[04:03:54] Ted Hardie introducing the meeting [04:03:54] Rememer the NOTEWELL [04:04:14] Ted: agenda bash? [04:04:35] Nothing heard, onto open mic topics [04:05:15] 2 WGs just, closed, OPES and WebDav [04:05:16] IRI Recgnition + robots.txt + new work for open mike [04:05:58] CRISP, IMAPEXT, WIDEX expected to close prior to Chicago [04:06:44] Paul Hoffman: ATOMPUB may close too [04:07:33] Mailing lists for these will stay open after WGs close [04:07:58] Congrats to Lisa [04:08:01] Newest addition: Lisa isn't here because of birth of Darwin [04:08:41] Chris Newman: Incoming apps AD replacing Ted (applause for Ted) [04:09:32] Chris is taking over a little early because of Lisa'a absence [04:09:51] Eric Burger now discussing apps area review team. [04:10:13] 7 documents reviewed [04:10:51] Reviews occured within 1 week [04:11:05] one doc longer [04:11:15] 1 doc per-person [04:11:33] apps-review@ietf.org [04:12:12] To volunteer, send an email to eburger@bea.com [04:12:47] web page to come, [04:12:55] Review of BoFs [04:13:03] FSM BoF [04:13:11] (oops) FSM [04:13:26] Stephane Bortzmeyer speaking [04:13:41] finite state machine description language [04:13:53] BoF tomorrow [04:14:35] BoF is to discuss need for such a language, and which of several contenders [04:14:56] Call for Questions [04:15:16] Pete Resnick will co-chair [04:15:48] Ted whether there will be a mandate or not> [04:16:09] expects it won't but will be available [04:16:38] lots X area but inside apps for wg [04:16:48] No questions. [04:17:01] Peppermint next, Rich Shockey [04:17:19] co-chairing with Andrew Newman [04:17:52] Various ENUM efforts, chartered narrowly (e.164 -> URI) [04:17:58] enum is charter to mapping tel # to uri in dns [04:18:30] SPEERMINT defined to allow interconnect of various service providers [04:19:12] need to epand provisiong protocols [04:19:48] trust issues wrt routing data [04:20:11] Building on rfc 4114 and 4414 [04:20:35] aim for chichago [04:20:36] Want to formalize in Chicago. [04:21:04] previous work crisp Provreg [04:21:34] Call for questions [04:22:00] Dave Crocker: Doesn't understand the problem yet [04:23:00] Trust and policy are issues [04:23:34] Ted: People that want to talk to multiple entities shouldn't have to talk differently to different carriers [04:23:52] data flow is a mesh [04:27:08] Crocker: acknowledges he's not a phone company guy, looking for a way to describe it to others like him [04:28:14] Issue is that there may be multiple entities providing service for a given name (E.164 num) [04:28:57] Another issue is "who owns phone numbers" [04:31:00] Klensin: Data structures highly indirected, many pointers, raises anxiety level [04:31:31] Mark Nottingham re HTTP next [04:32:07] Bar Bof in Montreal [04:32:10] BarBoF in Montreal, 15 interested inrevising rfc 2616 [04:32:43] Not to add new features, but to fix ambiguities, make it more stable [04:33:06] rfc2616bis draft written [04:33:28] over 50 issues outstanding [04:33:35] some editorial some substantive [04:33:51] issues list @w3c web site [04:34:29] Test suite for HTTP needed [04:34:31] possible test suite as outcom [04:35:11] Want to minimize creative interpretations of spec [04:35:37] Plan to request BoF in Chicago [04:35:55] Hope to have proposals for issues by then [04:36:22] Most problematic issue is mandatory-to-implement security [04:37:46] Various flavors of authentication, for example [04:38:10] Are new auth mechanisms needed? Many potential answers and constituents [04:38:32] Questions? [04:38:35] single solution may not be possible [04:38:58] Dave Crocker at mic [04:39:11] dave croker: heard fix current problems and more strigent security [04:39:22] mayne seperate efforts [04:39:51] Ted Hardie at mic [04:40:17] Would benefit community to have a document describing threat model for each [04:41:25] Turns into a cookbook of what auth to use for what [04:41:59] A lot has been layered on top of http [04:42:44] More of a description of how to use existing auth well rather than new methods [04:44:18] Chris Newman: As incoming AD, will fight to document what's done and their limitations [04:44:58] Barry Leiba next [04:45:07] regarding notifications [04:45:23] 7 classes of notifications relating to email [04:45:51] lemonade, context, idle, vfolder, msg-events, sieve notify, SNAP, etc [04:46:09] Human readable and machine readable [04:46:44] List of desirable properties...no polling, etc. [04:47:21] Extensible for other uses [04:47:28] besides email [04:47:40] notifications@ietf.org [04:47:42] BoF or WG in Chicagp? [04:48:04] Eric Burger at mic [04:48:25] Lemonade should have this done by Chicago [04:48:38] Is this to clean up lemonade or boil ocean? [04:49:04] Barry: Doesn't want to block lemonade [04:49:22] Dave Crocker: What about lemonade is insufficient? [04:49:36] Barry: Need to check for gaps and generalize for other things [04:50:21] Chris Newman at mic [04:50:54] People have worked on protocols for server-server and server-client notification