room
[09:30:08] <mnot> solve a problem: about the same
[09:30:29] barryleiba leaves the room
[09:30:30] <mnot> minimising dns lookups: none
[09:31:16] <mnot> mic lines filling up
[09:31:38] <mnot> paul hoffman at mic
[09:31:58] <mnot> paf at mic
[09:32:40] <mnot> crocker
[09:32:46] <JcK> If we really thought minimizing lookups was a hugely high priority, we would start deprecating NAPTR, URI, and probably some other things (like DNSSEC?).  Seems like we are too far down that slope to turn back.
[09:33:54] <mnot> peter
[09:33:58] cabo leaves the room
[09:35:32] <mnot> only one mechanism: one or two hands
[09:35:34] m&m leaves the room: Disconnected: connection closed
[09:35:48] mnot leaves the room
[09:35:52] Scott Hollenbeck leaves the room
[09:35:52] tlr leaves the room
[09:35:59] Ning Kong leaves the room
[09:36:03] <Franck Martin> the important is not to have one mecahnism but what happens when a registry does not care...
[09:36:17] yone leaves the room
[09:36:18] resnick leaves the room
[09:36:19] stpeter leaves the room: Offline
[09:36:26] Hugo Kobayashi leaves the room
[09:36:28] Joe Hildebrand leaves the room
[09:36:28] lef_jp leaves the room
[09:36:28] jtrentadams leaves the room
[09:36:28] sftcd leaves the room
[09:36:28] wseltzer leaves the room
[09:36:32] fneves leaves the room
[09:36:56] Karen O'Donoghue leaves the room
[09:36:58] tonyhansen leaves the room
[09:37:19] Julian lost the audio feed
[09:37:21] paf leaves the room
[09:37:28] bhoeneis leaves the room
[09:37:42] =JeffH leaves the room: Logged out
[09:37:43] <JcK> Audio died completely about 3 minutes ago.  Looks like a network failure because I'm also seeing a lot of Jabber drops from people in the room.
[09:37:58] Andrew Sullivan leaves the room
[09:38:02] bhoeneis joins the room
[09:38:28] bhoeneis leaves the room
[09:40:02] <SM> Yes
[09:40:20] yuioku.yj leaves the room
[09:40:28] Franck Martin leaves the room
[09:41:53] john.levine leaves the room
[09:42:20] <JcK> I assume the meeting has ended by now, but...  Hope the last minutes were at least recorded.
[09:42:39] sa10kan3@gmail.com leaves the room
[09:42:45] JcK leaves the room
[09:42:46] <SM> The people must have gone out for lunch:)
[09:43:41] SM leaves the room
[09:46:28] Jim Galvin leaves the room
[09:52:28] g.e.montenegro leaves the room
[09:53:58] m&m joins the room
[09:54:22] Julian leaves the room
[09:56:10] m&m leaves the room
[09:56:17] Jim Galvin joins the room
[10:00:23] Andrew Sullivan joins the room
[10:00:26] Andrew Sullivan leaves the room
[10:05:08] bhoeneis joins the room
[10:15:29] bhoeneis leaves the room
[10:23:36] sftcd joins the room
[10:34:51] sftcd joins the room
[10:34:59] sftcd leaves the room
[10:42:00] sftcd leaves the room
[10:48:31] bhoeneis joins the room
[11:01:56] cabo joins the room
[11:03:32] bhoeneis leaves the room
[11:07:17] bhoeneis joins the room
[11:15:30] Ning Kong joins the room
[11:28:39] tonyhansen joins the room
[11:35:04] Karen O'Donoghue joins the room
[11:40:39] paf joins the room
[11:43:01] paf leaves the room
[11:43:08] Karen O'Donoghue leaves the room
[11:46:26] Jim Galvin leaves the room
[11:53:36] bhoeneis leaves the room
[11:58:13] bhoeneis joins the room
[11:59:21] mnot joins the room
[12:09:09] Joe Hildebrand joins the room
[12:14:08] lef_jp joins the room
[12:19:35] Olafur Gudmundsson joins the room
[12:35:11] Joe Hildebrand leaves the room
[12:39:39] Joe Hildebrand joins the room
[12:52:16] Ning Kong leaves the room
[12:53:13] lef_jp leaves the room
[12:53:13] bhoeneis leaves the room
[12:54:25] bhoeneis joins the room
[12:54:53] tlr joins the room
[12:58:45] tlr leaves the room
[12:59:06] Olafur Gudmundsson leaves the room
[13:01:54] yuioku.yj joins the room
[13:05:13] tlr joins the room
[13:05:44] cabo leaves the room
[13:06:14] bhoeneis leaves the room
[13:06:22] mnot leaves the room
[13:06:25] jtrentadams joins the room
[13:07:24] Wendy Seltzer joins the room
[13:07:44] Joe Hildebrand leaves the room
[13:07:45] Ning Kong joins the room
[13:07:56] jtrentadams leaves the room
[13:08:18] Wendy Seltzer leaves the room
[13:09:28] lef_jp joins the room
[13:10:06] Joe Hildebrand joins the room
[13:13:15] bhoeneis joins the room
[13:13:26] mnot joins the room
[13:13:39] mnot leaves the room
[13:16:21] tonyhansen leaves the room
[13:16:25] cabo joins the room
[13:18:27] mnot joins the room
[13:19:07] lef_jp leaves the room
[14:08:22] mnot leaves the room
[14:12:23] tlr leaves the room
[14:14:53] Joe Hildebrand leaves the room
[14:16:44] Ning Kong leaves the room
[14:22:51] yuioku.yj leaves the room
[14:26:21] cabo leaves the room
[14:26:58] bhoeneis leaves the room
[14:32:38] tlr joins the room
[14:33:58] stpeter joins the room
[14:35:48] stpeter leaves the room
[14:49:02] bhoeneis joins the room
[15:17:57] tlr leaves the room
[15:22:41] bhoeneis leaves the room
[16:31:23] Ning Kong joins the room
[16:42:19] bhoeneis joins the room
[16:43:32] bhoeneis leaves the room
[17:24:22] tlr joins the room
[17:31:26] Ning Kong leaves the room
[17:45:50] tlr leaves the room
[20:12:57] tlr joins the room
[21:22:11] tlr leaves the room
[21:25:17] tlr joins the room
[21:25:40] tlr leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!