that aren't interested in the rest of lemonade. [04:51:07] Perhaps want to have a more focused list on the notification issue. [04:51:50] Eric Burger: lemonade is on a timeline, on the edge of relevance because running late [04:52:13] Lemonade has a lot of things to do, perhaps abdicate this [04:52:40] Barry: Didn't intend to have lemonade drop this [04:53:02] Eric: Will make people nervous if there is a sense it's being eclipsed [04:54:20] Chris Newman: leery of topic of a general notification solution. Wants a group doing email notifications in an extensible way [04:54:35] Better focused on things we understand [04:55:19] Do email first, then the general model [04:55:54] Ted Hardie (hat off): sailing across ocean, not boiling it. [04:56:58] Starting with a particular use case is good, but having it closer for other applications would be good [04:57:21] What about SIP notify? [04:57:51] DCrocker: Chris and Ted just said incompatible things [04:58:35] Work dies when focus is too broad [04:59:07] Eric stated a market confusion concern, which is valid [04:59:49] interprets Barry as saying "Lemonade notification can be broadened" [04:59:53] Barry nods somewhat [05:00:09] Eric is walking around trying to make eye contact with everybody... [05:00:42] Eric Burger wants to see who is interested that isn't a lemonader [05:01:12] If so, see Eric or Glenn Parsons [05:01:44] Show of hands: people not involved in lemonade that are interested in working on email notifications [05:02:08] (about 6-10 hands) [05:02:42] Similar show of hands for need for a separate WG [05:03:32] Ted at mic: Dave's point is valid, but broader question of whether we should still tackle this. IAB notifications workshop? [05:03:44] Barry will bring this up [05:04:14] Glenn Parsons: Concern that this will distract lemonade notifications work [05:04:41] Suggest that this not be chartered before Chicago [05:05:41] Pete Resnick: Suggests that Aaron might have a forum (IRTF) for woolier discussions [05:05:49] Ted disagrees that's a good idea [05:06:23] Pete: Not all IRTF things are all that researchy [05:07:08] Eliot Lear: confused. [05:07:27] Where to go to join mailing list? Not well posted on tools page [05:08:36] Chris: Wants to see (soon) somebody put together a model document for notifications [05:09:00] Part of this may fit well into lemonade, and part may lie outside [05:09:50] Barry: Trying to twist Philip Guenther's arm on that [05:10:36] Randy Gellens: Choice of venue will determine the scope that's addressed, how long it takes, etc [05:10:56] Happy to help [05:11:48] Eric: Points to lemonade agenda: Notifications being discussed about 5:15 today [05:12:27] Suggests interested parties come to that part of the meeting [05:13:26] Comment that lemonade notifications document is confusing. Eric: "fix it!" [05:13:49] End of topic. Next: Ted Hardie re ROAP [05:13:53] ROAP [05:14:04] Ted: Routing And Addressing Problem [05:14:07] BoF Wednesday [05:14:15] RRG last Sat [05:14:25] 60 or so attended, poorly publicized [05:14:43] second half of Plenary on Wed is really a ROAP BoF [05:14:58] Need lots of cross-area review [05:15:15] ID/Locator split needs early APPS area review [05:15:59] May have significant impact on referrals [05:16:23] But INT area meeting conflicts with lemonade and other things [05:16:46] Some proposals imply changing semantics of existing APIs [05:18:04] DCrocker: Ted may have underplayed seriousness of this [05:18:14] (ironic comment) [05:18:54] Abstract debate going on for quite some time (3-4 years thrashing). Dangerous. [05:19:02] Needs sanity; this group may help [05:19:19] Pete Resnick: Amplifies further on seriousness of this [05:19:42] multihoming [05:19:56] impact on roaming [05:20:06] Eliot Lear: Chaired IRTF effort this some time ago [05:20:29] 7 years later, we have some running code in this space (HIP) [05:20:53] Worth looking at HIP, what happens if your applications run on top of it [05:21:09] Things that look like IP addresses but aren't [05:21:37] Should collect info on what works and what doesn't into a document [05:21:57] Nark Nottingham at mic [05:22:02] Mark [05:22:14] Does this change things like URLs? [05:22:49] Ted: If the meaning of the IP address changes into an identifier that needs indirection, ability to pass referral to a specific point in the topology no longer works [05:24:57] Eliot: Problem is that the routing system on the Internet doesnt scale. Not just trying to add pain [05:25:26] Next: Open mic [05:25:58] IRI recognition (Yoshiro Yoneya) [05:26:31] [Aside: DCrocker looking for input for email architecture document] [05:26:38] Yoneya: [05:27:18] Many apps recognize URIs automatically [05:27:54] Utilization of IDNs and IRIs getting higher, [05:28:29] Apps need to recognize IRIs correctly, and take proper actions when users click on an IRI [05:29:20] URI delimiters (RFC 3986) [05:29:41] Whitespace delimiters ambiguous for IRIs [05:30:20] definition of wsp unclear [05:30:39] Perhaps use Unicode definition of wsp [05:31:13] Beginning of IRI probably unambiguous anyway [05:31:33] Inter-application IRI passing [05:32:23] Which form of IRI is used? IRI form, or ASCII-ized form? [05:33:33] Ted Hardie: Both forms are really IRIs. [05:34:53] Yoneya: Caller can't assume callee is IDN compliant [05:36:41] Discussion: Is whitespace discussion proposal acceptable? Assumptions re passing IRIs vs ACE form? [05:36:53] At mic: Pete Resnick. [05:37:33] Inter-process communication should use ACE for maximum compatibility. What is the concern with this> [05:38:38] Ted: Inter-protocol form should be the URI form of an IRI. There is always a URI form available. IRIs are really a presentation layer for the user. Protocol processing should use URI form. [05:39:09] Should do whitespace like URIs [05:39:35] But what you present to the user may very well be the IRI form [05:40:41] Yoneya: But many things can accept the IRI form [05:41:13] Pete: Not talking about inter-protocol comm, but inter app-comm [05:41:39] Good citizens who are callers do the ACE conversion first. [05:42:08] Patrik Faltstrom at mic [05:42:49] Copy and paste between apps typically has a negotiation to see what the recipient wants [05:43:00] Could have a similar thing here [05:43:36] Mac has done this for 15-20 years, perhaps in X windows too [05:43:59] Yoneya: Mac doesn't handle IRI correctly [05:44:05] John KLensin at mic [05:44:45] Applications in international environments should work in the way that makes sense locally [05:45:03] IETF doesn't typically specify user interfaces, it's not our specialty [05:45:50] Right-left writing example [05:46:00] People may be forced to ignore the protocols [05:46:54] Ted: URIs shown all have internationalization all in the domain portion [05:47:14] There must be a definition of how to make an IRI for each scheme [05:47:32] Cannot in general just say that any characters can go anywhere. [05:47:50] No such definition for http that Ted knows of [05:48:10] Discussions with W3C on this point [05:48:36] Some schemes have limitations [05:48:52] e.g., mailto: [05:49:12] Pete again: re whitespace [05:49:46] Whitespace a serious problem in free text URIs [05:50:07] How do you find the scheme in free text? [05:50:30] Full delimiting with angle brackets is the best appriack [05:50:33] approach [05:51:16] Perhaps push back to W3C to get something done about freetext [05:51:35] Randy gellens: Perhaps too late, whitespace delimiting is very common [05:51:56] (new speaker, didn't get name) [05:52:08] Also hope for delimiting [05:52:17] with angle brackets, for example. [05:52:44] My point is that I hope it isn't to late: delimiters (e.g., angle brackets) desperately needed [05:53:08] DCrocker: Repeating for emphasis [05:53:23] IETF doesn't have much track record on this. [05:53:45] This is a heuristic, so IETF shouldn't be working in that area [05:53:46] New speaker is Carrasco [05:54:20] Next topic: Harald Alvestrand [05:54:40] Request from a friend, re robots.txt [05:55:21] Shouldn't we standardize robots.txt? [05:56:26] Mark Nottingham: Agree, concerned that robots.txt is being cited in a legal context [05:56:48] With W3C liaison hat on, they have just opened a metadata working group [05:57:14] Another proposal called sitemaps being considered by some [05:58:04] Harald: We can define something new, but there is still a need to document what's there [05:58:46] Paul Hoffman: believes that he wrote a draft a decade or so ago, but there was no interest. [05:59:20] Reformatting should take less than an hour, probably in informational document. [06:00:25] DCrocker: Does IETF standardize existing practice? [06:00:49] Thinks standards-track may be useful [06:01:16] Mark N: What about addressing ambiguities? [06:01:29] Harald: If the spec is broken, fix it. If not, don't. [06:02:17] Paul H: General agreement that it's broken in some respects, but it's a really bad idea to make changes that would break something. [06:02:52] Do we want to do the real work on something improved? Is this the right place? [06:03:19] Chris N: If there are multiple interoperable implementations, a good spec, and community interest, then go for it. [06:03:30] Interest? 6 hands [06:03:49] (about 20 think something should be done) [06:04:23] Ted: Need to ask the person who started this work (Martin) about this. [06:04:34] Martijn Koster [06:04:38] Polling: [06:05:03] Just do it: 10-15. Mailing list: Smaller number. WG? no interest [06:05:29] Barry L: Start with ietf general mailing list, make a list if there's a lot of interest [06:05:48] Tony Hansen suggests using the apps list [06:06:12] coin toss results in decision to use of apps-discuss mailing list [06:06:28] End of defined topics [06:06:40] Now open open-mic [06:06:53] John Klensin: On apps-discuss, there are 2 drafts. [06:07:23] Network unicode things, please speak up sooner rather than later. [06:07:30] Last call is coming soon! [06:07:47] No other topics. [06:07:52] Meeting adjourned